Unique International College Pty Ltd v Australian Competition and Consumer Commission

Solicitor for the Respondents: Australian Government Solicitor

BETWEEN: UNIQUE INTERNATIONAL COLLEGE PTY LTD Appellant

AND: AUSTRALIAN COMPETITION AND CONSUMER COMMISSION First Respondent

COMMONWEALTH OF AUSTRALIA Second Respondent

JUDGES: ALLSOP CJ, MIDDLETON AND MORTIMER JJ DATE OF ORDER: 19 SEPTEMBER 2018

1. The appeal be allowed.
2. Declaration 1 of the orders of Perram J made on 8 November 2017 be set aside.
3. The first and second respondents pay the appellant's costs of the appeal.
4. The cross-appeal be dismissed with costs.
5. The matter be remitted to the primary judge for determination of the costs of the trial and for hearing as to further remedy. Note: Entry of orders is dealt with in Rule 39.32 of the Federal Court Rules 2011.

THE COURT: 1 This appeal concerns the unconscionability provisions in s 21 of the Australian Consumer Law (ACL), and in particular what it was necessary for the Australian Competition and Consumer Commission (ACCC) to prove in order to make out a case of a contravention or contraventions of s 21(1) by the appellant, Unique, based on allegations of a "system of conduct or pattern of behaviour" as those terms are used in s 21(4) of the ACL. The primary judge's reasons for judgment on the alleged contraventions of the ACL can be found in Australian Competition and Consumer Commission v Unique International College [2017] FCA 727. 2 We have concluded that there was an insufficient evidentiary basis before the primary judge to support the findings and declaration his Honour made that Unique engaged in a system of conduct or a pattern of behaviour in connection with the supply of online vocational education courses to consumers, and that the system or pattern of behaviour was unconscionable. Unique's appeal must be allowed. The ACCC's cross-appeal should be dismissed. 3 That conclusion as to the lack of evidentiary base for a system or pattern of behaviour case is to be contrasted with the findings that were clear, and uncontested on appeal, that the appellant had engaged in unconscionable conduct in its dealings with a number of individuals.

Some matters of background 4 Before outlining the facts of the case, it is appropriate to commence with a brief summary of the background to the courses offered by Unique. This background is relevant to some of the issues on the appeal, and to the question of what in Unique's conduct could properly be identified as unconscionable on the evidence.

The Federal Government's changes to the higher education system and the introduction of online courses 5 The Commonwealth introduced what is known as the "VET FEE-HELP Scheme" in 2007. The scheme has been amended from time to time. The acronym stands for the Vocational Education and Training Fee Higher Education Loan Program. Broadly, from 2009, the scheme provided for the Commonwealth to pay, in full, tuition fees for any approved course, on the basis that the amounts paid would be treated as a loan to the student, such loan to be repayable through the taxation system once a student earned above a specified income threshold. Relevantly to the period concerning the ACCC's allegations against Unique, that income threshold was set at about $50,000 per year. Approved course providers needed to have arrangements with an institution that offered a "higher education award" (such as a degree). That is because the scheme was designed to be a pathway into further higher education. 6 In early 2012, the Commonwealth foreshadowed substantial reforms to the VET FEE-HELP scheme. Those reforms were implemented through legislation enacted in the second half of 2012. The explanatory memorandum to the Higher Education Support Amendment (Streamlining and Other Measures) Bill 2012 (Cth) stated: The low take-up of VET FEE-HELP is an equity issue. People from identified demographic groups have a lower participation rate in education and training. These groups include Indigenous Australians, and people from a non-English speaking background, with disability, from regional and remote areas, from low socioeconomic backgrounds, and people not currently engaged in employment. Increased student take-up of VET FEE-HELP is key to lifting VET participation amongst these groups nationally. 7 One of the changes made by the 2012 legislation was to remove the need for a course to count towards a course at a higher education institution. This was consistent with the rationale for the changes expressed in the extrinsic material we have quoted above. Participation in vocational training was seen as an end in itself, and not only a pathway to higher education. Indeed, a media release from "Skills Australia" in May 2011 described the changes as a "sweeping overhaul of the country's vocational education and training (VET) system in order to help raise productivity, and address skills challenges that threaten future economic growth and prosperity". 8 The features of the policy that should be borne in mind in appreciating, and indeed not making unwarranted assumptions about, the evidence in this case were that (a) there was a decoupling of the courses from established tertiary institutions; (b) there was a focus of the policy on those from low socio-economic backgrounds, Indigenous Australians, people from remote and regional backgrounds, and the unemployed; (c) the courses could be delivered by the internet; (d) upon sign up the provider would be entitled to payment of the course fee from the Commonwealth; and (e) upon that payment it was the student who was liable to the Commonwealth for repayment of the tuition fee and a 20% loan fee, but with repayment only commencing with an income of above about $50,000 per annum. 9 Another aspect of the policy in practice under the scheme was that each student had an account with the Commonwealth which had a lifetime cap of $98,000. Thus, any taking up of the assistance diminished the likelihood of future assistance being available.

Unique's vocational education business 10 Relevantly, Unique was registered as a Registered Training Organisation by the NSW Vocational Education and Training Accreditation Board in October 2007. Initially it offered face-to-face vocational courses from its premises in Granville, NSW. Between 2008 (the end of the first financial year during which Unique offered vocational education courses) and 2013 (the end of the financial year after the changes to VET FEE-HELP were introduced), Unique's enrolment numbers were modest, ranging between 177 in 2008 and 789 in 2012, and reducing slightly to 631 in 2013. 11 During this period from 2008 to 2013, Unique arranged course credit transfer arrangements with a number of tertiary institutions, but was not able to obtain registration for VET FEE-HELP for its courses. All this changed, however, with the changes to the policy and Unique's registration. By the end of 2013, Unique had secured registration of its four vocational education training courses under the VET FEE-HELP scheme, which was made easier since those courses did not need to be eligible to count towards higher education courses, so there was no need for arrangements with higher education institutions to accredit Unique's courses, as there had been in the past. This was, we accept, how the new scheme was intended to operate. 12 From 1 January 2014, all courses offered by Unique were offered by way of online study. The courses relevant to this case (those eligible for VET FEE-HELP) were a Diploma of Salon Management; a Diploma of Management; an Advanced Diploma of Management; and a Diploma of Marketing. 13 It was not any part of the case made against Unique that any of these courses was inadequate or inappropriate. 14 It was not in dispute before the primary judge that Unique offered incentives for students to enrol in its courses. For example, one of its brochures in evidence described the incentives in the following terms: Enroll now and get a free I-pad or Laptop or $1,000 ($1,000 to buy i-pad or laptop of your own choice) (From 1 Jan 2014 till 30 Dec 2014) Enroll in any of our Diploma or advanced Diploma level courses and get first installment of special offer (Either an I-pad or a laptop or $500*) at commencement and get second installment of special offer (Either an I-pad or a laptop or $500*) at mid point. Ask us for more details. 15 Unique's case, and its evidence at trial, was that it had been told by various sources including the Department of Education and Training (as it is now described), that the offering of these incentives was permissible, and not in contravention of any standards applicable to Unique as an approved provider, nor in contravention of the ACL. 16 The offer, and provision, of "free" iPads and laptop computers was to feature prominently in the ACCC's case and in the primary judge's findings. The evidence was, and the primary judge found, that after 1 April 2015 the incentive programs were formally closed by Unique due to the change in Federal Government policy on the offering of such incentives and computers were not given away, but lent to students. However, during at least one enrolment session after this date (at Bourke), the primary judge found laptops were still being discussed as if they were being given away. Given the absence of any evidence from Unique that it sought the return of any laptop from the very many students who did not complete (or even commence their courses), the primary judge found that the laptop incentive program in substance continued. 17 Unique had also engaged in "referral" practices since it started offering VET courses in 2008. These referral practices involved students being paid money if they referred another student to Unique's courses, and that other student enrolled, and stayed enrolled, past the census date. This referral practice continued into the time when Unique was offering other incentives to prospective students. Unique also paid a commission to its staff when students were enrolled. 18 As we have noted, from 1 January 2014, the courses offered by Unique under VET FEE-HELP were all online courses; they were no longer tied to delivering learning at its Granville premises. Consequently, Unique could recruit and reach students in a very wide geographical area. Much of the evidence before the primary judge concerned recruiting and enrolment practices at a range of locations across New South Wales. The spread of Unique's enrolment practices outside New South Wales, and the primary judge's extension of declaratory relief to places outside New South Wales, forms part of Unique's grounds of appeal. 19 Nevertheless, it is not in dispute that Unique's enrolments jumped markedly after it came to operate under the changes that had been made in 2012 and after it became registered under the new VET FEE-HELP scheme. In 2014, 3,251 students were enrolled. In 2015, 4,677 students were enrolled, with corresponding increases in revenue as set out in the following table, compiled from the tables at [11] and [13] of the primary judge's reasons: Year Enrolments Revenue Net Profits after tax 2013 631 $1,702,612 $40,301 2014 3,251 $15,942,949 $8,214,031 2015 4,677 $56,183,632 $33,779,726

20 That revenue and consequent profits were, of course, obtained from the Commonwealth in accordance with the operation of the VET FEE-HELP scheme.

The ACCC's case at trial and the primary judge's findings in summary

The ACCC's case as pleaded 21 The case of the ACCC concerned contraventions of the ACL in connection with the supply of vocational education courses to consumers in relation to individuals, and in relation to a system. The period of time during which the ACCC alleged there had been contraventions by Unique ran from 1 July 2014 to 30 September 2015. 22 There were three causes of action, all contended to arise out of the same factual substratum: (a) contraventions of ss 18 and 29 of the ACL in relation to misleading or deceptive conduct; (b) contraventions of ss 74, 76, 78 and 79 of the ACL in relation to unsolicited consumer agreements; and (c) contraventions of s 21 of the ACL in relation to unconscionable conduct against the consumers, not against the Commonwealth. 23 The ACCC was successful to a large extent on each cause of action. Unique's appeal involves only part of the third cause of action, contraventions of s 21 of the ACL. 24 The primary judge emphasised (at [17]) that the ACCC's case did not involve allegations about the improper use of Commonwealth funds: But the Applicants' case is not a case about the misappropriation of Commonwealth funds, nor is it a case, despite the way it was sometimes depicted outside the courtroom, about rorting by Unique of a Commonwealth scheme. The Applicants' case was, from beginning to end, a case alleging the unconscientious exploitation and misleading of students. Such a case is by no means the same as a case which directly alleges that the Commonwealth has been defrauded. It is quite possible for the rorting of a scheme to happen without exploitation of students. It is also equally possible for exploitation of students to occur in the absence of rorting. … 25 The primary judge noted the ACCC's unconscionability case (the third cause of action) had two distinct elements: one focused on Unique's enrolment processes (set out in Part 2 of the Amended Statement of Claim dated 9 May 2016 (ASOC)), and the other (set out in Part 3 of the ASOC) focused on Unique's behaviour towards six named consumers. We shall describe these as the "system case" and the "individual consumer case". The appeal is only in relation to the system case. 26 Whilst we do not cavil with the primary judge's comments in [17] of his reasons (set out at [24] above), the assessment of unconscionability in relation to the consumers is to be made in the context of the scheme and Unique's obvious responsibilities under it. Such consideration may be important in understanding the full factual context in which unconscionability is to be judged.

The individual consumer case as pleaded 27 The individual consumer case can be briefly summarised, as it is not directly in issue on the appeal. However, an understanding of how that case was put at trial by the ACCC is necessary in order to understand the relevance of the categories of evidence which were adduced and relied on, and how that evidence did or did not overlap with the evidence relied on for the system case. As the primary judge explained (at [26]-[28]), the individual consumer case encompassed three sets of allegations, namely that: (a) Unique engaged in misleading or deceptive conduct in relation to five of the six individual consumers by not informing them its courses were not free and that they would be left with a debt to the Commonwealth under the VET FEE-HELP scheme. (b) In conducting its enrolment processes, Unique behaved unconscionably towards the consumers because it failed to explain the nature of what they were signing up to in circumstances where there were "elements of disadvantage" in relation to each of the individual consumers, and Unique adopted an "aggressive" enrolment process. (c) Unique had sought to have the six individual consumers enter into unsolicited consumer agreements without complying with the additional requirements of Division 2 of the ACL. As the primary judge noted, this conduct was also relied on as an indicator of unconscionability. 28 These cases were substantially made out, from which there was no appeal. 29 The six individual consumers referred to in the pleadings, in chronological order of the enrolment sessions held, were: (a) Ms Natasha Paudel, who enrolled in the session held at Walgett on Friday, 10 October 2014 (referred to in the ASOC as Consumer E); (b) Ms Kylie Simpson, who enrolled in the session held at Tolland, Wagga Wagga on Monday, 30 March 2015 (referred to in the ASOC as Consumer A); (c) Mr Tre Simpson, who enrolled in the session held at Tolland on Monday, 30 March 2015 (referred to in the ASOC as Consumer F); (d) Ms Jaycee Edwards, who enrolled in the session held at Bourke on Wednesday, 10 June 2015 (referred to in the ASOC as Consumer B); (e) Ms Fiona Smith, who enrolled in the session held at Bourke on Wednesday, 10 June 2015 (referred to in the ASOC as Consumer C); and (f) Ms June Smith, who enrolled in the session held at Bourke on Wednesday, 10 June 2015 (referred to in the ASOC as Consumer D).

The system case as pleaded 30 The system case should be explained both by reference to the primary judge's description, and to the pleadings. The system case was contained in paras 21 to 29 of the ASOC. 31 Paragraph 21 set out the ACCC's allegations about Unique's "process for marketing Unique courses, and enrolling consumers in Unique courses". The ACCC alleged that process had eleven features or categories of conduct: (a) targeting particular locations, including rural and remote towns, Indigenous communities and areas with significant populations of people with low socio-economic status. All the locations set out in the particulars were in New South Wales. They were not limited to Walgett, Wagga Wagga and Bourke and also "included" Bankstown (in Sydney), Boggabilla, Brewarrina, Emerton, Moree, Taree, Toomelah and Granville; (b) directing its employees to market to and enrol consumers in these locations; (c) having its employees visit these locations with "boxes of laptop computers to give to consumers" who signed up for a Unique course; (d) marketing at these locations including by way of calling on consumers in their homes and "group marketing" in private homes for the purpose of enrolling consumers; (e) employees representing to consumers, both at these locations and at Unique's campus in Granville, that in order to receive a free laptop, they needed to sign up to a course, and provide identification and personal information; (f) employees representing to consumers, both at these locations and at Unique's campus in Granville, that the courses were free, or free unless their income was, or reached, an amount they were unlikely to earn on completion or at all; (g) giving free laptops to consumers who signed up; (h) making incentive payments to its employees based on the number of consumers an employee enrolled; and (i) using (alleged over three sub-paragraphs) two kinds of forms (a VET FEE-HELP form and a student enrolment application/agreement), taking the personal identification and information provided by the students and having the consumers sign the two forms for the purpose of entering the consumers into an agreement with Unique and requesting VET FEE-HELP from the Commonwealth, which would constitute an agreement between the Commonwealth and each consumer. 32 Notwithstanding the terms of para 21, paras 22 and 23 of the ASOC also concerned the enrolment process. Paragraphs 21, 22 and 23 must be read together. 33 Paragraph 22 set out eight alleged failures of Unique, through its employees, during the enrolment process: (a) to ascertain whether consumers had the capacity to pay the course fees; (b) to explain, adequately or at all, the VET FEE-HELP scheme, the nature of a student's obligations under that scheme and that they would have a debt to the Commonwealth after the census date for each unit of the course; (c) to ascertain if consumers intended to undertake the course; (d) to ascertain whether or not consumers understood they were enrolling in a course; (e) to ascertain if consumers had read and understood the two forms they were presented with; (f) to explain, adequately or at all, what obligations consumers were assuming by signing the forms; (g) to ascertain whether consumers understood their obligations in signing the two forms; and (h) to provide consumers with copies of the two forms they had signed. 34 Paragraph 23 pleaded the consequences of the features or aspects of the enrolment process pleaded in paras 21 and 22. The consequences pleaded in para 23 focused upon attributes consumers were alleged to have had which (it can be inferred, and this allegation is made express in para 29) made them vulnerable. The attributes in para 23 were that the consumers Unique enrolled: (a) did not actually understand they had been enrolled in a Unique course; (b) could not read or otherwise understand the nature of the courses (including that they were online courses), and the nature of their obligations in agreeing to accept VET FEE-HELP assistance; and (c) could not access and use an email address, a computer or the internet, and had inadequate numeracy and literacy skills so that they were not capable of undertaking and completing the course they had been enrolled in. 35 For present purposes, what is important to note is that the ACCC's allegations in paras 21-23 were highly fact specific, as to the way Unique conducted its enrolment processes; what its employees failed to do; and what were the vulnerabilities of the consumers said to have been targeted by Unique. 36 It is also important to observe that at this point in the pleading there was no express allegation that Unique through its employees systematically failed to ascertain whether the consumer was suited to the course and the course to the consumer, though some of the specific allegations could be seen to relate to such a more general allegation. 37 The system case pleading continued in paras 24 and 25 with allegations concerning the consumer's ongoing exposure to VET FEE-HELP debt and the receipt by Unique of funding from the Commonwealth. At para 26 of the ASOC, the ACCC alleged that the reason, or motivation, for Unique's conduct was to maximise the number of consumers enrolled and receiving VET FEE-HELP, so as to maximise its revenue stream from the Commonwealth. 38 The allegation in para 26 carried with it an implied assertion that maximisation of revenue was at the expense of ensuring as far as possible that suitable people were being signed up for suitable courses. 39 Paragraph 27 of the ASOC alleged that the conduct concerning the six individuals (detailed at paras 31 to 142) occurred in the course of implementing the system pleaded in paras 21 to 26. 40 The ACCC alleged (at para 28 of the ASOC) that the consumers incurred three kinds of loss and damage as a result of Unique's conduct: (a) incurring a debt or potential debt to the Commonwealth in respect of a course that was unsuitable for them and which they were unlikely to be capable of completing; (b) anxiety and distress from an "unexpected debt"; and (c) reducing the entitlement of a consumer to further VET FEE-HELP in the future (assistance is subject to lifetime caps for each individual). 41 At para 29 of the ASOC, the ACCC alleged the unconscionability of Unique's conduct could be characterised in six ways: 29.1. Unique targeted vulnerable consumers in marketing and enrolling consumers in its courses; 29.2. Unique took advantage of its superior bargaining position relative to consumers, including vulnerable consumers, in order to enrol them in its courses; 29.3. Unique used unfair tactics, including offering inducements, in order to enrol consumers in its courses; 29.4. Unique gave consumers, including vulnerable consumers, misleading information, or failed to adequately disclose information regarding the nature of its courses; 29.5. Unique gave consumers, including vulnerable consumers, misleading information or failed to adequately disclose information in relation to their obligations if they received VET FEE-HELP assistance, including that they would incur a debt to the Commonwealth after the census date for each unit of the course, and the amount of the debt; and 29.6. consumers were enrolled in courses that were not suitable for them, and which they were unlikely to be capable of completing, having regard to their limited formal education, limited literacy and numeracy skills, and their lack of computer skills or access to email or internet connections at home. 42 It is important to note the terms of para 29.6, which drew out some aspects of the pleading which were hitherto less than fully explicit. The various inadequacies in what was alleged to have been done and not done can be seen to be worthy of criticism if persons are being targeted because of their vulnerable qualities, using unfair tactics and misleading or inadequate information to sign them up for courses which were unsuitable for them, which they would likely be incapable of completing and which would saddle them with a lifetime of debt, all in aid and furtherance of a purpose of maximising Unique's monetary revenues from the Commonwealth. The pleading has a certain deconstructed and particularised character to it, in circumstances where the unconscionability and the system can be better understood when described more coherently or holistically. The risk of deconstruction and particularisation is that it can lose the holistic interrelationship between all the factors, both as they affect the individuals and the system or pattern of behaviour. For instance, the targeting of the poor, ill-educated and socially and economically disadvantaged, whether these people are Indigenous, migrant or a member of some other category, was (when looked at alone) part of the aim of government policy. But it is an important and serious aspect of a case of unconscionability if such persons are targeted because of the ease of persuading them to sign up to courses for which they are unsuited by enticing them with lures of gifts and by misleading them, not so that they may be given the opportunity of obtaining a useful education, but so that Unique can maximise its revenue from the Commonwealth. 43 We will return to this question of the nature of unconscionability later in the discussion of the evidence. That is, the whole, rather than the deconstructed and particularised, nature of the concept. 44 Nevertheless, the declaration sought in relation to the system case, though deconstructed in its terms, sought to encapsulate the whole idea, as follows: 1. A declaration that during the period from July 2014 to September 2015, Unique, in trade or commerce, engaged in conduct in connection with the supply or possible supply, or marketing of the supply, of vocational education courses (courses) to consumers, that was, in all the circumstances, unconscionable within the meaning of s 21 of the ACL, by: 1.1 targeting particular locations, including rural and remote towns and indigenous communities and areas with significant populations of low socio-economic status (locations), to market its courses to consumers and to enrol consumers in its courses; 1.2 directing its sales representatives to market its courses to consumers at the locations, including by having the sales representatives call on consumers at their homes and also conduct group marketing sessions in private homes for the purpose of enrolling consumers in its courses; 1.3 its sales representatives representing to consumers that in order to receive a free laptop computer they needed to sign up to a Unique course and provide identification and personal information; 1.4 its sales representatives representing to consumers that its courses were free, or were free unless the consumer's income was in an amount which they were unlikely to earn on completion of a course, or at all; 1.5 offering inducements to consumers to enrol in its courses, including making cash payments to consumers and providing consumers with free computers and laptops; 1.6 providing financial incentives to its sales representatives in order to maximise the number of consumers enrolled by them in its courses; 1.7 its sales representatives failing to explain or adequately explain to consumers: 1.7.1 the nature of the Commonwealth VET FEE-HELP Assistance Scheme; or 1.7.2 the existence and nature of their obligations if they received assistance under the VET FEE-HELP Assistance Scheme; or 1.7.3 that they would have a debt to the Commonwealth after the census date for each unit of any course in which they were enrolled, and the amount of the debt; 1.8 enrolling in its courses consumers; 1.8.1 who could not access or use an email address; 1.8.2 who could not use a computer; 1.8.3 who could not use and did not have access to the internet; 1.8.4 who did not have adequate literacy or numeracy skills; 1.8.5 for whom the courses were not suitable, and which they were unlikely to be capable of completing. 1.9 Engaging in this conduct to maximise the revenue derived by Unique from payments to it from the Commonwealth under the VET FEE-HELP Assistance Scheme. 45 Relevantly for the appeal, the primary judge also noted at [29] of his reasons that: An interesting aspect of the case alleged in relation to the individual consumers is that it relied, in part, on the Applicants' more general system case pleaded in relation to Unique's enrolment processes. This can be seen, for example, at paragraph 48 of the ASOC (in relation to Ms Kylie Simpson) which explicitly picks up the system case pleaded at paragraphs 21 and 22. A related curiosity is that the system case at paragraph 27 of the ASOC alleges that the individual consumers' cases are to be seen as manifestations of the system case. It is possible this results in some circularity. However, it is best not to be distracted by these pleading issues in the abstract. … 46 As we have already noted, the ACCC's pleading at para 27 of the ASOC did indeed rely upon what was alleged to have happened in each of the individual cases as the conduct which occurred in the implementation of Unique's enrolment process: The conduct of Unique pleaded at paragraphs 31 to 142 below [the individual consumer case pleadings] occurred in the course of the implementation of Unique's marketing and enrolment process pleaded at paragraphs 21 to 26 above [the system case pleadings]. 47 It is this overlap in the way the case was framed, conducted and argued that appears to be responsible for the difficulties in identifying the evidentiary base for the system case.

The evidence at trial and the primary judge's findings 48 The primary judge examined in great detail the evidence of each of the lay witnesses called on behalf of the ACCC, and then those called on behalf of Unique. 49 Five out of the six named consumers were called. Mr Tre Simpson was not called, but his grandmother, Mrs Margaret Simpson, was a key witness for the ACCC. Aside from Mrs Margaret Simpson, and the five individual consumer witnesses, two other lay witnesses were called: one was a former employee of Unique (Ms Penny Martin) and the other was Ms Larissa Kidwell. 50 Ms Kidwell's evidence has a particular bearing on the system case. She gave evidence about an enrolment session at Taree on 27 November 2014: that is, towards the start of the period nominated by the ACCC as the period during which the alleged contravening conduct occurred and about six weeks after the events at Walgett that concerned Ms Paudel (one of the consumer witnesses). There were no allegations as part of the individual consumer cases about what happened at Taree. As the primary judge noted, Ms Kidwell's evidence about what happened there was only related to the system case. 51 Ms Martin's evidence was also relied on by the ACCC in its system case, and related to her attempts to contact, by email, the cohort of 80 students assigned to her after they had been enrolled in Unique's courses. In general terms, her evidence (as accepted by the primary judge at [218] and following) went to the high proportion of "bounce backs" she received on emails she sent out, with the remaining students not responding at all, together with the action she took with Unique's management to try to remedy, unsuccessfully, her inability to contact the students assigned to her. 52 The primary judge made a variety of findings, some favourable and some not, about the reliability of the applicant's lay witnesses, in terms of their evidence about the circumstances of the enrolment sessions, what they were told and not told, the filling out of the forms and the like. It is unnecessary to work through those findings in any detail. 53 Unique made the forensic choice to adduce considerable evidence to rebut the ACCC's allegations. That evidence included, but went well beyond, what occurred at the three locations on which the individual consumer case depended. Most of Unique's evidence was rejected by the primary judge on the basis of the view his Honour took of the witnesses' credibility and reliability. There was little in the way of corroborative evidence that his Honour was prepared to accept which added to the (small) quantity of direct evidence adduced on behalf of Unique that his Honour was prepared to accept. Thus, although there was a body of evidence adduced by Unique, in the end not much of it contributed to his Honour's findings. On appeal, Unique did not challenge his Honour's findings about its witnesses. 54 Unique called 14 lay witnesses, some of whom played a relatively minor role and some of whom played a central role in his Honour's fact-finding. In the latter category, for example, were Ms Mandy Kang, Unique's "student engagement manager" and the person who was at all or most of the enrolment sessions, Mr Amarjit Singh, the CEO of Unique, and Mr Manmohan Singh, a director of Unique and the father of Mr Amarjit Singh. The primary judge's adverse credit and credibility findings against almost all of Unique's witnesses were serious and involved a conclusion that he could not rely on anything they said. His Honour summarised his view about Unique's witnesses at [643]: I have concluded that all of Unique's witnesses were, in various ways, unreliable and have indicated my unwillingness to act upon their evidence unless corroborated. 55 We return to this part of the primary judge's fact-finding, his conclusions about what corroborative evidence existed, and what he was able to find, in our consideration of ground 1 of the appeal. However, as a general proposition, what can be said about the nature and state of the lay evidence at trial is as follows. The ACCC's evidence was focused on what occurred at the locations where the six individual consumers were enrolled. The ACCC primarily relied on the evidence about what happened at these locations (Walgett, Tolland (Wagga Wagga) and Bourke) for its system case as well as for its individual consumer case. Only 2 witnesses (Ms Martin and Ms Kidwell) gave evidence about matters going to the system case outside those locations. Ms Martin's evidence was in a small compass. Ms Kidwell's evidence related to one further site. It was not disputed at trial that during the relevant period Unique enrolled over 3,600 students across more than 428 locations. The ACCC ultimately relied to a considerable extent on the primary judge's adverse findings about Unique's witnesses. 56 The ACCC adduced expert evidence from Professor Tony Vinson. He was described appropriately by the primary judge as a distinguished social scientist. At the time he provided his report, Professor Vinson was Emeritus Professor in the School of Social Sciences at the University of New South Wales and Honorary Professor in the Faculty of Education and Social Work at the University of Sydney. Professor Vinson passed away while judgment was reserved. 57 There was no challenge at trial to Professor Vinson's expertise nor to the quality of his evidence. Rather, Unique challenged at trial, and on appeal, the way in which the primary judge used Professor Vinson's evidence in his findings. We discuss Unique's arguments in more detail in our consideration of ground 2 of the appeal below. Professor Vinson's expert report considered enrolment data concerning Unique's students, alongside data sourced from the Australian Bureau of Statistics. In the executive summary to his report Professor Vinson: (a) identified the locations for enrolments by Unique during the relevant period by reference to postcode, and indicated where there were postcodes with 15 or more enrolments. Included in this data were locations in Queensland and Victoria, not only NSW; (b) categorised which of these postcode locations would be described as "rural" or "remote", and as an Indigenous community, and why; (c) indicated what percentage of people in each location were Indigenous; (d) classified each of the locations by reference to the ABS Index of Relative Socio-economic Disadvantage, and then indicated the relative socio-economic disadvantage of the location where students who enrolled in Unique's courses resided; and (e) in relation to two specific areas studied (Wagga Wagga and Western Sydney) Professor Vinson concluded there was "clustered enrolment in sub-areas which also demonstrate higher levels of disadvantage relative to the broader location". 58 In other words, Professor Vinson conducted a data analysis of postcode information about Unique's student enrolments and used that to describe the socio-economic profile of the locations in which the students resided. 59 At [608], the primary judge described what Professor Vinson did as assessing the "socio-economic profile of Unique's students" (our emphasis). With respect, that might be something of an overstatement, because none of Unique's students outside the six individual consumers was specifically identified, and there was no evidence about their individual circumstances. In a way, this description by the trial judge reveals some of the difficulties which support Unique's arguments on ground 2 of the appeal. 60 As the primary judge pointed out at [610] of his reasons, the ACCC used Professor Vinson's evidence as the evidentiary foundation for submissions that students from remote locations, students from disadvantaged locations, and Indigenous students, were proportionally overrepresented in Unique's student enrolments. In answer, Unique made two submissions in substance. First, the comparisons drawn from Professor Vinson's evidence were the wrong comparisons, relying as they did on the general population rather than the overall VET population which Unique contended had proportionally higher numbers of Indigenous students, and students from disadvantaged and remote locations. Secondly, rather than providing any evidence of unconscionability (or a "system"), the proportions of students who were Indigenous, and/or from disadvantaged or remote locations, were a reflection of the student cohort the VET programs were designed and intended to reach. 61 The second of those submissions is a point well-made if proved "targeting" is said to be of itself indicative of unconscionable conduct. Without more, it may take the unconscionability case no further. However, that the government policy had the aim and purpose of running educational courses directed to marginalised or disadvantaged groups, does not make either irrelevant or virtuous the targeting of such groups if the conduct as a whole can be seen as the signing up of as many people as possible, to the maximum financial benefit of Unique, irrespective and uncaring of the suitability of the course for the consumer and the consumer for the course, using methods that involve misleading and inadequate information or other tactics that can be criticised.

The primary judge's factual findings on the individual consumer case 62 From [33] to [640] the primary judge undertook a comprehensive and careful discussion of each witness' evidence. From [641] to [709] the primary judge dealt with the individual consumer case. The primary judge structured his findings in this part of his reasons by reference to a heading entitled "Unique's Usual Practice", and then by reference to the four locations about which evidence had been given: Walgett, Tolland (a suburb of Wagga Wagga), Taree and Bourke. 63 It will be recalled that the evidence about Taree (from Ms Kidwell) was not adduced in support of the individual consumer case. It is unclear why his Honour considered it where he did in his reasons, since he expressly recognises at [690] that the evidence about Taree was only relied on to prove the system case. However, it may be explained by his Honour's earlier reference, which we have quoted at [45] above, to the ACCC's contention that evidence in the individual consumer case was relied on for the system case, so that his Honour considered it better to deal with all the location evidence in one place. The findings about "Unique's Usual Practice" self-evidently related to the system case. Logically, of course, as concerning "usual practice" they also concerned the individual consumer case. After rejecting the evidence of all of Unique's witnesses, all that the primary judge could take from the Unique witnesses was that there were short introductory sessions assisted with slides. This added little to the evidence concerning the individuals at Walgett, Tolland and Bourke. 64 In relation to each of the individual consumers at each of the locations, the primary judge made factual findings favourable to the ACCC's case. In respect of Ms Paudel and Unique's activities at Walgett (and reading his Honour's reasons at [671] with the declaration subsequently made in the Court's orders), it appears his Honour found Ms Paudel was not: (a) told that she would have a contingent debt; (b) given copies of the paperwork she signed; (c) given a copy of the VET FEE-HELP booklet; or (d) given written details of the courses. 65 However, the primary judge found that Unique's conduct towards Ms Paudel was not unconscionable. This was so because the primary judge found that Ms Paudel was told that she would be left with a debt if she signed up and she understood this. She was also told and she understood that she could withdraw from the course before the relevant census date. She knew "what was going on". See generally [759]. No case was made that she was deceived. On appeal, Unique relies on this finding to demonstrate how important the individual characteristics of each consumer can be in a case such as the present. We agree. We would add, however, that the fact that an individual was not deceived does not necessarily mean that conduct directed to that individual was not unconscionable. The conduct characterised as unconscionable may include unsuccessful, as well as successful, exploitation. 66 In respect of the enrolment session at Tolland, and the individual consumers Ms Kylie Simpson and Mr Tre Simpson, the primary judge made factual findings largely in accordance with the ACCC's allegations, although the primary judge did not uphold all of the ACCC's allegations in respect of Mr Tre Simpson. At one level for the purposes of this appeal, it is sufficient to extract his Honour's overall conclusion about Unique's conduct in relation to Ms Kylie Simpson and Mr Tre Simpson (at [689]): I therefore conclude in both Tre and Kylie's cases that the transaction was entirely exploitative. Unique received a large amount of money; Kylie and Tre received a lifetime contingent debt and an inexpensive device. 67 That said, it is illustrative of the seriousness of the behaviour and the clarity of the unconscionability in relation to the Simpsons to recite the findings at [686] to [688]: I make these additional findings: • no-one from Unique explained to Tre or Mrs Simpson the name of the courses involved; • they were not asked if they had an internet connection at home; • Unique's employees did not explain to them the obligations arising from the forms; • Unique's employees did not make any attempt to discern whether Tre understood what he was signing and in Kylie's case should have known that she did not; • Unique's employees did not give them copies of the forms they had completed; • Unique's employees did not tell them the cost of the course; • Unique's employees did not tell them they were incurring a substantial debt; and • Unique's employees did not tell them of their right to cancel before the census date. I find that it was entirely obvious to the Unique employees present that Kylie was totally unable to understand what she was doing. I do not make this finding in relation to Tre. I do not find that Unique employees believed him sufficiently competent, only that the Applicants have not proved the contrary which was their onus. On the other hand, I am satisfied that Unique's employees were quite indifferent to Tre's actual position. Actual and genuine inquiry on their part - as opposed to the form filling exercise embarked upon - would have revealed the problem soon enough. 68 In respect of the enrolment session at Bourke, and the individual consumers Ms Jaycee Edwards, Ms Fiona Smith and Ms June Smith, the primary judge again substantially accepted the factual allegations in the ACCC's case, and his conclusion about what happened at Bourke was (at [709]): I find that Unique's behaviour towards each of Jaycee, June and Fiona was exploitative and I do not accept that Adell was merely there to provide passive assistance to Johanne. She was actively involved in the process. That process involved luring Jaycee, June and Fiona with laptops to derive significant revenues from the Commonwealth which then saddled them with a lifetime contingent debt. Nor was any given a copy of the agreement. 69 Once again, some of the early factual findings mark out with clarity the starkly exploitative behaviour towards the consumers and its clearly unconscionable character. It is sufficient to refer to [704] to [708]: Although I have concluded that I should not rely upon Fiona's account on its own, Fiona's version in many ways resembles Jaycee's version which I am willing to act upon. Both were speaking to Johanne. In her discussions with Fiona and Jaycee, Johanne: • handed out a bundle of stapled forms which had X marked at various locations; • told them to sign at those locations; • took them through the forms quickly with no explanation; • asked them to provide identifying information; • told them that they did not have to complete the course if they did not want to; • did not tell them about the contents of the course; • did not tell them that they would be left with a debt; • did not tell them that they could cancel their enrolment prior to the census date; • did not give them a copy of the paperwork; and • did not inquire as to the level of their education. The forms were then signed and the laptops given to Fiona and Jaycee. June did not speak with Johanne, but rather Adell. June was shown a brochure but it was not explained to her. She was presented with the forms and told where to sign. They were not explained to her. She signed the forms because she believed she needed to do so to get a laptop. She was told the laptop was free. She understood that she was enrolling for some kind of course. No aspect of the VET FEE-HELP scheme was explained to her. She was not told she would be left with a debt. She was not told that she could withdraw from the course prior to the census date. She was not told that she could ask them to leave the premises immediately. She was not given copies of the paperwork she had signed. June had no use for the laptop which must have been apparent to Adell. Nevertheless, she was signed up for the course. And this was so even though she had no internet connection. 70 The primary judge's factual findings at [691] about Taree are properly to be seen as forming part of his findings on the system case, notwithstanding their location in the individual consumer case part of his reasons. Those findings were as follows: As I have already indicated at [227], I accept Ms Kidwell's account of the meeting at Taree. I therefore accept that Ms Kang's 'spiel' went for five minutes and omitted any explanation of the VET FEE-HELP arrangement or the right to withdraw from the course before the census date. It was only when she quizzed Rubbal further that Ms Kidwell discovered the true nature of what was being offered. I conclude that the process at Taree was, in essence, the same as at Tolland. Unique's employees' only interest was to get the paperwork completed without any interest in either explaining the true nature of what was on offer or the debt to which it would give rise. Whilst Ms Kidman [we think this is a reference to Ms Kidwell] was not ultimately exploited because she worked out what was going on, her evidence does allow me to conclude that what took place at Taree was, in general, exploitative of the large number of indigenous people there. 71 If the only interest of Unique's employees was to get the paperwork at Taree completed without any interest in explaining the offer or the debt, and it would appear without any interest in assessing the suitability of the courses for the consumers, one can conclude that at Taree, for financial motives and uncaring of the interests of the consumers, Unique, through its employees, did (and so had set out to) exploit a number of Indigenous people in Taree.

The primary judge's factual findings on the system case 72 On the system case, the primary judge described (at [710]) the "most controversial assertion" in the ACCC's case as the assertion that "Unique targeted particular locations including rural and remote towns and areas with significant populations of low socio-economic status for its enrolment process". At [711], the primary judge identified Professor Vinson's evidence as the principal evidence relied on by the ACCC for this assertion. Also at [711], the primary judge stated that he proposed to consider this case only as to NSW. This latter position is the subject of ground 1 of the ACCC's cross-appeal. 73 Noting the "targeting" case was a circumstantial one, the primary judge identified (between [713] and [718]) the factors he considered tended for and against the ACCC's assertion. We do not set them out here as we return to them when we explain our conclusions on the grounds of appeal. His Honour concluded (at [719]) that the ACCC had established the allegation that Unique targeted outlying areas, people of socio-economic disadvantage and Aboriginal people. 74 As to the other factual allegations made in the pleadings by the ACCC, the primary judge characterised them (at [720]) as largely involving "an attempt to extrapolate from what happened at the meetings in Walgett, Taree, Tolland and Bourke to a more general procedure", giving the examples of the allegations that the system included having Unique's staff represent to the proposed students that its courses were free, and that Unique had failed to explain adequately the contents of the sign-up forms. 75 Critically for some of the key issues on the appeal, the primary judge then said (at [720]-[721]): … At a high level of generality that kind of reasoning - if unsupported by anything else - would not permit one to conclude that the enrolment system had those features. To do that successfully, it would be necessary to conclude that the events at Walgett, Taree, Tolland and Bourke were somehow representative and to reach that conclusion it would be necessary to know how the Applicants had selected the nominated consumers in those towns from amongst the thousands of persons enrolled with Unique across the relevant period. I raised this with the Applicants before the trial commenced and Unique made the same point in its submissions. I do not think that any satisfactory answer by the Applicants was ever made to this point. Certainly, the Applicants made no attempt of which I was aware to explain how they had chosen the individual consumers and towns that they had. Without that information, I cannot rationally conclude that what took place in those towns was generally representative. (emphasis added) 76 Despite the conclusion at [721], in the same paragraph the primary judge then went on to find that he can identify "certain more limited systemic features" from Unique's own evidence: (a) The gift of computers to proposed students on signing up (either directly as a gift before 31 March 2015 or on a purported loan basis after that date); (b) The use of incentives for Unique's own staff to encourage them to sign up students; and (c) The holding of sign-up meetings at the targeted locations. There was no real debate that these findings were open on the evidence.

The primary judge's conclusions on the other alleged contraventions 77 As we noted at the outset of these reasons, the ACCC relied on three causes of action. 78 At [724]-[726], dealing with ss 18 and 29(1)(i) of the ACL (the latter of which relates to representations about the price of the goods or services), the primary judge made a number of findings about what Unique's employees or agents were required to explain to consumers in light of the nature of the "cohort" involved and the offer of free laptops. These findings are not the subject of any appeal and need only be noted. They do feature in the ACCC's cross-appeal. As we note below, the primary judge then went on to make findings about contraventions of ss 18 and 29 in relation to the individual consumers. 79 From [730] to [756] the primary judge explained the reasons for his conclusion that the ACCC had established a contravention of both s 79(b)(i) (failure conspicuously to include a notice informing the consumer of their right to terminate an agreement) and s 79(c)(i) (failure to include a form of notice for so doing). The primary judge further concluded that Unique's contention that its conduct fell within the "party plan" exception in s 69(4) of the ACL, read with reg 81(1)(c) of the Competition and Consumer Regulations 2010 (Cth), should not be accepted. In relation to the individual consumers, the primary judge found that there were some further contraventions of the unsolicited agreements provisions in ss 76 and 78 of the ACL and, in relation to some individuals, s 74. 80 Along the way to these conclusions, the primary judge found that the decision in Australian Competition and Consumer Commission v ACN 099 814 749 Pty Ltd [2016] FCA 403; 344 ALR 61 was plainly wrong. That disagreement concerned whether it was necessary, for the application of s 69 of the ACL, for the dealer to initiate negotiations. The primary judge concluded that it was not; Reeves J in that case had concluded that it was. Neither the primary judge's findings about this decision, nor his findings about contraventions of ss 74, 76, 78 or 79 of the ACL, are the subject of any appeal and also need only be noted. We do not set out below the details of the further findings made by the primary judge about the unsolicited agreements provisions of the ACL.

The primary judge's conclusions on s 21 81 At [757], the primary judge noted there had been little judicial consideration of the phrase in s 21(4) "system of conduct or pattern of behaviour", and noted that there were at least two limbs. First, where an "internal process is deliberately adopted", and second where a "process emerges without necessarily ever having been expressly articulated". His Honour stated that both limbs had relevance to the matter before him. 82 On the individual consumer case, the primary judge: (a) At [758]-[761], did not accept there were contraventions of s 21 in relation to Ms Paudel at Walgett. His Honour also appears to find (at [760]) no contraventions of ss 18 and 29 in relation to Ms Paudel. (b) At [762]-[765], accepted there were contraventions of ss 18 and 29 in relation to Mr Tre Simpson at Tolland, and also contraventions of s 21. (c) At [766]-[768], did not accept there were contraventions of ss 18 or 29 in relation to Ms Kylie Simpson at Tolland, but did find there was a contravention of s 21. (d) At [769]-[770], accepted there were contraventions of ss 18 and 29 in relation to Ms Jaycee Edwards at Bourke, and also a contravention of s 21. (e) At [771], accepted there were contraventions of ss 18 and 29 in relation to Ms Fiona Smith at Bourke, and also a contravention of s 21. (f) At [772], accepted there were contraventions of ss 18 and 29 in relation to Ms June Smith at Bourke, and also a contravention of s 21. 83 From [773] to [778] the primary judge set out his conclusions on the system case. Comparatively, it is a brief section of his Honour's reasons. A great deal of focus in argument on the appeal fell on these paragraphs. 84 It is necessary, however, to remind oneself that the primary judge concluded at [720] and [721] that he could not rationally extrapolate anything as to the system case from the findings as to the four locations. That was not the subject of complaint by the ACCC on appeal; nevertheless a thread of the argument that these events were somehow representative of a broader system or pattern remained. At [774], the primary judge drew together from the factual findings (and [719] and [721] in particular) the four features of the ACCC's factual case that he had accepted could constitute a system of conduct or pattern of behaviour: (a) the strategy of targeting disadvantaged people by reference to indigeneity, remoteness and social disadvantage (whether deliberate in its original conception or not); (b) the use of gifts of laptops or iPads to students signing (or loan computers after 31 March 2015); (c) the use of incentives to staff to encourage them to sign up students; and (d) the holding of sign-up meetings. 85 At [775], the primary judge explained why he declined to deal with the s 21 case in relation to allegations about conduct in Queensland and Victoria, finding a case about those States "was not clearly run". This is the subject of one of the ACCC's grounds of cross-appeal. 86 At [776], the primary judge explained why each of those features constituted a system or pattern of behaviour. As to the features (b)-(d), his Honour expresses a brief conclusion. As to feature (a), his Honour expands his reasons a little more: Matters (b) to (d) constituted a system within the meaning of s 21(4). They were each the result of considered decision making by senior management within Unique. This is not clear, however, in the case of (a). As I have indicated above, whilst I accept that targeting of the disadvantaged is what took place I remain unclear as to how it came about. But I am certain that it was happening. In that circumstance, it is appropriate to describe what took place as a pattern of behaviour within the meaning of s 21(4). However, I do not need to decide precisely what the mechanism was. For the purposes of considering the Applicants' case it is sufficient that I am satisfied that there must have been such a mechanism. I am satisfied of that proposition in New South Wales. 87 At [777], the primary judge reiterates that he sees the four features, together, as a "system": The Applicants therefore succeed in establishing within the meaning of s 21(4), the existence of both a system and a pattern of behaviour with the four features above. I will call the system and pattern thus identified compendiously, 'the system'. 88 Having made his finding about the existence of a system or pattern of behaviour, the primary judge turned to the characterisation issue, and concluded that the system was unconscionable (at [778]): The next question is whether this system was unconscionable. I do not think that (b) to (d) by themselves would necessarily be unconscionable. With the correct student cohort and management practices this style of operation may well have been permissible. However, when the practices in (b) to (d) are deployed against a targeted group of disadvantaged persons very different issues arise. In terms of s 22(1), it seems to me relevant to note in an assessment of the system that the targeted cohort consisted of people who were unlikely to understand the documentation involved (s 22(1)(c)) and that the use of the gift of a free (or 'lent') computer was apt to confuse this particular cohort into thinking a very bad deal was a good one - in my opinion an unfair tactic within the meaning of s 22(1)(d). The effect of the system in (b) to (d) was to supercharge the exploitation of the disadvantaged group which was being targeted (and also Unique's remarkable profits). The system was unconscionable within the meaning of s 21. 89 The reasons conclude with an explanation of arrangements for determination of appropriate relief.

Relief granted 90 Having invited submissions and after hearing the parties, the primary judge granted declaratory relief, over the objection of the ACCC: see Australian Competition and Consumer Commission v Unique International College (No 7) [2017] FCA 1289. The first declaration related to Unique's conduct constituting a system or pattern of behaviour. The next six sets of declarations related to contraventions of the ACL by Unique in relation to the six named individuals. 91 Since this is the only order impugned on the appeal, the form of the declaration made in relation to the system case should be set out: During the period from 1 July 2014 to 30 September 2015 (relevant period), Unique engaged in a system of conduct and a pattern of behaviour in connection with the supply or possible supply, or marketing of the supply, of online vocational education courses (courses) to consumers in New South Wales that was unconscionable within the meaning of s 21 of the Australian Consumer Law (ACL), by: 1.1. targeting disadvantaged people by reference to indigeneity, remoteness and social disadvantage; 1.2. offering gifts of laptops and iPads (or loan computers after 31 March 2015) to consumers to sign up; 1.3. providing financial incentives to its sales representatives to encourage them to sign up consumers; and 1.4. holding sign-up meetings.

Unique's grounds of appeal 92 There are eight grounds of appeal, all concerned with the primary judge's findings and conclusions on the system case. Ground 8 is something of a rolled-up ground, as we understand, designed to invite the Full Court to reach its own conclusion on the evidence on the system case if it finds error in the primary judge's reasoning, rather than to remit any aspect of the system case. We are satisfied that is the appropriate approach to take. That is because we are persuaded the evidence is insufficient to prove the system case on the balance of probabilities, both as to the conduct said to constitute the system or pattern of behaviour, and as to the characterisation of that conduct as unconscionable. 93 There was, however, a range of other relief sought by the ACCC in relation to other contraventions the primary judge found had been established, and for which no relief has yet been granted. The matter will need to be remitted to the primary judge in order to deal with the balance of the relief. 94 Aside from ground 8, Unique's submissions divided the grounds of appeal into three groups. First, there was a contention that the primary judge erred in his findings about the system case, because there was no evidence led as to the appellant's conduct towards the vast majority of its students (Ground 1). This ground occupied the most time in oral argument on the appeal and it is the principal basis on which we allow the appeal. Secondly, grounds which challenged the primary judge's first factual finding on the system case: namely that Unique had a strategy of targeting disadvantaged people by reference to "indigeneity" (the term used in evidence at first instance), remoteness and social disadvantage (Grounds 2 to 5). As we understand Unique's arguments, these grounds went mostly (if not entirely) to the primary judge's fact-finding, rather than his characterisation of the conduct as unconscionable, although we consider Unique recognised, correctly, that the unconscionability finding as made by the primary judge would inevitably fall away if these grounds succeeded. The reality of this is apparent from the terms of [778] of the reasons (as to which see [88] above). Thirdly, two grounds which challenged the characterisation of what was, as we understand the case, uncontentious evidence about the provision by Unique of iPads and laptops, as unconscionable conduct (Grounds 6 and 7).

What is not challenged on the appeal 95 We emphasise there are no grounds of appeal relating to: (a) the primary judge's declarations, or findings, in relation to the individual consumer cases under s 21; (b) the primary judge's findings on contraventions of ss 18 and 29 of the ACL; and (c) the primary judge's findings on contraventions of ss 74, 76, 78 and 79 of the ACL.

The ACCC's cross-appeal 96 The ACCC raised two grounds on its cross-appeal, both going to the primary judge's findings concerning contraventions of s 21 of the ACL. Both would have the effect of widening the relief the ACCC submits should be granted in relation to the contraventions, by widening the basis for the contraventions. 97 Ground 1 of the cross-appeal contends the primary judge should have found, on the ACCC's system case, that Unique's unconscionable conduct occurred in Victoria and Queensland, as well as New South Wales. 98 Ground 2 of the cross-appeal contends that the primary judge should have found there were five additional features to Unique's unconscionable system or pattern of behaviour, making nine features in total. The ACCC contends those additional five features were that Unique's representatives: (a) made false, misleading or deceptive representations to consumers; (b) failed to explain the VET FEE-HELP Assistance Scheme to consumers; (c) failed to explain the nature and content of the forms they had consumers sign; (d) failed to ensure that consumers were enrolled in courses that were suitable for them; and/or (e) failed to comply with the requirements for making unsolicited consumer agreements in Division 2 of Part 3-2 of the ACL. 99 During oral argument, it became apparent that the ACCC also contended that if any or all of the four features the primary judge found constituted the system or pattern of behaviour could not be upheld on the appeal, then these five features (or a combination of two or more of them, or a combination of one or more of these five features with one or more of the four features found by the primary judge) could nevertheless still constitute a system or pattern of behaviour and so the Court's declaration of a contravention of s 21 by Unique should stand. 100 We consider both grounds of the cross-appeal should be dismissed, and we explain our reasons for this at [232]-[249] below.

Resolution: Unique's appeal 101 In summary, we uphold the appeal on ground 1, which means the first declaration concerning contraventions of s 21 by reason of the system case must be set aside. We uphold ground 8, on the basis that, having considered the evidence for ourselves, but primarily taking into account our reasoning on ground 1, there is no basis in the evidence for any findings of contravention against Unique on the ACCC's s 21 system case as pleaded and as conducted. 102 Those conclusions are sufficient to dispose of the appeal. Nevertheless, we have considered the other six grounds of appeal, and would uphold most of them. It is necessary only to deal with those grounds briefly, given our conclusions on grounds 1 and 8.

The current state of authority concerning contraventions of s 21 by reason of a system of conduct or pattern of behaviour 103 In Kobelt v Australian Securities and Investments Commission [2018] FCAFC 18 at [176]-[178] the Full Court explained the introduction of s 21(4) and its equivalent provision in s 12CB of the Australian Securities and Investments Commission Act 2001 (Cth), in terms we respectfully adopt: Section 12CB of the ASIC Act was amended by the Competition and Consumer Legislation Amendment Act 2011 (No 184 of 2011) to add (among other amendments) s 12CB(4)(b) which is in the following terms: (4) It is the intention of the Parliament that: (a) … (b) this section is capable of applying to a system of conduct or pattern of behaviour, whether or not a particular individual is identified as having been disadvantaged by the conduct or behaviour; … This amendment commenced in operation on 1 January 2012. The Explanatory Memorandum for the Bill, when dealing with an identical amendment to the Australian Consumer Law (Schedule 2 to the Australian Competition and Consumer Act 2010 (Cth)), stated that the unconscionable conduct provisions of the Australian Consumer Law are not limited to individual transactions. After referring to the decision of the Full Court of this Court in Australian Securities and Investments Commission v National Exchange Pty Ltd (2005) 148 FCR 132 (National Exchange) (a case we will consider shortly), the Explanatory Memorandum stated: 2.23 … However, it follows from the principle that a specific person need not be identified that a special disadvantage is not a necessary component of the prohibition. 2.24 To emphasise this point, paragraph 21(4)(b) of the ACL indicates Parliament's intention that the provision may apply whether or not there is an identified person disadvantaged by the conduct or behaviour. This ensures that the focus is on the conduct in question, as opposed to the characteristics of a particular person, or the effect of the impugned conduct on that person. In his Second Reading Speech, the Parliamentary Secretary to the Treasurer said: The final interpretative principle to be introduced by the bill is that the prohibition on unconscionable conduct applies to systemic conduct or patterns of behaviour and that there is no need to identify a person at a disadvantage in order to attract the prohibition. Unconscionable conduct is not limited to individual transactions or events … This interpretative principle ensures that conduct, rather than individual transactions or events, is the focus of the provisions. 104 The extension of s 21 by para (4)(b) to a "system of conduct or pattern of behaviour" which is unconscionable removes the necessity for revealed disadvantage to any particular individual. A "system" connotes an internal method of working, a "pattern" connotes the external observation of events. These words should not be glossed. How a system or a pattern is to be proved in any given case will depend on the circumstances. It can, however, be said that if one wishes to move from the particular event to some general proposition of a system it may be necessary for some conclusions to be drawn about the representative nature or character of the particular event. The notion of unconscionability is a fact-specific and context-driven application of relevant values by reference to the concept of conscience: see Paciocco v Australia and New Zealand Banking Group Ltd [2015] FCAFC 50; 236 FCR 199 and Commonwealth Bank of Australia v Kojic [2016] FCAFC 186; 249 FCR 421. It is an assessment of human conduct. A system of conduct requires, to a degree, an abstraction of a generalisation as to method or structure of working or of approaching something. If s 21(4)(b) is to be engaged, it is the system that is to be unconscionable. Nevertheless, the concept of unconscionability (even of a system) is a characterisation related to human conduct by reference to conscience, informed by values taken from the statute. As Cardozo J said (speaking for the Court) in Lowden v Northwestern National Bank & Trust Co 298 US 160 at 166 (1936) (albeit in a very different context): "A decision balancing the equities must await the exposure of a concrete situation with all its qualifying incidents. What we disclaim at the moment is a willingness to put the law into a straitjacket by subjecting it to a pronouncement of needless generality." This expression of legal technique in the firmly gentle style of that great judge only reflects what other great judges of the tradition of Equity have said, such as in the passage of the judgment of Dixon CJ, McTiernan J and Kitto J in Jenyns v Public Curator (Qld) [1953] HCA 2; 90 CLR 113 at 119 adopting what Lord Stowell had said in The Juliana (1822) 2 Dods 504 at 522; 165 ER 1560 at 1567: "A court of equity….looks to every connected circumstance that ought to influence its determination upon the real justice of the case." These expressions of legal technique should be recalled when the temptation arises to seek to re-define in short terms the words chosen by Parliament that require the application of general values to factual and contextual circumstance by reference to the notion of conscience. 105 Although there has been comparatively little consideration of the function and operation of s 21(4) in a principled way, a number of previous "system" case decisions bear examination, in terms of how the regulator sought to prove a system case, and how the Court dealt with the evidence adduced. 106 Australian Competition and Consumer Commission v Excite Mobile Pty Ltd [2013] FCA 350; ATPR 42-437 was a proceeding brought under s 51AB of the Trade Practices Act 1974 (Cth), concerning the marketing, sale and supply of mobile telephone services through a telemarketing company based in India and other call centres in Asia. Potential customers were contacted by telemarketers and offered, through the use of scripts, an enticement to contract, namely the "gift" of a phone and holiday vouchers. Excite itself did not appear to contest the allegations, but the named individual respondents did appear and contested the case against them. Mansfield J did not consider what constitutes a system or pattern of behaviour in his reasons for judgment, but the level of evidence provided by the ACCC provides some occasion for comparison with the present appeal. 107 In terms of the nature of the evidence, on the uncontested case, Mansfield J described it in the following terms at [16]: The ACCC provided 10 examples of telemarketing calls and / or mobile contracts that were selected as follows. The first and last customers of Excite Mobile were chosen. The remaining eight customers were selected by choosing the last customer of every second month in which there was a recording available. 108 His Honour also noted at [131]: … There was no submission that the documentary or electronic evidence was inaccurate or selective. In particular, the sampling of various telemarketer calls as random, and therefore representative, in general terms, was also not challenged. 109 Another aspect of the unconscionable conduct in this case was Excite's debt recovery process. In terms of the evidence relied upon, Mansfield J stated (at [48] and [49]) the ACCC submitted evidence of 136 recordings with 98 different customers involving one of Excite's employees, and three recordings involving another employee, and the evidence about which consumers succumbed to the debt recovery pressure related to the agreements of approximately 90 customers. 110 There is no magic in these numbers: however, on any view, they demonstrate two matters. First, the selection of evidence of individual consumers through a random or representative process, with the process disclosed on the evidence. Second, the use of evidence about a reasonable number of individual consumers so that, even if intuitively, a Court could exclude hypotheses of coincidence or lack of representativeness, and could, because of the number of individual consumers about which evidence was led, safely perceive a "pattern". What happened with 80 consumers, even on a hypothetical sample of thousands (the estimated total number of consumers in Excite is not set out in the evidence), may well be sufficient evidence for a Court to decide whether or not there is a pattern. 111 In contrast, can conduct in relation to six consumers establish a "pattern of behaviour" or a "system"? As we explain below, in a known class of more than 3,600 consumers, and on the limited evidence before the primary judge, we would find they could not, at least not without persuasive evidence about how the six could be said to be representative of the 3,600. 112 The allegations in Australian Competition and Consumer Commission v Keshow [2005] FCA 558; ASAL 55-142 have some factual parallels with the current proceeding. Keshow was also a Trade Practices Act case, concerning a person Mansfield J described as a "humbugger", who travelled through Indigenous communities in the Northern Territory, purporting to sell and supply children's educational materials and, at a later time, household goods as well. At [3], Mansfield J described the conduct in the following terms: … He took advantage of the lack of education and commercial experience of those in the communities in doing so. In many instances, the educational materials were not needed or useful having regard to the age of the child or children of the consumer. The products he contracted to provide were most commonly not supplied, or not supplied in their entirety. Whether a contract to provide educational materials was met was haphazard. The payment arrangements in each instance involved an open-ended periodic payment authority, procured at the instance of the respondent, and authorising payment on the day which Centrelink or like benefits were regularly received by the particular complainant or other community resident. The respondent in a number of instances continued to receive periodic payments well after the value of the goods to be provided by him (whether or not they had been provided) had been received. In fact, there is no evidence to show that the respondent maintained adequate records of what products had been sold to which consumers, whether the products had been provided, as agreed, or what had been paid for them. 113 There were eight individual complainants from three Aboriginal communities, together with evidence from three people with some level of responsibility with one of the communities (Amoonguna). The evidence was that the respondent also visited other communities in the Northern Territory, including Hermannsburg, Meekatharra and communities at Katherine and Tennant Creek. However, it would appear the total number of consumers was not nearly as large as in the current appeal. 114 There was also a transcript of a lengthy interview with the respondent: it was from this evidence that Mansfield J said (at [76]) that there was other evidence of the respondent's dealings with customers other than the complainants. At [78], Mansfield J drew a series of inferences from the respondent's interview about his dealings with all his customers. The respondent's bank records and an analysis of those records were also in evidence. As Mansfield J noted, the evidence was ultimately uncontested by the respondent. There was also opinion evidence from an anthropologist, given on several levels: about the complainants individually, then concerning the complainants as a group and as members of their communities, and lastly concerning Indigenous consumers in remote communities in the Northern Territory as a class. On this last matter, the anthropologist's evidence extended to evidence of published socioeconomic information as to the characteristics of the communities in which the complainants lived, and published information about the educational and financial literacy of persons living in such communities. 115 However, as Mansfield J pointed out at [87], no generalisations could be made from this kind of evidence about all the residents in these communities, and their levels of education and financial literacy. We respectfully agree, and this is part of the difficulty with the point at which the ACCC's evidence before the primary judge stopped, in terms of what it disclosed about Unique's students. The Court must be careful to resist invitations to make findings that may involve the stereotyping of people living in certain locations, or in certain kinds of communities. 116 The same point was made by Drummond J in Australian Competition & Consumer Commission v IMB Group Pty Ltd (ACN 050 411 946) (in liq) [2002] FCA 402. IMB Group concerned a long-running and elaborate development scheme for a sporting and entertainment venture in Logan City, a local government area in the southern Brisbane metropolitan area in South East Queensland in the 1990s. The proponents of the scheme conceived of forming a syndicate for the purpose of entering a team in a national rugby league competition, and then developing a sporting and entertainment project centred on that team, whilst also taking advantage of the (then) recent relaxation of laws in Queensland about poker machines in sporting venues. Funding was to be generated through the sale of investment policies to individuals. It was the marketing and sale of these investment policies which were at the centre of the unconscionable conduct allegations. 117 More than 3,200 policies were sold by IMB, generally after purchasers attended seminars conducted by IMB, but also through door-to-door canvassing and telemarketing. Although initially the ACCC foreshadowed calling about 70 people who had purchased policies, in the end it called only 15 witnesses in this category. In contrast, the respondents foreshadowed calling 1,340 witnesses, 1,302 of whom were respondents to questionnaires, having attended meetings convened by IMB's legal representatives, and having then answered questionnaires either at the meetings or at home. From this, the respondent proposed to call 439 witnesses to give evidence. Drummond J did not allow this, and instead what happened was that the District Registrar randomly selected forty of these witnesses from the entire 439 and the parties agreed to Drummond J relying on his Honour's assessment of their evidence as an indication of the likely effect of the evidence of all 439 (see reasons at [40]). The ACCC is recorded in Drummond J's reasons as having made "trenchant criticism" of this exercise and the reliability of these witnesses because of the way they were "recruited". In the end, the respondents called 71 witnesses, including the "questionnaire" witnesses who had been randomly selected. The respondent enjoyed some success in defending many of the ACCC's allegations about the way the investment policies were marketed and sold, Drummond J concluding at [291]: There is such a mass of evidence contradicting the limited body of evidence produced by the Commission in relation to the issues concerning Council approval raised in pars 22 and 23 of the Commission's pleading that I am not prepared to find that the Commission has made out any case within pars 22 and 23 of its pleading. 118 We set this out as another example of how a representative sample of witnesses might be procured during a trial. IMB Group also demonstrates that once a regulator embarks on an undertaking to prove a system case, or a pattern of behaviour, then a number of forensic issues and challenges will arise, which in some cases may lead to the system case not being capable of being proved. Respondents are entitled to be given an opportunity properly to test the evidence of a system case against them - which is precisely the considerable forensic exercise IMB undertook in the case before Drummond J. 119 At [20], Drummond J said this about the "class" of consumers said to be the target of the sales strategies: … This class can be identified in general terms as residents of the Logan City area possessing the characteristics referred to in this "2001 Rugby League Syndicate" document. Some marketing occurred further afield, in Rockhampton and Moranbah in central Queensland (in about March 1992), in the Camu Aboriginal Community in far North Queensland (in about July 1992) and in parts of New South Wales, though most activity was concentrated in Logan City. But, as the evidence in this case demonstrates, it is wrong for the Court charged with determining whether conduct contravenes the provisions of s 52 the TPA to assume that the class of persons likely to be exposed to the relevant conduct fits some homogeneous stereotype conveyed by expressions such as "mum, dad and the kids in the middle to lower socio-economic group". As Hill J said in Tobacco Institute of Australia Limited v Australian Federation of Consumer Organisations Inc (1992) 38 FCR 1 at 49: "In the end, the question is not whether account is to be taken of the effect of the conduct upon the gullible, but whether the conduct in question is misleading or deceptive." 120 At [134], Drummond J said: While no one suggests that the Community was made up entirely of sharp business people, there is no justification to assume, on the limited evidence before the Court, that the Camu Community conformed to the stereotype of a non-functional indigenous community subsisting on welfare benefits and open to exploitation. If the Commission had established by evidence that the Camu Community was like that, then there would be little doubt that Glenn Ivers and Backo's activities would warrant the Commission's description of "appalling exploitation". Residents, as I have said, were members of an Aboriginal co-operative community engaged in banana growing. Mothe said he joined the IMB scheme, not because of club membership, but because he thought the scheme offered a good investment opportunity. That contradicts point (vi) in the Commission's submission. He was aware, however, that investment success was not guaranteed and that the scheme was still "at the planning stage". The reaction of the thirty to forty of the Community who attended Ivers and Backo's presentation was that about twenty members, including the administrator Mothe and the treasurer/secretary Shepherd, bought a total of about five policies while ten to twenty other members were not persuaded to invest. That does not suggest an obviously gullible audience. The Community's response to the seminar does not appear, on the limited evidence available, to be much different from the reaction of Logan City residents who attended IMB seminars at Springwood. 121 Returning to Keshow, it is of some significance that the kind of expert evidence adduced before Mansfield J extended well beyond the individual complainants. Nor was it simply statistical data, as Professor Vinson's was. At [101]-[106], Mansfield J made findings of unconscionable conduct in relation to the individual complainants. At [107], Mansfield J set out his findings in relation to all the respondent's customers (other than the complainants): I have also reached that view with respect to each of the respondent's dealings with the respondent's customers in the Northern Territory. Obviously, those persons would have differing degrees of exposure to commercial and business transactions and differing facility in communicating in English; some would in general terms have understood the nature of the periodical payment forms, and many would not. Consequently, as with the transactions involving the complainants, not all the factors to which s 51AB(2) directs attention would operate to point to unconscionable conduct on the part of the respondent in his dealings with each of the respondent's customers. But my findings as to the respondent's way of operating, to which I have referred in detail above, also lead to the conclusion that his conduct in relation to each of the respondent's customers was unconscionable as that term is used in s 51AB(1) of the TP Act. 122 At [108], Mansfield J explained why he did not rely on the anthropological evidence for his conclusions as to the individual consumers, although his Honour stated that the anthropological evidence supported his findings. He did not rely on the anthropological evidence for his findings about the respondent's other customers, we assume, for the reasons given at [87] of his reasons, to which we have referred above. 123 Australian Competition and Consumer Commission v EDirect Pty Ltd [2012] FCA 1045, a decision of Reeves J, concerned a mobile phone telemarketing scheme, again using overseas call centres. It was decided after sub-s (4) was inserted into s 21 of the ACL. Features of the scheme included offers of bonuses to potential customers if they entered into mobile telephone contracts, such as mobile telephones, televisions and laptop computers, and claims that the telemarketer had been "referred" to the consumer by someone the consumer knew. Obtaining names and numbers of people consumers knew, so that further "referrals" could be undertaken was a key part of the scheme. The ACCC's allegations concerned EDirect's conduct in relation to eight identified consumers. It also ran a system case. Aside from evidence about the transactions relating to the eight consumers, there was also evidence of the scripts of the sales component of the telemarketing calls, and there were recordings of 3,000 calls to potential customers. EDirect did not defend the proceeding at trial, but had filed a defence in which it made substantial (but far from complete) admissions going to the ACCC's system case. Reeves J records (at [88]) the ACCC's submissions: Relying on what the Full Court said in National Exchange (see at [70]-[71] above), Mr Burnside submitted that it was not necessary for a particular individual to be identified to make out this type of unconscionable conduct. Instead, he submitted EDirect employed a high pressure sales system that was directed to an unnamed group or class of persons that could be expected to include members who were vulnerable or susceptible to EDirect's sales process in the way described in the ACCC's statement of claim. (emphasis added) 124 Reeves J found the ACCC had not proven a system of high pressure sales. He found (at [94]-[95]) that the scripts did not establish enough features of any such system, and further held that the ACCC's evidence about the 3,000 calls also did not establish there was a high pressure sales system. Reeves J rejected (at [97]-[107]) the ACCC's detailed evidence (from several witnesses within the ACCC) about their analysis of the 3,000 call recordings, where (for example: see [98] of Reeves J's reasons) six ACCC officers "randomly listened to and noted features of [the] telemarketing call recordings". Spreadsheets of the results were created and adduced in evidence. It is not necessary to describe the detail of Reeves J's criticism of this evidence: suffice to say his Honour found it lacking in several respects. Part of his Honour's conclusions were reached after listening, as part of the trial, to approximately a dozen recordings played in court (see [103]). Reeves J concluded at [107]: It is true, as Mr Burnside submitted, that the ACCC does not need to show that EDirect employed the system of conduct in question in the majority of the 3,000 recordings, but only in a "significant number" or "a non-trivial number over the period". However, focusing on the number of telemarketing calls tends, in my view, to distract from the fundamental issue. That is, as stated above (at [91]), whether the ACCC has shown that the critical features of EDirect's sales system as pleaded in its statement of claim were present in combination in a sufficient proportion, or number, of the recordings such that it has proved, on the balance of probabilities, that the high pressure and relentless system which is central to this issue, existed as a fact. Because of the deficiencies in the ACCC's analysis I have outlined above, I do not consider that analysis proves that fact. … 125 Of course, these findings are specific to the proceeding in which they are made, and depend on the evidence in that proceeding. However, the comparison with the evidence before the primary judge in this case illustrates the difficulties we see with the ACCC's approach in the present proceeding in three ways. First, in the present case, there was no analytical evidence at all of the kind in EDirect. Secondly, EDirect provides an illustration (and only an illustration, but nevertheless instructive) of how an individual consumer case and a system case may require very different kinds of evidence. Thirdly, even if there had been distinct analytical evidence, the approach taken by Reeves J in EDirect illustrates that the Court's assessment of such evidence may well be a complex task, and the presence of such evidence is far from certain to discharge the ACCC's burden of proof. 126 None of the above is to set the burden of proving a system case too highly. As we discuss below unconscionability is a serious issue to which the terms of s 140(2) of the Evidence Act 1995 (Cth) are relevant. Some assertions of unconscionability will be more serious than others. That is inevitable due to the fact- and context-specific evaluation that it is necessary to undertake by reference to the values recognised by the statute. Regard should also be had to the ability of a party to prove or disprove the existence or nature of any system: Blatch v Archer (1774) 1 Cowp 63; 98 ER 969. 127 The AMI case (Australian Competition and Consumer Commission v ACN 117 372 915 Pty Ltd (in liq) (formerly Advanced Medical Institute Pty Limited) [2015] FCA 368; ATPR 42-498) was brought by the ACCC against a corporation trading as Advanced Medical Institute (AMI), its Chief Executive Officer, director and shareholder, Dr Jacov Vaisman, and two doctors engaged by AMI. We note that an appeal against this decision was dismissed: NRM Corporation Pty Ltd v Australian Competition and Consumer Commission [2016] FCAFC 98. The contravening conduct was said to be the offering of treatments for premature ejaculation and erectile dysfunction using nasal spray and later oral strips. AMI's revenue for this program ranged between $40 and $55 million between 2008 and 2010. Its method was telephone sales: men were "treated" over the telephone. The ACCC seized some 3 million recordings of these calls, and adduced a sample of them in evidence. Amongst the Court's findings, relevantly to the question of the existence of a system or pattern of behaviour that was unconscionable, the Court found AMI had designed a selling strategy to target the anxiety and distress of men seeking treatment for these conditions, and its salespeople used "high pressure selling techniques", including telling men of adverse physical consequences if they did not agree to the treatment (such as that their penis would shrink and they would suffer "psychological impotence"). The Court found there was no medical evidence that the sprays were effective treatments, nor were men warned about the side effects. 128 The AMI decision is of some assistance because in that proceeding the respondent made not dissimilar submissions in closing argument to the first two grounds of Unique's appeal in this proceeding. North J records that submission at [83]-[84] of his Honour's reasons: In their final written submissions the respondents argued 80. The ACCC brought forward a case based on particular patients, not a business practice or tendency case. For that reason, the case must be determined in relation to each transaction. There can be no elision or inferences drawn or tendency reasoning. The respondents repeatedly asked that the ACCC make its case clear, and a tendency or business practice case was expressly disavowed. It is the particularized case that the respondents are meeting. Therefore, while it is possible for there to be a finding of unconscionable conduct where no particular consumers are identified, that is not the case brought and is not the relief sought by the ACCC. Instead, the ACCC seeks relief in relation to particular transactions with particular patients identified in Annexures B and C. Only those transactions are relevant. This submission repeated an argument which was raised by the respondents on several occasions in directions hearings and in a strike-out application filed by the respondents on 12 October 2011 in the pre-trial management process. On each occasion the ACCC explained that its case was not limited in the way submitted by the respondents. 129 His Honour rejected AMI's submission, finding that at all times the ACCC had conducted the proceeding both on the basis of allegations of unconscionable conduct in relation to a number of individuals (approximately 168 "consultations" in all), but also as a system case. His Honour's approach to the argument about tendency evidence (at [92]) is one with which we agree, and return to in our consideration of ground 4 below. 130 At [95], North J extracted part of the ACCC's closing submissions, where the ACCC examined how it sought to use the evidence it had adduced, both as to individual consumers and as to AMI's "generic" conduct: In its written submission dated 10 July 2013 regarding the scope of its case, the ACCC sought to justify selecting the 168 cases as follows: By saying "AMI/NRM had, and implemented a business model" we are not suggesting, nor do we need to say or prove, that AMI or NRM engaged in this conduct (implementation of a plan) all or most of the time. To conclude that AMI and NRM implemented a business model does not, of necessity, require or imply reliance on tendency (similar fact) evidence. We do not mean, nor do we say, that "this is how AMI/NRM usually or typically acted", or "these examples are typical of what you would find if you sampled any instance or time during the period". We simply say, and mean, that AMI and NRM did certain things generically and certain things 170 times, and that those things, taken together (or separately or in more limited combination) reflected their intention - their 'model', and therefore their attitude towards consumers. 131 North J accepted this approach: see [939]. It is worth extracting his Honour's findings at [939] to [942] on the system in total, because they illustrate the nature of the evidence he considered: The ACCC demonstrated that AMI and NRM designed a system of conducting business which included some general elements. The nature of the advertising, the process by which patients interacted with AMI and NRM, the role of salespeople and their remuneration by commission, the contract terms and relevant refund policies, and the length and cost of treatment programs, were all matters which AMI and NRM implemented systemically and marked the way they conducted business. These systemic features could be seen in the individual cases which the ACCC proved. In addition, the 168 individual cases demonstrated some features which occurred in most of those cases, such as the offering of medications being limited to the AMI and NRM medications, the failure of doctors to diagnose an underlying cause or refer patients to GPs or specialists for those causes or presenting conditions, and the use of high-pressure selling techniques by salespeople. The individual cases also involved some features which were specific to those cases, such as the refusal of refunds in particular cases. The conduct which has been found to have been unconscionable existed in most of the 168 individual instances. Certainly, the most serious unconscionable conduct existed in most cases. That is not to say that all of the unconscionable conduct existed in all 168 cases. However, all the conduct found to have been unconscionable was sufficiently widespread to justify the relief sought, as discussed later in these reasons for judgment. Although the point has been made elsewhere in these reasons for judgment, it is useful to repeat at this point that the ACCC did not set out to show that the patients in each of the 168 cases suffered some disadvantage. Proof of unconscionable conduct depended on what AMI and NRM did, whether or not it had an effect on the individual patient. For instance, some of the Annexure B patients did not enter into agreements at all. That did not mean that, for example, in using high-pressure selling techniques in attempting to procure contracts, AMI and NRM did not engage in unconscionable conduct. The circumstances that no contract was concluded may bear on whether the Court would grant relief, but it does not prevent a finding that AMI and NRM engaged in unconscionable conduct. So much is clear from National Exchange and, now, s 21(4)(b) of the ACL. 132 With respect, we consider his Honour was correct to do so. In principle, there is no reason why evidence about what happened in the circumstances of a number of individual consumers cannot also be adduced as evidence to prove a system. 133 Nevertheless, the circumstances of the alleged unconscionable conduct, and the evidence adduced, will be critical. The proportion or distribution or some other feature that the individual consumers can be seen to represent of the entire consumer class may be important. Whether the class can be said to have substantially common relevant characteristics, or not, will also matter. In AMI, the class (men seeking treatment for impotence) all had a number of common characteristics which were what gave rise to their vulnerabilities. More individual attributes - level of education, literacy, socio-economic grouping - did not, or could be seen not to, matter. How many of the transactions involving individual consumers reveal features said to be part of the "system" will be important. Obviously, the more features which self-evidently have an unconscionable character, the easier it will be to prove an unconscionable system. For example, in AMI, some of the features included: (a) the training of salespeople in methods which were likely to frighten men into agreeing to buy the treatment programs; (b) the remuneration of salespeople by commission and the failure to disclose this fact in a context where men believed they were consulting a medical practice, which would characteristically have patient welfare as a primary concern; (c) the "unduly harsh" refund terms, which required parties to try all treatment options (including invasive self-treatment procedures) before becoming entitled to a refund or cancellation of debts, and the strict enforcement of such terms; (d) the exploitation of the doctor/patient relationship, which created an inherent power imbalance and was combined with concealment of the lack of scientific basis for the medications sold; and (e) the use of long-term contracts with consumers, in circumstances where this was without medical justification and was not necessary to protect the legitimate interests of suppliers. 134 Further, the nature of the evidence adduced from those responsible for the alleged design and implementation of the system will be important. In AMI at [891], North J summarised his findings about the evidence about Dr Vaisman: Dr Vaisman knew that men who suffered from ED or PE generally felt anxious about their condition. He believed that they felt frustrated, ashamed, dissatisfied, distressed, and that they suffered from low self-esteem. Dr Vaisman targeted these characteristics as a method of pressuring patients to agree to treatment programs. Salespeople were trained by him to take advantage of these feelings as a means to sell treatment programs. 135 Most critically, the nature of the allegations of unconscionable conduct will govern how probative the evidence of individual consumers will be. The more generic the alleged conduct, and the less the unconscionability depends on the attributes of consumers, the more probative evidence about what happened to a number of consumers may be. An example is EDirect and its telemarketing scripts and calls: the attributes of consumers did not play such a central role in those allegations. The facts of AMI are another example. The vulnerabilities of the male consumers were generic vulnerabilities, which could be said to arise from the very situation of seeking treatment from AMI: they were concerned about their sexual performance. 136 In contrast, in the present appeal, the vulnerabilities of the consumers were very much dependent on their individual circumstances: their levels of education, their literacy and numeracy, whether they had intellectual impairments, what was explained to each of them and what was not, and whether they had access to the internet and whether they understood how to operate a computer. These were not matters about which inferences could be drawn without sufficient evidence. 137 In AMI, the Court again had the scripts used in the telemarketing calls, and samples of the recordings of the calls. The unconscionable nature of the conduct was, at least in part, demonstrable from the contents of this evidence. That is not the case in the current appeal - for example, the scripts about the enrolment process and the programs were not found by the primary judge to contain, and were not suggested on appeal to contain, anything that could be said to be a feature of unconscionability. 138 The final case we should consider is the Full Court's decision in Kobelt (noting that special leave to appeal has been granted). Relevantly, Kobelt concerned findings of contravention of s 12CB(1) of the ASIC Act, which is materially identical to s 21. The appellant, Mr Kobelt, was a long term resident of Mintabie, an opal mining area excised from, but surrounded by, the Anangu Pitjantjatjara Yankunytjatjara Lands (APY Lands). Mr Kobelt had conducted a general store at Mintabie (called "the Nobbys") since the mid-1980s. He also conducted a second-hand car business from the premises, mostly selling to Aboriginal people who were residents of the APY Lands. If customers wanted credit (whether to buy the second-hand cars Mr Kobelt sold, or for other purchases such as food and groceries), they had to participate in what was called a "book-up" system. Mr Kobelt also used the "book-up" system to provide cash advances to customers, or to issue purchase orders customers could use at other stores on the APY Lands, if the distance to travel to Mintabie was too great. Briefly, the book-up system involved customers giving Mr Kobelt their debit card (or "key card"), and their PIN for this card. This card accessed the account in which customers' Centrelink payments were made. Mr Kobelt would keep the card until the debt was paid. As amounts were paid into the account, Mr Kobelt would withdraw all of the amounts, or nearly all of them. He would keep about half, and apply it to the customer's debt to him, and would make the other half (more or less) available to the customer to use at the store, or take in cash. However, the customer had to come to the store to get his or her half. Without going through the details of the primary judge's findings, the "book-up" system was conducted relatively informally, with little by way of written records, no application form and so forth. 139 Just in relation to the sales of second-hand cars, the evidence was that, between 1 July 2010 and 30 November 2012, Mr Kobelt withdrew a total of just under $1 million from the accounts of 85 customers to whom he had provided book-up. All but one of the customers who used Mr Kobelt's book-up facility in the relevant period were Indigenous, and nearly all were residents of the APY Lands. 140 ASIC's unconscionability case rested on the book-up system as the "system" or "pattern" of behaviour for the purposes of s 12CB(1) of the ASIC Act. The system was alleged to have two components: the giving of credit/deferral of payment on provision of the customer's debit card and, secondly, the way Mr Kobelt withdrew money from the accounts to satisfy the debt. ASIC also relied on individual contraventions of s 12CB(1) in relation to five individuals, four of whom gave evidence, although ultimately ASIC did not press for findings about these five customers. ASIC's case expressly relied on the disadvantages of Indigenous people living on the APY Lands, who, the Full Court said at [67], "in the overwhelming majority of cases, had very limited or no net assets, had very limited net income and had low levels of financial literacy". The Full Court noted that, by the end of the case, much of this was not disputed by Mr Kobelt. 141 At [69] the Full Court said: The primary judge considered the evidence with respect to the characteristics of the Nobbys' customers at some length. Before doing so, he indicated that he accepted that proof of ASIC's system case did not require proof of the individual circumstances of each customer to whom the system applied. We do not think that that observation had any effect on his conclusions, but we would note that the observation is true, but only to a point. Section 12CB(4) of the ASIC Act provides that s 12CB is capable of applying to a system of conduct or pattern of behaviour whether or not a particular individual is identified as having been disadvantaged by the conduct or behaviour. However, where an applicant's case is that the unconscionable conduct relates to a group which is disadvantaged, then it must prove that disadvantage. In any event, that was the case pleaded and conducted by ASIC. 142 The same anthropologist as had given evidence before Mansfield J in Keshow gave evidence in Kobelt. The Court conducted views of the communities in which some of the customers lived, and the stores (including Mr Kobelt's store). The Court also heard evidence from a financial counsellor working in these communities, including evidence about how the financial counsellor assisted about 30 Indigenous people to cancel their key cards and to obtain replacement key cards and PINs. The Court also had evidence from the operator of another general store on the APY Lands, who did not offer any such book-up facility, or any other form of credit. 143 Mr Kobelt called evidence from an academic from Adelaide University, with expertise in mathematical statistics. His evidence was that, using the Full Court's description at [104]: …one could not draw inferences concerning the Nobbys' customers unless they were randomly selected. They were not randomly selected and were the subject of a form of "sample selection bias". Professor Golnek gave his opinions as to why the Nobbys' customers could not uncritically be regarded as a random and, therefore, representative sample of all Anangu. 144 He also gave evidence that no valid inferences as to the characteristics of the Nobbys' customers, generally, could be drawn from the circumstances of the four individual witnesses called by ASIC. 145 While acknowledging the criticisms of Professor Golnek had force, the primary judge eventually concluded that most of the 117 customers had the characteristics alleged by ASIC and that included low levels of financial literacy. 146 On appeal, the Full Court appeared to accept the method adopted by the trial judge, but disagreed with the primary judge's characterisation of the appellant's conduct as unconscionable: see for example [268]-[269], [288]. 147 There were 14 grounds of appeal, and most were particular to the circumstances of the case and do not inform our consideration of the present appeal. It is fair to say that in Kobelt, as we have explained, there was a range of different kinds of evidence, tending very much in different directions. As the Full Court's discussions of the grounds of appeal reveal, there were aspects of the evidence which supported the way Mr Kobelt operated, such as the book-up system and his holding of people's debit cards, being a way for people to avoid the "demand-sharing" practices which operated in many Indigenous communities: see for example [239]-[252]. In other words, the evidence on both sides fully explored the characteristics of a majority of the consumers/customers who used the book-up system offered by Mr Kobelt, as well as exploring the merits and flaws of the system itself, from an evidentiary perspective that included expert evidence from a range of sources going to the substantive characteristics of both the system and the consumers who used it. The evidentiary picture was more complex and nuanced than in the present proceeding. 148 It is worth, however, referring to ground 8 of the appeal, which dealt with issues not dissimilar to those raised by Unique. The Full Court dealt with this from [228] of its reasons. Mr Kobelt contended that only six Anangu gave evidence and a finding of special disadvantage because of a low level of financial literacy was not open with respect to the 117 Nobbys customers. The argument was rejected at [232]: We reject the appellant's argument. The primary judge had evidence which supported his conclusion that most of the Nobbys' customers had low levels of financial literacy. The appellant's own evidence was that half of his customers could not read and more than half could not add up (at [283]). Furthermore, it was reasonable for the primary judge to conclude that even in the case of those who could read, their ability to do so was likely to be compromised. The primary judge also referred to the evidence of remoteness, his observations on the view, the evidence of Messrs Stauner and Kilpatrick and the evidence of the appellant himself. Of course, each item of evidence might be limited in effect. For example, it could be said that by virtue of his position, Mr Stauner was most likely dealing with the most vulnerable. Nevertheless, the primary judge was entitled to consider the accumulated force of the evidence and when that is done, we do not think that there is any error in the finding which he made. Ground 8 of the Notice of Appeal is rejected. 149 In our opinion, the nature, scope and extent of the evidence in Kobelt was qualitatively different from that adduced in the present case. 150 In conclusion, what the authorities demonstrate, unsurprisingly, is that the debate about whether or not a corporation or an individual has engaged in conduct that reveals a "system" or "pattern of behaviour" will be highly fact-specific, and will rely to a significant extent on the forensic exercise the regulator chooses to undertake to prove the existence of the system, as well as any forensic exercise the respondent undertakes by way of answer. The same is true of the characterisation of conduct as unconscionable. In a case like Kobelt, both parties adduced lay and expert evidence about whether it was correct to describe what Mr Kobelt did as unconscionable. ASIC persuaded the primary judge, but not the Full Court. 151 In upholding the appeal, we are not to be taken as deciding that a regulator needs to adduce evidence which covers a majority of consumers affected by the alleged conduct, nor as deciding that some kind of representative or random sampling method cannot be adopted to prove the existence of a system. Rather, it is clear from the cases we have discussed that all those methods and others are open, and may be sufficient depending on the particular circumstances of the case. 152 Nor are we to be taken as deciding that evidence of what occurred to individual consumers is incapable of being used to prove a system. As North J said in AMI, depending on how that evidence is presented, and the use sought to be made of it, it may not be tendency or coincidence evidence at all. If it is (and whether or not this is so will need to be determined on a case-by-case basis) then a regulator should be aware of the need to comply with the provisions of the Evidence Act. 153 Proving that a characterisation of unconscionability is appropriate also requires careful attention to precisely what it is about the class of consumers that is relied upon. The more that individual characteristics of consumers are said to be what makes them vulnerable, the greater the need for evidence about individual consumers is likely to be. The more that unconscionability depends on specific interactions between the respondents and consumers, the more likely it is that a level of representative evidence will be required. Where, as in the current appeal, the features of the alleged "system" depended on what happened at various enrolment sites, what was said and what was not, and what attributes those who attended and enrolled had, it is unlikely that an unconscionable system case can be proven without more attention being paid by the regulator to the need to prove representativeness of individuals, or to have a sufficient sample of individual consumers, or expert evidence, which addresses these matters.

Unique's Briginshaw argument and unconscionability 154 It is unnecessary to determine Unique's arguments about how the approach set out by Dixon J in Briginshaw v Briginshaw [1938] HCA 34; 60 CLR 336 at 362 (and now s 140 of the Evidence Act), as a general proposition, should be applied to allegations of contraventions of s 21 of the ACL, although some observations should be made. Unique contended its submission was supported by observations of the Full Court in Kobelt, as well as by the inclusion of unconscionable conduct in r 16.42 of the Federal Court Rules 2011 (Cth). 155 There is no third standard of proof: see Neat Holdings Pty Ltd v Karajan Holdings Pty Ltd [1992] HCA 66; 110 ALR 449 at 449-450 (Mason CJ, Brennan, Deane and Gaudron JJ). No doubt an allegation of unconscionable conduct is a serious allegation, to use the plurality's characterisation in Neat Holdings, and this explains its inclusion in r 16.42. (Though a court rule is not helpful in the construction and interpretation of a statute). To behave unconscionably should be seen, as part of its essential conception, as serious, often involving dishonesty, predation, exploitation, sharp practice, unfairness of a significant order, a lack of good faith, or the exercise of economic power in a way worthy of criticism. None of these terms is definitional. The Shorter Oxford Dictionary on Historical Principles (1973) gives various definitions including "having no conscience, irreconcilable with what is right or reasonable". The Macquarie Dictionary (1985) gives the definition "unreasonably excessive; not in accordance with what is just or reasonable". (The search for an easy aphorism to substitute for the words chosen by Parliament (unconscionable conduct) should not, however, be encouraged: see Paciocco at 266 [262]). These are descriptions and expressions of the kinds of behaviour that, viewed in all the circumstances, may lead to an articulated evaluation (and criticism) of unconscionability. It is a serious conclusion to be drawn about the conduct of a businessperson or enterprise. It is a conclusion that does the subject of the evaluation no credit. This is because he, she or it has, in a human sense, acted against conscience. The level of seriousness and the gravity of the matters alleged will depend on the circumstances. Courts are generally aware of the character of a finding of unconscionable conduct and take that into account in determining whether an applicant has discharged its civil burden of proof. We see no reason to doubt the primary judge was conscious of this: so much is apparent from some of the passages in his Honour's reasons to which we have earlier referred. We reject Unique's invitation to make some broader statement of principle about Briginshaw in the context of alleged contraventions of s 21. We do not consider that the findings of the Full Court in Kobelt at [206] and in following paragraphs, on which Unique relied in its written submissions, assist Unique's arguments: those passages deal with grounds of appeal in relation to the pleadings in that case. 156 The primary judge dealt with the nature of unconscionability at [727] of his reasons as follows: So far as the concept of unconscionability in s 21 of the ACL is concerned, the principles are also clear. What constitutes unconscionable conduct for the purposes of s 21 is an inquiry undertaken by reference to an evaluative statutory standard. One should not test the matter by reference to other non-textual evaluative standards such as 'moral obloquy' but should remain focussed on the word 'unconscionability' in s 21 in its statutory context. Thus one must survey the values and norms recognised in the statute itself, the place of norms and values in equity and commercial law and the guidance the statute itself gives as to its values. As Allsop CJ said in Paciocco v Australia and New Zealand Banking Group Limited [2015] FCAFC 50; (2015) 236 FCR 199 at 274 [296] the concept of unconscionability (which informs the related statutory definition of 'unconscionable conduct') : '…includes a recognition of the deep and abiding requirement of honesty in behaviour; a rejection of trickery or sharp practice; fairness when dealing with consumers; the central importance of the faithful performance of bargains and promises freely made; the protection of those whose vulnerability as to the protection of their own interests places them in a position that calls for a just legal system to respond for their protection, especially from those who would victimise, predate or take advantage; a recognition that inequality of bargaining power can (but not always) be used in a way that is contrary to fair dealing or conscience; the importance of a reasonable degree of certainty in commercial transactions; the reversibility of enrichments unjustly received; the importance of behaviour in a business and consumer context that exhibits good faith and fair dealing; and the conduct of an equitable and certain judicial system that is not a harbour for idiosyncratic or personal moral judgment and exercise of power and discretion based thereon.' 157 There was no submission put that the instructions his Honour gave himself were in any way inadequate. Reference may also be made to Paciocco at 265 [259] to 276 [306] and Kojic at 434 [55] to 436 [59] per Allsop CJ and 442 [85] to 442 [87] per Edelman J. Whether or not, and, if so, to what extent the conduct impugned as unconscionable involves a degree of reprehensibility or "moral obloquy" (see Paciocco at 266 [262]), depends on the nature and character of what is alleged. But there can be no doubt here that the essential criticism of Unique was the exploitation of, and predatory conduct in relation to, vulnerable people in order to maximise monetary gain from the Commonwealth. As senior counsel said, somewhat colourfully (albeit substantially accurately), his client was said to have been engaged in a scam. It may be, however, that the conduct can be characterised as unconscionable and articulated as such without going as far as counsel put it.

The ACCC's submission about "other evidence" to support the system finding 158 A plank of many of the ACCC's responses to the grounds of appeal was the submission that there was "other evidence demonstrating the operation of Unique's marketing and enrolment process, and the trial judge made findings about that process which were not limited to findings as to what occurred at those four sign-up meetings". The ACCC supported this submission, not by reference to the evidence so much as by reference to the primary judge's findings, at [6] of its written submissions as follows, to be read with [34]-[38] of its cross-appeal submissions: First, contrary to UAS [4] and [15], the evidence adduced by the Respondents below was not limited to the evidence of witnesses at the sign-up meetings in Walgett, Taree, Tolland and Bourke and the expert evidence of Professor Vinson. For example, there was evidence concerning arrangements for numerous other sign-up meetings, which his Honour (at J [650]) observed was "very relevant to determining what was going on" and evidence of the vast sums of money Unique was making, which his Honour found provided a motive to engage in the kind of consumer targeting which his Honour found was occurring (J [716]-[717]). More significantly, the trial judge made findings about Unique's conduct that were not limited to what occurred at those particular meetings, and those findings are not challenged by Unique in its appeal. As explained in the Respondents' Submissions in support of their Cross-Appeal filed on 11 April 2018 (RXAS) at [34]-[38], the trial judge made much wider-ranging findings about Unique's marketing and enrolment process, including that: Unique was involved in a scramble to sign up as many consumers as possible (J[650]); the introductory sessions at sign-up meetings were limited to only five minutes (J[650]) (and not 20-45 minutes, as variously claimed by Unique's witnesses); that in having consumers sign various forms Unique (through Ms Kang) had no interest in whether consumers had been given proper explanations of what they were signing up for, but was only interested in creating the appearance of a properly operating system (J[353]-[356]); and that Unique employees were willing to sign declarations falsely attesting to having provided consumers with explanations and information (J[276], J[279], J[446], J[483], J[527], J[556]). In many respects, the most powerful evidence of the unconscionability of Unique's system at trial came not from the Respondents' witnesses (its victims) but from the trial judge's rejection of the false evidence given about Unique's behaviour by its own employee witnesses. (references omitted) 159 The ACCC relied on this group of references both in its response to the appeal grounds about a "system", and about whether it was an unconscionable system. We do not consider the matters referred to take the ACCC's case very far on either aspect. We deal with his Honour's "scramble" finding at [197]-[198] below: it does not relevantly identify a pattern of behaviour, other than enrolments themselves, which is not how the ACCC's case was put. Other of these findings may reflect poorly on the credit of Unique's witnesses (and there is no challenge to his Honour's credit findings), but the findings the ACCC refers to are mostly findings by his Honour about what happened, again, at particular sites. For example, the ACCC relies on [353]-[356] where the primary judge makes findings about both what happened at particular sites (at [354]), and what the primary judge thought of Ms Kang's evidence about why she filled forms out (at [353]). However, a negative finding about Ms Kang's explanation for ensuring forms were filled out does not take the system case very far. We accept there are specific findings such as the one at [353] that can be found in his Honour's reasons. However, they are either inextricably mixed with findings about what happened at the sites (and so are difficult to disentangle from this context) or they are isolated and incapable even on a generous reading of supporting the finding made about the system at [774]. 160 The ACCC also set out at [39]-[50] of its cross-appeal submissions a number of what it contended were aspects of the evidence before the Court relating to Unique's marketing and enrolment process that the primary judge failed to address in his reasons. These were: complaints by consumers about Unique; the lack of ACL training given to Unique's staff; evidence concerning log in and completion rates for students in the courses, and the scale of Unique's referral and commission payments practices. We deal with all of these in relation to ground 2 of the ACCC's cross-appeal at [236]-[249], and conclude neither singly nor in combination are they matters that, had the trial judge dealt with them, would have resulted in a finding of a contravention of s 21 by Unique on the ACCC's system case.

Ground 1: insufficient evidence 161 Unique is correct to place at the forefront of its submissions the following facts arising from the evidence and not disputed by the ACCC: (a) During the relevant period, Unique enrolled in excess of 3,600 students. (b) Those students came from 428 different postcode locations across Australia. (c) The six named consumers represented 0.016% of the enrolled students. (d) There was no evidence that these students were representative in any way of the approximately 3,600 students. (e) There was no evidence as to how the students had been selected and indeed the ACCC expressly declined to adduce evidence about how they were selected. The primary judge expressly found he could not rationally conclude the six were in any way representative (see the primary judge's reasons at [720]-[721]). (f) Three of the six came from one family group (Ms Jaycee Edwards, Ms Fiona Smith and Ms June Smith at Bourke) and two of the remaining three were also from one family (Mr Tre Simpson and Ms Kylie Simpson at Tolland). (g) The enrolment events relied on from the individual consumer case were only three in number. 162 On the case as presented by the ACCC, it was erroneous for the primary judge to rely on the evidence from the individual consumer case as a significant part of the evidentiary basis for the system case. As we have sought to explain in our discussion of the authorities above, there will be a certain level of evidence necessary to distinguish alleged contravening conduct in relation to individuals, from alleged contravening conduct against a group of consumers that is contended to be a system or pattern of behaviour. While recognising the likelihood the evidence may overlap, the focus in the second category must be on what justifies the description of a system or pattern. Where the nature of the system and the nature of unconscionability depends on the attributes of individual consumers, or the specifics of transactions with individual consumers, there will need to be evidence about either a material proportion of individual consumers; or evidence about how and why the individual consumers were chosen; or evidence about the representativeness of the individual consumers, or a combination of all three. These categories may well not exhaust the possible categories of evidence. On any view, what was presented by the ACCC to the primary judge fell into none of these categories. 163 The system case findings were not supported to any material extent by the evidence of what occurred at Taree, nor by Professor Vinson's evidence. We explain why in our consideration of ground 2, below. 164 For the same reasons, we do not consider it was open to the primary judge to be satisfied that, singly or cumulatively, there was evidence that, in conducting itself in this way, using a "system" with the four identified features (see [84] and [91] above), Unique acted unconscionably towards the group of more than 3,600 consumers who were enrolled as students in its courses, or even a majority of them sufficient to clothe all its conduct in relation to these students with the description of "unconscionable". 165 We will deal with so-called targeting under ground 2. Even if Indigenous or poor or unemployed people were "targeted", that would at one level only be to focus upon the expressed aims of the government policy. It is important also not to make unwarranted stereotypical assumptions about these groups of people. It was a feature of many of the submissions of the ACCC, even if sometimes only implicit, that either the individual consumers and the conduct towards them could be taken as representative or, if not, assumptions of vulnerability should be made about all those "targeted". It was this latter silent assumption that not only underpinned many of the ACCC's submissions but also [778] of his Honour's findings. The use of gifts of laptops or iPads, the use of incentives to staff to encourage them to sign up students and the holding of sign-up meetings (see 774, (c) and (d) of the primary judgment), which could be an otherwise inoffensive system (see the second sentence of [778]), became unconscionable because of the targeting of Indigenous, remote and socially disadvantaged communities (without knowing anything at all in the evidence about those communities). 166 Unique submitted, and we accept, that the equivocal nature of the evidence was, in fact, recognised by the primary judge when at [653] he found there were matters that were left "rather up in the air" by the evidence. We accept Unique's submissions that those matters were central to the system case, and in particular whether Unique's conduct could be described as unconscionable. 167 What could have brought the evidence down out of the air? It should not be thought that we are of the view that the task was Herculean. Apart from Ms Martin there was no educational expert. In Kobelt, for example, there was considerable expert evidence about the Indigenous communities themselves. An educational expert might have explained how online courses operate in an education sense: the kinds of facilities necessary for a student to tackle them successfully and the levels of literacy necessary. 168 An educational expert might have given an opinion on the fact, well-established in the evidence as the ACCC submitted, that although Unique's revenue increased from $40,301 in 2013 to $33,779,726 in 2015, its expenditure on staff and resources did not increase at all. How, one might have asked an expert, would a program in a business such as this, enrolling more than 3,600 students to participate in online courses from remote locations and knowing many students might face literacy and social challenges, have been appropriately planned, resourced and delivered? Assertions from the bar table about the accounts were of no assistance. The absence of this evidence is exacerbated when one appreciates the evidence of Ms Martin. Ms Martin was a Unique employee whose evidence was dealt with at [215] to [224] of the primary judge's reasons. Ms Martin held qualifications in education and hairdressing. She had worked at Unique between 2007 and 2009, and worked there again during the relevant period from November 2014 to February 2015. In this latter period she was an "online assessor". After her training she was assigned some 80 students. She became concerned at the lack of response from these students. Her experience as a teacher led her to expect a better response. She spoke to Unique's CEO, Mr Amanjit Singh, who struck her as "passionate about education". Ms Martin's evidence was inconclusive but not on its face indicative of a system that was running exploitatively. The primary judge said at [222] and [223]: Gagen told Ms Martin that the college was going to give the students individual college email addresses. It is not clear to me how this would have solved the problem since at least half of the emails had been bouncing back from their intended addressees. My impression was that Ms Martin never really got to the bottom of her concerns about this. Despite that, she was aware that she had seen other trainers in the system receive feedback in the forums. There were also two occasions when she had met students face-to-face but she had no way of knowing if they were her students. Both occasions involved a group of retired, largely Filipino, students. Ms Martin gave evidence that her dealings with the education staff at Unique had been satisfactory in the sense that all of her queries had been answered promptly and the personnel were very nice people. 169 Little could be drawn from such snippets. What was required to characterise a system as unconscionable was some coherent evidence of what a proper system would look like compared to any inadequacy of Unique's system (on this hypothesis). 170 An expert familiar with the government policy could also have contextualised the policy which led to the changes in the VET FEE-HELP program, the assumptions underlying how it would be delivered and to whom, and what support the policy contemplated would be given to students, especially those in remote locations or with particular educational, or economic, disadvantage. The "failures" alleged in [22] of the ASOC could have been explored, at a systemic level, and through Unique's own documentation, with Unique's witnesses. A larger cohort than six individual consumers and four locations could have been selected. At the very least, evidence about how and why these consumers and locations were selected could have been adduced. A properly prepared representative sample could have been used. 171 The failure of the ACCC even to explain how the six consumers were chosen was a matter noted by the trial judge, but perhaps in the mire of evidence about the six consumers which reflected very poorly on Unique, by the time his Honour came to assess the system case, the absence of any proof of their representativeness had been overwhelmed by his Honour's impression of how unconscionable the behaviour of Unique's employees was to the individuals concerned. 172 It is true that there are a number of findings by the primary judge in that part of his Honour's reasons where he makes findings on the individual consumer case which are expressed in general terms and are capable of being seen as findings which also go to the system case. After all, his Honour's reasons disclosed a clear consciousness of the way the ACCC had pleaded its case, and had sought to use the allegations made (and therefore the evidence adduced) on the individual consumer case in support of the system case as well. Two examples of such passages are the following (we have bolded the general findings). 173 At [683], in the findings about what happened at Tolland: In relation to Tre, I find that some parts of the forms were filled out by Mrs Simpson with some assistance from Mel Connors who, it will be recalled, is Mrs Simpson's friend. Other parts were completed by Unique's employees. I conclude that the form filling exercise in Tre's case was not directed by Unique at ascertaining any information about Tre and was purely for the purpose of ensuring that paperwork was complete from Unique's perspective. Ms Kang, Jasmeen, Ms Merritt and Mr Bell were entirely indifferent as to Tre's suitability for the course. Their only interest was to make sure he enrolled. To the extent that the forms suggest a process by which Unique determined that Tre was suitable, I reject their reliability as evidence of that kind. The large number of enrolments which occurred during this period shows that this meeting was for the purpose of signing up as many people as possible. It had little to do with education. It was principally about raising as much money as possible before the curtain came down on incentives on 31 March 2015, the very next day. I am unable to make findings about what was said or not said to Kylie. The evidence from both sides is not sufficiently reliable. I do, however, conclude she would have appeared as patently unsuitable for the course for which she was enrolling due to my findings at [93], [137] and [320] in relation to her intellectual capabilities. (emphasis added) 174 At [693], in the findings about what happened at Bourke: Three introductory aspects of this matter should be noted. First, this meeting was not conducted by Ms Kang or Jasmeen. Nor were Rubbal or Mr Bell present. It therefore differs from the meetings at Walgett, Tolland and Taree. To the extent that the practices revealed at Bourke appear similar to those at Walgett, Tolland and Taree that may be material capable of suggesting that those events were not isolated occurrences. (emphasis added) 175 Nevertheless, we consider isolated findings of this kind are not capable of answering the point made by the appellant in ground 1 (and grounds 5-8) of the appeal. That is because, even though expressed more generally, these findings still rest only on the basis of what was proven to have occurred at three sites (Tolland, Walgett and Bourke) in relation to six consumers, and at one additional site (Taree) in relation to one additional consumer (Ms Kidwell). 176 The evidence about what occurred at Taree, principally from Ms Kidwell, adds something of probative value to the ACCC's allegations, but we consider it was materially inadequate to make out the existence of a "system" or "pattern of behaviour" on the targeting feature. It could not add much in terms of weight to the unconscionability aspect of the ACCC's case, as the only consumer from this site to give evidence, Ms Kidwell, was not vulnerable in any relevant sense, and was not taken advantage of in any relevant sense on the evidence. Whilst this does not mean the conduct was necessarily appropriate, as even unsuccessful attempts at exploitation can be unconscionable (see [65] above), it also does not, in isolation, materially advance the ACCC's allegations of an unconscionable system. 177 We agree with Unique that another way to approach the error in the primary judge's approach would be to evaluate how his Honour dealt with the factors set out in s 22 of the ACL, bearing in mind, as Unique submitted, that the statutory task is to determine whether the conduct was unconscionable "in all the circumstances": see, in relation to s 12CB(2), the equivalent provision in the ASIC Act, the observations of Keane J in Paciocco v Australia and New Zealand Banking Group Limited [2016] HCA 28; 258 CLR 525 at 620 [294], and Gageler J at 587 [189]. This was put expressly as a ground of appeal in grounds 6 and 7, and we have considered the detail there. 178 The ACCC contended that Unique's submissions "fly in the face" of recent authority of this Court which said, in Paciocco at 276 [304]: In any given case, the conclusion as to what is, or is not, against conscience may be contestable. That is inevitable given that the standard is based on a broad expression of values and norms. Thus, any agonised search for definition, for distilled epitomes or for shorthands of broad social norms and general principles will lead to disappointment, to a sense of futility, and to the likelihood of error. The evaluation is not a process of deductive reasoning predicated upon the presence or absence of fixed elements or fixed rules. It is an evaluation of business behaviour (conduct in trade or commerce) as to whether it warrants the characterisation of unconscionable, in the light of the values and norms recognised by the statute. 179 That statement of principle is not in doubt: in particular, it is not in doubt that for a system case under s 21, there will need to be a close evaluation of the business behaviour of the respondent or respondents. However, that evaluation will need to be conducted on the basis of probative evidence. That is the point of ground 1 of the appeal, and it is properly made.

Ground 2: error regarding expert evidence 180 This ground should also be upheld. The error can be seen at [627]-[629] of the primary judge's reasons. The table produced by his Honour at [629] should be set out: ….VET FEE-HELP statistics Australia wide against Unique's results in NSW: Indigenous Lowest 20% Very Remote Cities Unique in relevant period (NSW) 36% 54.7% 1.31% 0.03% VET general 2015 (Aus) 9% 27% 0.56% 69.17%

181 It will be recalled that the purpose of considering this evidence was to address what his Honour found to be the first feature of Unique's system - namely, the targeting of disadvantaged consumers. The relevance of that targeting in his Honour's reasoning and conclusions can be seen at [774] to [778] of the reasons. 182 The primary judge drew on material in evidence through the Commonwealth's Discussion Paper, released in April 2016 and entitled "Redesigning VET FEE-HELP: Discussion Paper", which included data about students who had accessed the scheme in 2014 and 2015. As his Honour noted at [622]-[623], of the 2014 VET cohort (nationally) approximately 6% were Indigenous. In 2015, that percentage was 9%. In 2014, approximately 25% of the VET student cohort were from "low socio-economic areas", and in 2015 it was 27%. As his Honour observed, and some might expect for a program whose objective was to draw in students from such cohorts into higher education, these proportions were greater than the proportions of these cohorts when compared to the general population. On one view, the higher these proportions within the VET program, the more the program might be said to be achieving its objectives. 183 The point made in response by the ACCC to the primary judge (and recorded at [625] of his Honour's reasons), is that the proportion of Indigenous students enrolling in Unique's programs in New South Wales was much higher - around 36%, so that (the argument appeared to go) Unique's level of enrolling Indigenous students far outstripped even the higher proportions of Indigenous students enrolling in VET programs across the country. That might be so, but again as Unique submitted on the appeal and we accept, evidence of the proportions - in and of themselves - are equivocal in terms of what they prove. The numbers are just as capable of proving that Unique was "targeting" its enrolments to those whom the Commonwealth had identified should be targeted. Further, a conclusion that the figures bespeak exploitative conduct makes an unarticulated and suppressed assumption about the necessary vulnerability of the Indigenous communities approached. 184 Where the error creeps in is, as the primary judge pointed out (at [626]) the ACCC had not provided any figures during final submissions. On appeal, Unique submitted, and we accept, the ACCC did provide correct figures in its final submissions. Those submissions do indeed set out in detail the percentages which can be drawn from Professor Vinson's evidence, as to the proportion of Unique's students who came from areas classified by the ABS as "remote" or "very remote", the proportion of students who were Indigenous and the proportion who were from low socio-economic backgrounds. What the submissions did not do was compare these proportions with what was in the Commonwealth Discussion Paper about the national VET student cohort. 185 However the appellant is correct to identify, and the ACCC did not dispute, that there are two difficulties in what the primary judge did. First, he compared the national rate (from the Discussion Paper) with the NSW rate for Unique. Secondly, his mathematical calculations were not correct, because he used the rate of enrolment of the appellant's students in very remote (1.31%) and major city areas (0.03%) and compared that to the proportion of students enrolled from very remote (0.56%) and major city locations (69.17%) from the Discussion Paper. 186 On appeal, Unique produced an amended table to the one created by the primary judge at [629]. Once the mathematical errors are corrected, the position is not as stark as the primary judge's table suggests in respect of the data as to major cities: 187 Contrary to the ACCC submissions on the appeal, the errors by the primary judge were material to his reasoning. His Honour's description of the figures as "stark" at [630] is an indication that he used the figures as contributing to his finding Unique was targeting these students. That is also true of his Honour's finding at [633] that Unique's enrolment profile featured "markedly more" people from these three categories. At [634], the primary judge rejected Unique's responsive submission by express reference to the "cities" figure in his table, which was incorrect. At [639], the primary judge relied on the table for the finding that Unique's student enrolment in these 3 categories "substantially exceeded" the corresponding profile of VET students. While even on the correct table produced by Unique this statement might still be made, we are satisfied that his Honour's impression about the level of disproportion was affected by the miscalculation. That view is confirmed by the primary judge's reasons at [712], in his findings on the system case. He refers to Unique's enrolments being "4 times higher" than the national data. In this paragraph the primary judge uses this kind of numerical comparison ("double", "twice as high") frequently. He also refers to the "cities" enrolment rate by Unique as being "2,300 times less than the general VET population across Australia". Plainly, the proportions affected his Honour's view of Unique's conduct. Further, the mathematics in the table had earlier caused the primary judge (at [346]) to reject Ms Kang's evidence about why Unique's employees had changed to visiting houses and holding sign up meetings away from the Granville premises. The primary judge used the erroneous "cities" percentage (0.03%) as part of his reasoning to reject Ms Kang's explanation. 188 At [714], the primary judge set out the hypotheses advanced by Unique at trial for the figures being as they were: namely, that its referral system and the fact some of its employees had contacts in certain areas, together with a small number of visits to remote areas resulted in a large numbers of enrolments in certain locations. We do not need to accept or reject this hypothesis. The point on appeal, which we have accepted, is that the ACCC had the burden of proof of an unconscionable system, and the figures - when accurately depicted - without any further expert analysis on the basis of them, were not capable of demonstrating any deliberate targeting (even if targeting was in the face of the stated government policy probative of or relevant to the proof of an unconscionable system). 189 Indeed, as Unique submitted, the figures themselves could tend in different directions, depending on how they were used. Unique submitted that: Nearly 98% of the Appellant's enrolments during the Relevant Period were from areas not classified as Very Remote, and 92% were from areas not classified as either Remote or Very Remote. That does not, on any rational interpretation of the data, reflect a strategy of "targeting" consumers by reference to the remoteness of where they lived. 190 Again, we do not need to accept the competing hypothesis from Unique. It is enough to find that the raw numbers tell one very little at all about a strategy or plan of targeting vulnerable consumers.

Ground 3: connecting demographic data with Unique's referral system 191 This ground challenges the primary judge's conclusions at [715], read with his Honour's conclusion of targeting at [719]: It is not totally clear what the skewing effect was. No-one tried to show its extent i.e. how much it reduced the effect set out above. Even if, however, the skewing effect is sufficient to be taken seriously, it is not inconsistent with the Applicants' case. The relevant data points are Taree, Tolland and Bourke where there were large numbers of enrolments (and to a lesser extent, Walgett). The evidence about those meetings, however, shows conduct exploitative of the persons it was directed at. That exploitative conduct is inconsistent with the hypothesis of a proper enrolment system having had the unexpected, but innocent, consequence of causing large numbers of enrolments at these locations. On the findings I have made, Unique's hypothesis must therefore become not only that the unusual enrolment profile was an unexpected by-product of its referral system and the identity of its staff, but that the exploitative enrolment practices revealed at Walgett, Taree, Tolland and Bourke were unfortunate coincidences caused, presumably by some sort of internal system failure. The Applicants' simpler hypothesis is that they were not a coincidence at all. 192 That Unique had a referral system, and used contacts that its staff had, as part of its enrolment process was not contested by Unique. Each of the visits to Bourke, Taree, Tolland and Walgett came as a result of personal contact or invitation between individuals and Unique's staff. It would appear, from his reasons at [715], that the primary judge accepted these facts. It can be accepted, as the ACCC set out in written submissions on the appeal at [16], that a word of mouth method was used to encourage people to attend the meetings, and the news of free laptops spread from one person to another. These matters demonstrate how news spreads in any community, especially small communities. 193 There are three points we consider sufficient to uphold this ground of appeal and find that the primary judge erred in the way he used the existence of a referral system, and how that referral system played out at the four sites (including Taree) as a foundation for his Honour's finding that there was deliberate targeting of locations which had the three characteristics in the demographic data of remoteness, "indigeneity" and low socio-economic status. 194 First, the primary judge's erroneous impression about how far Unique's enrolment statistics departed from the VET national statistics appears to have infected this aspect of his Honour's fact-finding as well. His Honour described the enrolment numbers at the four sites as "large" when, as the appellant submitted on the basis of calculations made using Professor Vinson's evidence, there were 104 enrolments at these sites, representing only 2.6% of Unique's enrolments. As with the other grounds of appeal and our overall conclusions, the disproportionate weight given to what happened at these sites is apparent, in circumstances where there was no attempt to show how the sites were representative. 195 Secondly, the primary judge again drew heavily on the evidence about what happened to the five individuals (recalling his Honour did not find Ms Paudel's enrolment unconscionable) to label, in [715], the enrolment practices as "exploitative" of the person they were directed to but this again relates only to the four sites. 196 Thirdly, on appeal the Court was not taken to any cross-examination of Unique's witnesses concerning this targeting allegation. There does not appear to have been any direct evidence about what Unique's employees (whether at a managerial level or at the level of people like Ms Kang) knew about the demographic make-up of the locations they went to, whether they chose particular locations (even the four in evidence) over others because of demographic characteristics, or whether they recruited Indigenous staff for the purpose of exploiting their family relationships or relationships with their communities. Again, the fact that Unique employed Indigenous staff would seem to fit with the objectives of the VET FEE-HELP program. No direct evidence was adduced by the respondents to support its allegation against the appellant. Insofar as the ACCC responds to this kind of argument by contending that Unique did not adduce evidence that what had occurred at Walgett, Taree, Tolland and Bourke were "errant failures of process" or that what happened was atypical, these submissions by the ACCC again ignore where the burden of proof lay. We consider that this was a common theme in many of the ACCC's submissions on the appeal. Unique was not obliged to prove its enrolment processes were "proper", or to prove it had an innocent or ordinary commercial explanation for why it enrolled people at particular locations. The burden lay on the ACCC. 197 The ACCC placed considerable weight on the content of text messages between Ms Kang and other employees such as Nyomi Whitton and Adell Richardson, which it contended revealed the nature of the referral "system". We are unable to see what probative value the text messages have. They reveal the interstices of Unique's employees conducting a mobile enrolment process across large areas of rural and regional NSW. They show how much the employees travelled late in the night and early in the morning, how they needed to arrange for laptops to be delivered, for forms to be available, how they dealt with complaints from people whose laptops were not working, how they were trying to schedule visits responsively to requests to visit particular locations and how enrolment appointments were being cancelled and rescheduled because of staff not being in particular locations. 198 It is true that at [650] the primary judge briefly mentioned the text messages, describing them as revealing a "scramble to sign up as many people as possible", in particular prior to the Commonwealth's deadline on 31 March 2015, after which providers could no longer offer "free" laptops and iPads. The primary judge does not otherwise rely on, or analyse, the text messages in Exhibit 46 to which the Court was taken during the appeal. The likely explanation is that they show little more than the level of (perhaps expected) chaos in attempting to enrol people from disparate locations over a comparatively short period of time. All of which on one view might indicate a level of commercial greed, but we fail to see how it takes the targeting allegations any further, and consider the primary judge was correct not to spend much time dealing with that evidence. Further, we were not taken to any cross-examination of Unique's witnesses about their text messaging practices and how it was said to reveal the strategies alleged by the ACCC.

Ground 4: impermissible use of tendency/coincidence reasoning 199 We would not uphold this ground of appeal, but the fact that Unique has raised it illustrates the forensic gaps in the ACCC's case in this proceeding. 200 Unique's argument again focused on what the primary judge said at [715], and his Honour's statement that "Unique's hypothesis must therefore become ... that the exploitative enrolment practices revealed at Walgett, Taree, Tolland and Bourke were unfortunate coincidences". 201 As Unique submitted, its argument had been rather differently framed: it contended the ACCC had "cherry-picked" the four sites. That submission was made in the context that all through the trial (and prior to it commencing) the ACCC had elected not to file any evidence, or give any indication by way of submissions, about how the individuals and sites were chosen, or how (if at all) they were said to be representative. In its closing submissions at [618], Unique submitted: In the absence of any evidence as to a random selection process, it is likely that the named consumers and Larissa Kidwell have been cherry-picked, following identification as specific (and therefore aberrant) instances of non-compliance (in the event their evidence were to be accepted uncritically, which the Respondent submits it should not) and/or merely the product of them being the only people who actually made complaints to the Applicants (or Fair Trading in the case of Ms Kidwell) during the Relevant Period. 202 Unique contended that his Honour's statement indicates that he was, impermissibly, reasoning that if Unique had engaged in the kind of conduct he had found occurred at the four sites, then it must have engaged in the same conduct at least at a material proportion of its other sites, and with a material proportion of its other more than 3,600 students. 203 If anything, the argument seems to be one relating to tendency reasoning (s 97 of the Evidence Act) rather than coincidence, or similar fact evidence (s 98). The evidence of what happened at one site was not adduced by the ACCC to prove what happened at other sites - because what happened at other sites, with other individuals, was not the subject of any express allegations by the ACCC. The Court was not invited to compare, for example, evidence adduced about what happened at Tolland, to prove what happened at another nominated town in New South Wales with other nominated students. 204 The ACCC could have relied upon evidence of conduct at a particular site with particular students to prove that, because Unique conducted its enrolment processes at a particular site, and recruited a particular kind of prospective student, it conducted its enrolment process in the same way, and in respect of similar students, at other sites for at least a material number of its approximately 3,600 students. It is possible that evidence of what happened at four sites, and in relation to six students, is capable of rationally affecting the likelihood that the same process occurred, and the same kind of students were recruited, at other sites: see Hughes v The Queen [2017] HCA 20; 344 ALR 187 at 193 [16]. The task is evaluative and somewhat open-ended: Hughes at 200 [42]. The question is whether the evidence supports the alleged tendency, and the extent to which the tendency makes more likely the material facts alleged: Hughes at 199 [41]. 205 The real question is: in order to prove which material facts? If, for example, what the ACCC had sought to do was to prove further aspects of an individual consumer case in this way, then we accept the process in s 97 would have needed to be followed, and any reasoning by his Honour towards fact-finding for other individual consumers would have been erroneous. 206 That was not the ACCC's case, as it submitted on the appeal. Rather, the ACCC relied on the same kind of analysis as that employed by North J in AMI at [92]: The evidence however is not tendency evidence subject to the Evidence Act. It is evidence of a system of operating which the ACCC sought to establish was itself unconscionable conduct. A similar situation was referred to Sackville J in Jacara Pty Ltd v Perpetual Trustees WA Ltd (2000) 106 FCR 51; [2000] FCA 1886 at [67] as follows: If, for example, the evidence in a shopping centre misrepresentation case shows that the lessor's agent gave instructions that particular representations should be communicated to prospective tenants, that evidence would [not be tendency evidence]. The evidence, if accepted, would go beyond proving that the agent had a propensity to make representations of the kind alleged. Rather, it would establish that the agent had set in place a system which, if implemented in the particular case, would have resulted in the representation being made to the applicant. The existence of the system, in the absence of evidence to the contrary, readily supports an inference that it was implemented in the particular case. The evidence of the system makes it more likely that the fact in issue (the making of the representation to the applicant) occurred, independently of the agent's propensity to act in a particular way. 207 This approach seeks to distinguish between a "system" or "pattern of behaviour" as an element of a contravention of s 21, and a s 21 cause of action based only on conduct directed at individual consumers. It is an important distinction. Unlike the circumstances in which s 97 and s 98 usually arise, which is where material facts (usually of conduct rather than state of mind) are alleged and evidence is adduced of other behaviours to attempt to prove the conduct alleged, in the case of reliance on s 21(4) the material fact is, itself, the existence of a system, or pattern of behaviour. Evidence of what occurred on numerous occasions, in different places and at different times, with different students, would all be admissible to prove the existence of such a system or pattern. There would be no requirement to adhere to the process in s 97 (or s 98 for that matter). 208 The real difficulty, which takes us back to the other grounds of appeal which we have upheld, is that evidence from four sites out of 428, and about six students out of more than 3,600, in the absence of any evidence about how the sites or students were chosen, or how they might be representative (or random), is in our opinion not capable of proving on the balance of probabilities to a degree of satisfaction necessary that Unique had a "system" or engaged in a "pattern of behaviour" that was unconscionable. The evidence proves what happened at those sites and to those students, but there is no real basis to extrapolate much beyond that, save where evidence was not in dispute or the subject of admissions about how what occurred at these sites also occurred elsewhere. 209 Therefore, the error in this passage of his Honour's reasons is the same error as we have already identified - the primary judge was erroneously prepared to accept the evidence from these sites as sufficient. His Honour expressed it differently here - by rejecting any notion of coincidence (even if wrongly attributed to the appellant), but the underlying error is of the same kind. 210 We would not uphold ground 4 of the notice of appeal.

Ground 5: no evidence of "targeting" 211 Ground 5 repeats many of the same arguments as put on the other grounds, but focuses on the primary judge's finding (in [774(a)]) of "targeting" as a principal feature of Unique's system. 212 We uphold this ground substantially for the same reasons as the other grounds upheld - we agree with Unique there was an insufficiency of evidence. We accept, as Unique submitted: While it may have been possible to find that the Appellant had a pattern of enrolling students from disadvantaged locations based on the evidence of Prof. Vinson (which may have been caused by any number of factors), to conclude that the Appellant had a pattern of targeting disadvantaged students would have required wide-scale evidence of the Appellant's deliberate intention to approach such students because of their disadvantage. No such evidence was produced at trial. (emphasis in original) 213 We do not accept the ACCC's submissions that some distinction can be drawn between the primary judge's observations at [776] (that it was unclear whether Unique's targeting was the result of considered decision-making by senior management and it was not clear how the targeting came about) and his conclusion (in the same paragraph) that he was "certain" that targeting was occurring. The latter was a central finding of fact in the primary judge's findings on the existence of a system, and also on its unconscionable character. It does connote deliberate conduct on the part of Unique and there was no probative evidence of any deliberateness. Again, we were not taken on the appeal to any cross-examination of Unique's witnesses about this matter. These aspects were central to the ACCC's case but appear to have been left to inference. Reference could also be made to the primary judge's findings at [350]: No doubt, the referral system was likely to bring in leads in various, possibly quite separate, areas. But it was Ms Kang (or at least Unique) who made the choice about which leads would be followed up. Ms Kang did not give evidence in her affidavit explaining how that particular decision was made. 214 This passage makes it clear there simply was no evidence about targeting - Ms Kang was the obvious witness to have given it, or to have been cross-examined about it.

Ground 6: free laptops as an unfair tactic 215 As we note earlier in these reasons, there was no dispute that Unique did offer "free" laptops and iPads (in the sense that there was no charge for them) as an incentive for students to enrol. There is no doubt on the primary judge's findings that at the four sites, there was considerable evidence that the "free" laptops did motivate people to attend the sign up meetings, and to enrol. 216 We note again that 104 students enrolled at these sites, and that some of the evidence about laptops went beyond the six named consumers, to other people who attended the sign up meetings. For example at [680]-[681] the primary judge found (in relation to the Tolland site): Following an earlier chance meeting between Mrs Simpson and Aunty Vennie, Mrs Simpson was at her home with Tre. Aunty Vennie arrived by car. She said something to Tre which caused him to run through the house saying 'Aunty Vennie has laptops'. Aunty Vennie then drove Mrs Simpson and Tre to her house. Aunty Vennie then returned to Mrs Simpson's house about 20 minutes later and brought Kylie and Regina back to her house. The evidence as to whether Kylie knew that laptops were on offer is not entirely satisfactory. However, I can imagine no other plausible reason why she would have signed up without the offer of one. It seems clear that laptops were on offer and that Aunty Vennie told Tre they were on offer. I infer that she must have been told something similar by someone. I do not find that the person who so informed her was from Unique. It was more likely a word of mouth occurrence. 217 Unique's challenge is to the finding by the primary judge at [778] that the use of a laptop as an incentive was an "unfair tactic". 218 Putting to one side the debate between the parties about whether the Commonwealth "approved" the use of laptops, it is clear on the evidence that at least until 31 March 2015 there was no prohibition on laptops being provided as an incentive and, the Department of Education knew (from Mr Amarjit Singh's email inquiry of them) that Unique was providing them. 219 It is clear (see [580]-[584] of the primary judge's reasons) there was a significant "spike" in Unique's enrolments just before 31 March 2015 (when it was required to cease giving incentives such as laptops) for courses, and the primary judge did not believe Amarjit Singh's explanations for the spike. This ties in with his Honour's later "scramble" finding about the way Unique was enrolling students at Tolland, which occurred only one day before the 31 March 2015 cut off. 220 There is no doubt that there was evidence of seeking to maximise the effects the incentives were having: Mr Amarjit Singh's own affidavit evidence, to which the primary judge referred at [585], was that he knew the laptop offer assisted in recruiting students. There is also evidence (and his Honour found) that the offer of laptops continued in substance (if not in form) after 31 March 2015, despite the prohibition on it. At [694] the primary judge found: Secondly, the incentive programs had been closed on 1 April 2015. Thereafter, Unique's documentation suggests that laptops were not given away but merely lent for the purposes of the courses. As I will shortly explain, it is tolerably clear that at the Bourke meeting they were still being discussed as if they were given away. Combined with an absence of any evidence from Unique that it sought the return of a single laptop from (the very many) students who did not complete (or even commence) their courses, this suggests - and I find - that the laptop incentive program in substance continued. 221 This finding is not challenged on appeal. 222 The real question is: what was the evidence the primary judge relied on to find the laptop incentive was an unfair tactic (as opposed to that it continued to occur)? 223 In relation to the individual consumers, the primary judge made findings such as those he made about Ms June Smith and Ms Fiona Smith at [697]-[699]: namely, that neither had an internet connection and both did not know how to use a computer when they were given one. However, his Honour does not find expressly that it was unfair to persuade a person to enrol in an online computer course by offering a free laptop when that person did not have an internet connection and did not know how to use a computer. That is, his Honour does not connect this evidence with his finding at [778]. It is not difficult to see how such a conclusion could have been reached: plainly this conduct could have been seen as an unfair tactic because people who were not experienced with computer technology were encouraged to incur a debt for a course they needed to undertake using a computer and using the internet, when they had neither skills nor access to the internet, and there was no evidence Unique was giving them any computer training. At one level, this kind of reasoning could have been accepted, although his Honour did not expressly reason this way. 224 The difficulty comes back to the one we have now identified many times. Proving that it may have been an unfair tactic to "lure" (the primary judge's word) Ms June Smith and Ms Fiona Smith to enrol through the gift of a device they could not at that time use to do the course due to their personal circumstances may establish the use of an unfair tactic for those two individuals. However, without direct evidence, or an evidentiary base for an inference, about how many other students enrolled by Unique also had no internet connection and did not know how to use a computer, it was not open to the primary judge to make a finding Unique had in a systemic sense used an unfair tactic.

Ground 7: confusion whether a bad deal was a good one 225 Like ground 6, this ground of appeal turns on the primary judge's conclusionary finding at [778] of his reasons. Indeed, this finding is rolled together with the "free" laptop finding. 226 Unique's submissions link this finding at [778] back with the primary judge's findings on the ACCC's misleading or deceptive conduct case at [725]-[726]: The debt created by the VET FEE-HELP scheme was readily able to be understood by people who had no expectation of earning more than $50,000 as being the same as free. But free it was not. It reduced their ability to enrol in future vocational training courses. Such a scheme, when combined with the attractive offer of a free laptop, created a powerful impression of a very good deal. That powerful impression was incorrect, however. Having engendered such an impression with the laptops, it was misleading not to explain in the clearest terms precisely how the VET FEE-HELP scheme worked and the fact that it would leave each person who took the laptop with a lifetime debt as well as a reduced ability to access the VET FEE-HELP system in the future. And, the kind of explanation which was called for was one which was tailored to the audience which had been persuaded to attend the sign-up sessions by the lure of a free laptop in the first place. The cohort involved, even on Unique's case, was a cohort in which disadvantaged persons featured. Such a group was even more exposed to the lure of the laptop than the general community. Thus, an even clearer explanation was called for. Perhaps put a little less formally, it was misleading to offer free laptops to groups of poorly educated and/or illiterate people on the basis that they sign up to VET FEE-HELP courses without explaining in the plainest of terms what the ramifications of this would be. 227 As we understood how Unique's case was put on appeal, there was no challenge to the primary judge's findings of contraventions of s 18 and s 29 of the ACL. That being the case, this ground must be understood not as challenging the primary judge's findings at [725] and [726] directly, but rather challenging the way the primary judge went on to use those findings as a basis for his findings in the ACCC's unconscionable system case. 228 Seen in that way, we would uphold the ground on two bases, now becoming familiar in our reasoning. First, the findings depend on the attributes of the six individual consumers. It may have been unconscionable not to explain to people with their attributes exactly how the scheme worked, the debt students would acquire and their reduced future entitlement to access the VET scheme in aid of having them sign up to obtain the Commonwealth revenue. That is because, at least on his Honour's findings of some of them, they would not have understood these things themselves, and would not have read, and some might not have been capable of reading, the material which they were given. It is not possible to extrapolate that approach out to a material number of Unique's approximately 3,600 students. 229 Secondly, unlike some of the other authorities, there was no evidence that the product the students received was flawed. Unlike AMI, where there was no evidence the treatment worked, in this proceeding no evidence was adduced to suggest the quality of the courses Unique provided was inadequate, or that students would not be properly qualified on completion. The impugning of the quality of the courses was not part of the ACCC's case. Unlike National Exchange, there was no evidence that students were financially short-changed; and there was no evidence Unique's course pricing (and therefore the debt students incurred) was unreasonable or excessive. There was no suggestion the courses were in fact unavailable. Provided the students were equipped (in terms of resources and education level) to access the course, the evidence appeared to be students who did Unique's courses would receive more or less the same quality of product as students doing comparable VET FEE-HELP courses elsewhere. And, as we have now said many times, the ACCC did not attempt to prove how it was that a material number of the 3,600-strong student cohort did not have the resources or education levels to benefit from the courses.

Ground 8: no unconscionable conduct by way of a system or pattern of behaviour 230 It will be apparent, as we said at the outset of these reasons, that we accept Unique's submissions that in all the circumstances, there is an insufficient evidentiary basis for a finding of the kind set out in the first declaration made by the primary judge. The declaration must be set aside. For our own part, having considered the material to which we were referred and the submissions made, in exercising a rehearing function, we do not consider any other or differently framed contravention of s 21 can be found, by way of any system or pattern of behaviour as alleged by the ACCC. There were four features of the ACCC's factual case that were accepted by the primary judge as constituting a system of conduct or pattern of behaviour (see [84], above). We have dealt with some of these features in our findings in relation to grounds 5 and 6 of the appeal. We accept that as a matter of fact there was indeed a pattern of providing to students "free" laptops or iPads, that there was a pattern of using incentives for staff to encourage them to sign up students and that there was a pattern of holding sign up meetings. All of these were not disputed by Unique's own witnesses, and there was sufficient evidence that, to this extent at least, what occurred at the four sites was typical of what occurred at other sites. 231 However, to make a limited finding of that nature takes the ACCC's s 21 case nowhere near far enough. There is nothing inherently unconscionable about the last two aspects of the pattern. As to the laptops, as we have set out above, in order to characterise the provision of laptops as unconscionable, one would need evidence about why they were, in effect, useless to the students and did not enable them to start or finish the course for which they had incurred a debt, but rather were nothing more than lures to enrol. Even for the six individual students it is not clear that could be found for all of them. Beyond the six, there is simply no evidence at all. There can be no alternatively framed relief to replace the declaration we have set aside.

Resolution: the ACCC's cross-appeal 232 The cross-appeal should be dismissed.

Ground 1 233 As to the first ground, it is not tenable. The ACCC relies on its pleadings at [21.1] which gives particulars of locations in New South Wales. It contends this list is not exhaustive. However, in substance, the ACCC pleaded and ran a case about what happened in New South Wales. It called no witnesses about enrolment outside New South Wales. It made no allegations about consumers outside New South Wales. In its submissions on the cross-appeal, the ACCC relied on some exchanges during closing submissions as raising what had happened in New South Wales, together with some cross-examination of Unique's witnesses. That is far from sufficient. Unique's counsel were entitled to see the cross-exanimation as peripheral or unimportant given the way the pleadings were structured. In the absence of a formal application to amend the pleadings and put the case squarely, various exchanges during closing submissions are not sufficient to entitle the ACCC, as it seeks on its cross-appeal, to wider declaratory relief. 234 In any event, as a matter of substance and putting to one side the way the ACCC introduced, late, what happened in Queensland and Victoria, the argument itself has little merit. It is true that in some of the evidence (such as Professor Vinson's report) there are references to correlations between the students enrolled by Unique in its courses and relative socio-economic disadvantage in, for example, Queensland. However, as we have noted above, the statistics in Professor Vinson's evidence take the ACCC's unconscionable system case not very far at all. As we have noted, the higher proportions of students with any one or more of the three characteristics of "indigeneity", low socio-economic grouping or remote or very remote locations could be said to indicate Unique was enrolling students whom the government had intended would be offered access to such courses. 235 It is not only not possible, but impermissible, to draw any inferences, by the making of stereotypical assumptions, about the attributes and characteristics of individual students who live in such areas, let alone students who are Indigenous. The ACCC should not invite the Court to engage in such a stereotyping exercise. Stereotypical assumptions - which often carry unstated prejudices - are no substitute for admissible evidence.

Ground 2 236 As we set out at [98] above, the ACCC identifies five additional features it contends the primary judge should have added to his Honour's finding about what constituted Unique's "system": (1) Unique's false, misleading or deceptive representations to consumers (as his Honour found them); (2) a failure to explain the VET FEE-HELP Assistance Scheme to consumers; (3) a failure to explain the nature and content of the forms they had consumers sign; (4) a failure to ensure that consumers were enrolled in courses that were suitable for them; and/or (5) a failure to comply with the requirements for making unsolicited consumer agreements in Division 2 of Part 3-2 of the ACL. 237 In its notice of cross-appeal, under the heading "Particulars", the ACCC has a list of matters to which it contends the primary judge failed to have regard. It is unclear how these represent particulars of any of the five features, and if so, which. Rather, they appear to be separate aspects or categories of evidence that the ACCC complains the primary judge did not consider, but should have. 238 This ground should be rejected, on two bases. First, each of the "features" set out in the notice of cross-appeal, and which we have set out above, suffers from the same infirmity as that which has occupied most of our reasons: the evidence derives from what happened to six consumers at four sites, out of over 3,600 consumers at more than 428 sites, with no evidence about how or why the consumers and sites are representative, or whether they are random, or indeed any evidence at all about how they were selected. As Unique submitted, in contrast to other s 21 cases, the ACCC did not rely on representations in advertising or marketing material: rather, it relied on what was allegedly said to the consumer. However, it sought to prove only what was said to six of them. These additional "features" focus more closely on what is set out in [22] of the ASOC and described as "failures", but the problem is the same. The way the "failures" are alleged required proof of what happened in individual consumer transactions - there was nothing generic about them. For example: a failure to explain the nature and contents of forms consumers had to sign needed to be proved in respect of the class of consumers as a group - we have already explained some of the ways in which that might have been done. 239 Second, the additional "particulars" do not advance the ACCC's system case either, even if the ACCC were now to be permitted to rely on them when, as Unique pointed out, these matters were not pleaded as aspects of Unique's system, or as indicators that it had a system, at first instance. Insofar as the "Particulars" rely on matters in his Honour's reasons (such as the "scramble" finding - see [2.6], or the finding that Unique's witnesses falsely declared they had provided advice about VET FEE-HELP when they had not), we have dealt with these in our reasoning on Unique's appeal and, once again, the substance of the evidence to which the primary judge refers relates only to the six consumers and the four sites. Where his Honour's reasons go beyond this (such as the finding at [691] that generally the "spiel" at the start of the meeting by Ms Kang only lasted five minutes and did not include an explanation of the VET FEE-HELP Assistance Scheme or consumers' rights to withdraw from a course before the census date), they do so without evidence about what the effect of this spiel was on a material number of consumers, and what those consumers did or did not understand from other sources. Again it does not advance the ACCC's case very far at all. 240 The ACCC raises three new aspects of the evidence on the cross-appeal, which it invites the Court to consider. It submits the primary judge erred in not finding, and this Court should find, that what took place at Walgett, Taree, Tolland and Bourke was "characteristic of the Appellant's marketing and enrolment process" because: (A) the Appellant received complaints from consumers that the Appellant's representatives had engaged in the type of unlawful conduct that took place at Walgett, Taree, Tolland and Bourke in other places; (B) the Appellant did not provide its representatives with any training as to how to market and enrol consumers in the Appellant's courses in a manner that complied with the requirements of the ACL; (C) the log-on and completion rates of the consumers enrolled in the Appellant's courses were low. 241 In its cross-appeal submissions (at [48]-[50]), the ACCC added a contention about the primary judge's failure to make findings about the scale of referral fees or commissions which Unique paid, and Unique's practice of paying referral fees and commissions in the course of considering the elements of Unique's marketing and enrolment process. This contention does not appear in the cross-appeal and no leave was sought to add it. We accept Unique's submissions that the evidence showed the referral fee system had been in operation since Unique commenced operations in 2008, which tends against the referral practice being part of a "system" post 2012. Moreover, the evidentiary gaps which we have focused on in the ACCC's system case mean that it would not have been possible for the primary judge to infer that the existence of the referral practice formed part of an unconscionable system, because there was no evidence about which students were signed up as a result of such referrals, what their individual characteristics were and how they were vulnerable. 242 As to the complaints, there is nothing particularly probative about this collection of complaints, which number 19 in total, in terms of what was lodged during the contravention period. There is no evidence of correlation between those students said to be vulnerable (because of their location, their race or the like) and the authority of the complaints. The complaints were untested. Their number is very small, out of a cohort of over 3,600 students. Indeed, the Court might just as easily speculate that a small number indicates most students were content with the course. We decline to speculate at all about this evidence, which is what the ACCC effectively invited us to do. 243 The contention about the absence of specific training on the ACL is tangential, to say the least. How the absence of training was said to affect the way Unique's officers conducted themselves (especially given the ACCC's primary case was one based on a conscious targeting and unconscionability) is not explained. 244 Finally, the ACCC relied on the evidence from Ms Martin about how she received no responses to her emails, and also relied on some overall statistics about completion rates of the Unique student cohort. It will be recalled, and the primary judge found (at [218]-[221]) that Ms Martin sent an email to the 80 Unique students she had been assigned, and received no responses at all from her class, getting "bounce back" emails from 40 addresses, and noting some of the email addresses she had been given for different students were the same. 245 That evidence is not inconsistent with the overall statistics about completion rates for Unique's courses. The evidence was that out of 3,631 students only 493 completed a course. Only 717 students logged on, even once, to their course. 246 These statistics are sobering. In its cross-appeal submissions, the ACCC submitted this evidence suggested that Unique was at the least: … indifferent to whether the students it enrolled were suitable to undertake, or were likely to be capable of completing, online studies for a diploma-level qualification, whether by reference to the students' aptitude, education level, computer literacy, or their access to the internet. 247 Again, senior counsel for the ACCC did not take the Court to any cross-examination of Unique's witnesses to this effect. As Unique submitted, although the numbers are low, they should be seen in the context of the overall performance of the VET FEE-HELP program. When that is done, the absence of a direct challenge to Unique's witnesses on the basis of these numbers might be explicable. 248 Unique's completion rate was 13.6%. That may seem remarkably low, but when considered relatively quite a different picture emerges. According to the Commonwealth's statistics in its Discussion Paper, the completion rate for external online courses across the VET sector was 7%. The Discussion Paper accepts at many points that completion rates were low, despite the "extraordinary growth" in the scheme: more than 50-fold since 2009. The Discussion Paper includes a table of five anonymised providers, described as "major" providers. The course completion rates for 2014 range between the five providers from a low of 1.3% to a high of 7.8%. For 2013, the lowest rate was 2.6% and the highest 16.3%. 249 These statistics are appalling, in particular when one considers the outflow of Commonwealth revenue and the corresponding increase in lifetime personal debt to so many people. It is little wonder the Commonwealth sought to revise the scheme. However, placing the completion rates for Unique in context illustrates how important it was for the ACCC to go well beyond abstract numbers, into some kind of qualitative evidentiary exercise that was not only representative to some degree across Unique's student cohort but also placed within the specific context of the VET FEE-HELP scheme.

SUMMARY AND Conclusion 250 By way of summary, our conclusions can be seen clearly by reference to the pleadings. 251 The system said to be unconscionable was pleaded in paras 21 to 26 of the ASOC. Paragraph 21 pleaded aspects that were individually and collectively neutral; para 22 pleaded the systematic failures; and para 23 pleaded the consumers who were systematically enrolled - those who did not understand certain things. 252 Declaration 1 made by the primary judge in its terms did not refer to any failures, other than by using the word "unconscionable". The basis for the declaration was the finding of the four features in [774] and the conclusion of unconscionability in [778]. 253 But, with respect, the conclusion on targeting was flawed. The conclusion (even if not flawed) was also an inadequate foundation for the conclusion as to unconscionability, unless the failures in para 22 and the consequences in para 23 were proved. These were not dealt with by the primary judge. 254 The gaps in the evidence in proving these things included the following: There is no evidence that the individual cases were representative of the failures pleaded, accepting the evidence does show the "process" or the "system" pleaded in para 21 as accepted by the primary judge at [774] and [777]. The emails of complaints are of little weight and may (contrary to the submissions of the ACCC) help show over the relevant period (1 July 2014 to 30 September 2015) very little dissatisfaction overall. There was no documentary evidence of "failures" as a system. The evidence of the expert did not deal with the causal relationship between the locations and type of enrolments and the conduct (or failures) alleged against Unique. There was no evidence that the targeted group were disadvantaged unfairly as a group, as distinct from the various individuals before the Court who had varied vulnerabilities (e.g. age, literacy etc.). There was no evidence showing the giving of laptops was "unfair" or that the contract entered into by the students was a "bad deal". All were premised on the general vulnerability of the class of consumers, which was not established. There was no evidence, nor was it suggested, that the courses offered were either not properly described or of bad quality in themselves. There was no qualitative expert evidence as to the system, the completion rates, the structure of Unique's business or other matters which could form a reliable or rational basis to conclude that Unique's system was unconscionable. 255 It would appear the ACCC's allegation of unconscionable conduct tracked closely the observations in the Commonwealth's Discussion Paper at p 14: While data indicates significant growth in VET FEE-HELP take up across all student cohorts, the growth in disadvantaged students is markedly higher compared to non-disadvantaged students. In itself, this is one of the objectives of the scheme. However, as outlined below, it has been accompanied by poor outcomes in some aspects. One explanation for this is the proliferation of unethical actions by a small number of providers offering inducements such as iPads, cash and vouchers to prospective students to enrol in a course and request VET FEE-HELP. These behaviours specifically targeted vulnerable people through cold calling or door knocking neighbourhoods of low socio-economic status. Those targeted are signed up to a course which they may not have the academic capability to complete and may not understand the loan must be repaid. 256 These were observations in a policy document. To prove a case of unconscionability of a system in a court of law against a particular provider, against whom serious forms of relief were sought, required more than the evidence adduced by the ACCC in this case. 257 Noting that the question of costs was reserved by the primary judge on 8 November 2017, the orders that we would make are as follows:

1. The appeal be allowed.
2. Declaration 1 of the orders of Perram J made on 8 November 2017 be set aside.
3. The first and second respondents pay the appellant's costs of the appeal.
4. The cross-appeal be dismissed with costs.
5. The matter be remitted to the primary judge for determination of the costs of the trial and for hearing as to further remedy. I certify that the preceding two hundred and fifty-seven (257) numbered paragraphs are a true copy of the Reasons for Judgment herein of the Honourable Chief Justice Allsop, Justices Middleton and Mortimer.