COURT FILE NO.: CV-20-00647858-00CP

DATE: 20220715

SUPERIOR COURT OF JUSTICE - ONTARIO

RE: JARED LE FEUVRE, Plaintiff

– and –

ENTERPRISE RENT-A-CAR CANADA COMPANY, Defendant

BEFORE: E.M. Morgan, J.

COUNSEL: Jonathan Ptak, Jamie Shilton, Jean-Alexandre De Bousquet, and Daria Chyc, for the Plaintiff

Christopher Naudie, Geoffrey Hunnisett, and Elie Farkas, for the Defendant

HEARD: June 14-16, 2022

CERTIFICATION motion

I. Context of the case

[1] The Plaintiff claims that the Defendant, a Canadian subsidiary of a U.S.-based car rental company, has systemically misclassified its branch managers since 2010. He seeks to represent a class in the range of 2,500 individuals who have from that time until the present been a Branch Rental Manager (“BRM”), Assistant Branch Rental Manager (“ABRM”), or Station Manager (“SM”) employed by the Defendant.

[2] The Defendant operates under the brands of Enterprise Rent-a-Car Canada, National Car Rental, and Alamo Rent-A-Car in the provinces of Ontario, British Colombia, Alberta, Saskatchewan, Manitoba, Prince Edward Island, and Nova Scotia. There are approximately 500 such rental car branches nationwide.

[3] The Plaintiff’s position is, in essence, that the BRMs, ABRMs, and SMs have duties and operational tasks that are indistinguishable from the Defendant’s non-managerial employees. The Fresh As Amended Statement of Claim alleges that every employee of the Defendant’s that has been labelled a “manager” at its many branches has not exercised job duties that were “managerial in character”. Accordingly, the Plaintiff pleads that each of those class members is entitled to overtime pay – i.e. something which hourly employees receive and managerial employees do not receive – in addition to the salary and profit-based compensation that they have received as BRMs, ABRMs, or SMs.

[4] The evidence in the record is divided on the fundamental issue of whether the proposed class members perform mostly managerial or non-managerial functions. While both sides concede that a store or branch manager might be expected to divide her time between daily operations or customer service on one hand and supervisory or operations management tasks on the other, there is a stark division among the class member affiants themselves as to the content of their workday. This is partly due to the fact that the affiants reflect a mix of urban centre vs small town locations, airport vs non-airport branches, downtown vs suburban storefronts, and unionized vs non-unionized workplaces.

[5] The Plaintiff’s manager-affiants are for the most part drawn from outside of the largest corporate operating groups. By way of illustration, a witness who is an ABRM in Regina indicates that he spent an average of 70 work hours each week on non-managerial functions. Likewise, the Plaintiff himself deposed that his experience as a manager at a northern Ontario branch was that he supervised 2 employees and managed about 270 vehicles. Perhaps not surprisingly, this group overall deposes that 65%-90% of their time is spent on non-managerial functions such as serving customers at the counter and even washing and parking cars.

[6] The Defendant’s manager-affiants, by contrast, are for the most part drawn from larger centres. By way of illustration, a witness who is a BRM in downtown Toronto estimates that she spends 70% of her time on managerial tasks like financial and fleet management. Likewise, a witness who works as an SM at the Pearson Airport branch deposes that she supervises 23 employees and manages 2,700 vehicles. It is again unsurprising that this group overall deposes that they spend only 10%-30% on non-managerial tasks.

[7] It is against this somewhat complicated factual background that the class action certification criteria must be applied.

II. The certification test

[8] It is by now well settled that as a legal hurdle, certification presents a relatively low bar. The Plaintiff must establish that there is “some basis in fact” on which to base the claim. This threshold is set to reflect the fact that “[t]he certification stage does not involve an assessment of the merits of the claim and is not intended to be a pronouncement on the viability or strength of the action; rather, it focuses on the form of the action in order to determine whether the action can appropriately go forward as a class proceeding”: Pro‑Sys Consultants Ltd. v. Microsoft Corporation, 2013 SCC 57, [2013] 3 SCR 477, at para 102.

[9] Before delving into the stage-by-stage certification test itself, it is worth noting that the claim was issued on September 18, 2020. As set out in section 39(1), Schedule B of the Class Proceedings Act, 1992, SO 1992, c. 6 (“CPA”), the recently enacted amendments to the CPA which otherwise impact the certification test do not apply to his proceeding.

a) Cause of action – section 5(1)(a)

[10] The Plaintiff claims that the Defendant breached its statutory and/or contractual duties, that it acted negligently, and that it was unjustly enriched. Under s. 5(1)(a) of the CPA, the court must presume that the facts pleaded are true; moreover, the claim will pass the cause of action threshold unless it is plain and obvious that it will fail: Pro-Sys, at para 63.

[11] The heart of the claim is that the Defendant breached its obligation to pay the Plaintiff and putative class members overtime for hours worked above the applicable statutory threshold. In Ontario, the relevant threshold is set out in the Employment Standards Act, 2000, SO 2000, c 41 (“ESA”); each Canadian jurisdiction in which the Defendant operates has its own statutory standard, the specifics of which vary although the principle is the same. The claim relies on each of the relevant employment standards statutes across the country and seeks damages for those alleged statutory breaches. This is possible in a national class action, provided that none of those other jurisdictions specifically bar adjudication by an Ontario court.

[12] Defendant submits that, in fact, British Columbia jurisprudence does just that. In Macareag v. E Care Contact Centres, 2008 BCCA 182, the British Columbia Court of Appeal held that the exclusive jurisdiction for determining issues of overtime pay lies with the Director under B.C.’s Employment Standards Act, and that statutory minimums are not enforceable as implied terms of an employment contract as any judicial enforcement is supplanted by the regulatory mechanism.

[13] Accordingly, the statutory causes of action may be valid for Ontario, but it is plain and obvious that they are not proper causes of action in respect of B.C. class members: Hollick v. Toronto (City), 2001 SCC 68, [2001] 3 SCR 158, at para 25. The certification process cannot be used as a means of granting rights to those individuals that they would not otherwise have: Pioneer Corp. v. Godfrey, 2009 SCC 42, at para 116.

[14] As Plaintiff’s counsel points out, the claim for overtime pay obligation originates in provincial employment legislation, but (other than in British Columbia) the claim to enforce it can be understood as a claim for breach of contract, breach of statute, or both. Most often, courts have held that statutory minimum standards are implied terms in employment contracts. As contractual terms, they can be enforced in actions for breach of contract, and that class actions can be certified on that basis: Stewart v. Park Manor Motors Ltd., 1967 CanLII 243 (ON CA), [1968] 1 OR 234 (Ont CA); Kolodziejski v. Auto Electric Service Ltd., 1999 CanLII 12264 (SK CA), [1999] SJ No 276 (Sask CA); Hutlet v. 4093887 Canada Ltd., 2015 MBCA 82; Walter v. Western Hockey League, 2017 ABQB 382, aff'd 2018 ABCA 188; Dominguez v. Northland Properties Corporation, 2012 BCSC 328.

[15] The Fresh As Amended Statement of Claim also pleads the elements of liability in negligence. The Plaintiff states that the Defendant owed each of its “managers” a duty of care arising from their close relationship, and that the Defendant breached this duty by misclassifying them as managers and failing to pay overtime. The pleading claims that each of the putative class members suffered damages in the form of unpaid overtime as a result of this breach. Similar claims, albeit in somewhat different factual contexts, have been certified on the basis that such negligence claims constitute a viable cause of action: See Berg v. Canadian Hockey League, 2019 ONSC 2106 (Div Ct); Fulawka v. Bank of Nova Scotia, 2011 ONSC 530 (Div Ct).

[16] The pleading passes the section 5(1)(a) hurdle, with the exception of British Columbia. There is both a contract and a tort basis for the claim. In addition, there are statutory breaches alleged in provinces across the country which, outside of B.C., appear to present causes of action which are at least potentially viable.

b) Identifiable class – section 5(1)(b)

[17] In Cloud v. Canada (Attorney General), 2004 CanLII 45444 (ON CA), [2004] OJ No 4924, at para 45, the Court of Appeal explained that under section 5(1)(b) of the CPA, the proposed class must be defined by “objective criteria which can be used to determine whether a person is a member without reference to the merits of the action.” With that in mind, the Plaintiff proposes a class definition defined objectively by the Defendant’s own classification.

[18] The proposed class definition is framed to allow a class member to self-identify as a member based on whether they were employed as a BRM, ABRM, or SM. This can be done without regard to whether it was legally correct for any of those positions to be classified as overtime-exempt.

[19] The Plaintiff’s proposed class definition is as follows:

All persons who held or hold the position of Branch Rental Manager, Assistant Branch Rental Manager, Station Manager, or an equivalent position to any of the foregoing at Enterprise Rent-A-Car, National Car Rental, and Alamo Rent-A-Car branches operated by the Enterprise Rent-A-Car Canada Company in British Columbia, Alberta, Saskatchewan, Manitoba, Ontario, Nova Scotia, and Prince Edward Island between January 1, 2010 and the date of the order of the Court certifying the action as a class proceeding.

[20] Counsel for the Defendant argues that the class definition stretches too far back in time, and that as a consequence it violates the applicable limitation period. Counsel for the Plaintiff responds that it is not for the certification court to decide whether any class members’ claims may be barred by limitations statutes: See Navaratnarajah v. FSB Group Ltd., 2021 ONSC 5418, at para 12.

[21] As the Court of Appeal observed in Fresco v. Canadian Imperial Bank of Commerce, 2022 ONCA 115, at para 101, the question of whether a class member can rebut the presumption of discoverability “is really a matter best left to individual assessment”. That individualized analysis can take place after certification of the class and as part of the ultimate exercise in determining who meets or does not meet the class definition. In the meantime, and for the purposes of the certification motion itself, the Court of Appeal has concluded that in the workplace context the employees’ reliance on their employer’s representations as to the legality of its overtime polices can be a basis to suspend applicable limitation periods: Fresco, at paras 99-101.

[22] Accordingly, limitation periods, which turn on the issue of discoverability, are best left for another post-certification day. I agree with the Plaintiff that there is some basis in fact for the contention that the employer’s representations of the legality of the management labels here might impact on the applicable limitation period. In all, I am confident that the Plaintiff has presented facts on which to base a viable class that passes the section 5(1)(b) hurdle.

c) Common issues – section 5(1)(c)

[23] As the Supreme Court of Canada has pointed out, certification, including approval of the common issues, “does not require a preliminary merits showing”: Hollick, at para 16. Rather, “the factual evidence required at this stage goes only to establishing whether these questions are common to all the class members.”: Pro-Sys, at para 101. What is required is that the Plaintiff provide some basis in fact on which the court can conclude that common issues arise in each class member’s claim.

[24] To be clear, not only must a certifiable common issue be a substantial ingredient of each class member’s claim, but it must be such as to avoid duplication of fact-finding or legal analysis and, in addition, must not depend on individual findings of fact that have to be made with respect to each claimant: Fairview Donut Inc. v. The TDL Group Corp., 2012 ONSC 1252, at para 229. Furthermore, a proposed common issue should not be certified where it is stated in such general terms as to be unhelpful in advancing the analysis of the claim: Rumley v. British Columbia, 2001 SCC 69, [2001] 3 SCR 184, at para 29.

[25] The proposed common issues are:

Liability

1. Do the requirements in the Applicable Provincial Employment Standards Legislation relating to the payment of Overtime Pay form express or implied terms of the contracts of employment with the Class Members?

2. Did the Defendant breach its contracts of employment with the Class Members by denying eligibility for overtime compensation to the Class Members who held the position of:

a. Branch Rental Manager or an equivalent position?

b. Assistant Branch Manager, Station Manager, or an equivalent position?

3. Are any applicable policies or contractual terms, which purport to exclude the Class Members from eligibility for Overtime Pay, void and unenforceable?

4. Did the Defendant owe Class Members statutory, contractual, or common law duties, and/or duties of good faith, to:

a. properly advise Class Members of any entitlement to Overtime Pay for hours worked in excess of the applicable provincial thresholds per week?

b. ensure that the Class Members’ hours of work were monitored and accurately recorded?

c. ensure that Class Members were appropriately compensated for their overtime work during the class period?

5. If so, did the Defendant breach any of these duties? If so, how?

Aggregate Damages

6. If Common Issues 1-5 are answered in the affirmative, can the amount of damages owing to the Class Members be assessed on an aggregate basis, in whole or in part?

7. If the answer to Common Issue 6 is answered in the affirmative, what is the quantum of aggregate damages owed to the Class Members?

8. What is the appropriate method or procedure for distributing any aggregate damages award to the Class Members?

Unjust Enrichment

9. Was the Defendant unjustly enriched by failing to compensate Class Members with Overtime Pay for hours worked in excess of the applicable overtime thresholds?

Punitive and Exemplary Damages

10. Are the Class Members entitled to an award of punitive and/or exemplary damages based on the Defendant’s conduct?

11. If Common Issue 10 is answered in the affirmative, what is the quantum of punitive and/or exemplary damages owed to the Class Members?

12. What is the appropriate method or procedure for distributing any punitive and/or exemplary damages award to the Class Members?

[26] In Plaintiff’s counsel’s view, the common issue requirement here is met simply and elegantly. All the Court must do is to answer the misclassification allegation against one of the BRMs, ABRMs, or SMs, and it will have answered the question for them all. Since the class members share the same three titles, and since the claim is itself based on the title given to the class members, Plaintiff’s counsel submit that the common issues criterion is easily met.

[27] In objecting to this series of proposed common issues, counsel for the Defendant raises a number of arguments: structural deficiencies in the questions themselves, difficulty in answering or analyzing the issues on a common basis, lack of similarity between all BRMs, ABRMs, and SMs, and material differences in provincial statutes and regulatory schemes. The Defendant also takes issue with the feasibility of assessing aggregate damages on a class-wide basis, and to this end challenges the methodology proposed by the Plaintiff’s experts.

[28] Each of these arguments will be addressed in turn.

[29] The central issue raised by section 5(1)(c) – i.e. the commonality requirement – is placed in starkest relief by proposed common issues 2, 4, and 5. By contrast, questions 1 and 3 present issues that are, indeed, common, but require no real answer. Those questions ask about the application of statutory employment standards to the issue of management overtime pay posed in this case, and the enforceability of those standards. Those issues are raised equally by each class member’s claim; at the same time, the Defendant concedes that it is obligated to comply with all applicable employment standards legislation.

[30] This Court has said in similar cases that it is “self-evident…that the ESA cannot be contracted out of, and it is part of every contract of employment that its terms and those of the Regulations thereunder must be complied with”: Azar v. Strada Crush Limited, 2018 ONSC 4763, at para 21. Accordingly, to ask in questions 1 and 3 whether statutory employment standards apply to the Plaintiff and are enforceable against the Defendant does not, as the Supreme Court of Canada requires, “significantly advance the action”: Hollick, at para 32. The questions engage no analysis or controversy.

[31] Questions 2, 4, and 5, on the other hand, pose the crucial question of whether by denying overtime pay in management contracts, the Defendant in fact breached its contractual obligations – including the contractual obligation of good faith – to the BRMs, ABRMs, and SMs. As indicated, counsel for the Plaintiff submits that this question can be answered in the identical way for every one of the proposed class members. Counsel for the Defendant, on the other hand, submits that this question, important as it is to the merits of the claim, must be answered in respect of each class member on an individual basis.

[32] The problem with the Plaintiff’s view of commonality is that what is common is the potential for a claim, but not the factual basis of an actual claim. While the possibility or risk that the Defendant has breached its obligation is common to the entire class, the “risk that you may not be paid overtime because of an employer’s impugned practice is not a breach of the employment agreement”: Baroch v Canada Cartage Diversified GP Inc, 2015 ONSC 40, at para 36. The breach occurs if, and only if, the claimant in fact worked overtime and in fact was denied pay for that work.

[33] That question turns on specific fact finding that can only be answered claimant by claimant, or putative class member by putative class member. As Strathy J. (as he then was) observed in Singer v. Schering-Plough Canada Inc., 2010 ONSC 42, at para 140(h): “A common issue cannot be dependent upon individual findings of fact that have to be made with respect to each individual claimant.”

[34] Thus, “[w]hile the theory of liability can be phrased commonly, the actual determination of liability for each class member can only be made upon an examination of the unique circumstances with respect to each class member’s [employment]”: Williams v. Mutual Life Assurance Co. of Canada (2000), 2000 CanLII 22704 (ON SC), 51 OR (3d) 54, at para 39 (SCJ), aff’d 2001 CanLII 62770 (ON SCDC), [2001] OJ No 4952, 17 (Div Ct), aff’d 2003 CanLII 48334 (ON CA), [2003] OJ No 1160 and 1161 (Ont CA). Here, every BRM, ABRM, and SM might be misclassified, but a specific examination of their particular job functions and routine workday would form the basis for that determination. Merely carrying the label of BRM, ABRM, or SM does not provide any basis in fact for the analysis that must ensue.

[35] Plaintiff’s counsel address this problem by stressing that proposed common issues 2 and 4 focus attention on the question of misclassification. Counsel for the Plaintiff submit that the issue is a class-wide one since the managerial classifications in issue do not vary class member to class member. It is the Plaintiff’s view that the misclassification claim conforms with the Court of Appeal’s description in Brown v. Canadian Imperial Bank of Commerce, 2014 ONCA 677, at para 40, that “[e]ligibility for overtime pay could only be certified as a common issue under s. 5(1)(c) of the CPA if the evidence on the motion demonstrated that the alleged misclassification of the employees as managers/supervisors could be resolved commonly for all employees in the proposed class.”

[36] The Court of Appeal has made it clear that commonality requires a factual underpinning to be present in the certification record. Thus, common issues will be found to exist where the evidence on the certification motion establishes “some basis in fact to find that the functions, responsibilities and duties of all of the employees in the putative class are sufficiently similar that the classification of those employees as eligible or ineligible for overtime pay could be made for the class as a whole and without regard to the specific circumstances of individual employees”: Brown, at para 41.

[37] Again, this analytic standard requires a review of the work actually performed by each potential class member: Ibid., at para 41, citing McCracken v. Canadian National Railway Co., 2012 ONCA 445, at para 91. Substantial variability in the “autonomy, duties and responsibilities of employees having the same job title or classification” will lead to the conclusion that there is no “core commonality” in the functions and duties under examination: Brown, at para 44.

[38] The cases on which Plaintiff’s counsel most heavily rely – Fulawka v. Bank of Nova Scotia (2012), 2012 ONCA 443, 111 OR (3d) 346, at paras. 78-84 (Ont CA), leave to appeal ref’d, [2012] SCCA No 326, and Fresco v. Canadian Imperial Bank of Commerce (2012), 2012 ONCA 444, 111 OR (3d) 501, leave to appeal ref’d, [2012] SCCA No 379 – were what the Defendant calls ‘off the clock’ cases, not misclassification cases. In those cases the employer conceded that the employee class was entitled to overtime pay, and the issue was the employer’s imposition of system-wide barriers to collecting that pay. Likewise, the employment misclassification case on which Plaintiff’s counsel primarily rely – Rosen v. BMO Nesbitt Burns Inc., 2013 ONSC 2144 – is a case where the question was different enough from the present case that it was susceptible to a class wide analysis. There, the issue was one of tradeoff – i.e. whether the opportunity for higher income was a fair and equitable tradeoff for the plaintiff and class of employees losing the right they otherwise would have to overtime pay.

[39] In the case at bar, by contrast, the controversy is over the variability of the actual job functions and duties performed by the class. It is this factual variability that makes up much of the affidavit evidence in the record, and that makes the issue of classification far more problematic in terms of being addressed on a class-wide basis. It is fundamental to the Plaintiff’s argument that the “essence of the job” of potential class members is to process customer rentals and to perform basic branch operational duties. However, given the significant differences outlined in the introductory portion of these reasons for judgment, it is not clear in the record that there is a unified “essence” to the jobs in issue.

[40] In their written submissions, Defendant’s counsel juxtapose two Manitoba members of the putative class as a further illustration of the problem. Tiffany Johnson-Mailey, a former ABRM at several locations in that province, deposes that, on average, 90% of her typical workday during her time with the Defendant was spent performing non-managerial tasks. Kristina Corbett, a BRM at various downtown Winnipeg locations and the Winnipeg airport, deposes that non-managerial tasks do not constitute a regular part of her work, and that 80% of her time is spent on tasks that undoubtedly qualify as management duties.

[41] Those vastly differing experiences are to be expected, according to the Defendant, since the scale of business they were responsible for managing differ markedly. Ms. Corbett supervises up to 39 employees, while Ms. Johnson-Mailey only supervised between 3 and 6 hourly employees during the years of her employment and, in fact, she states that she spent much of her time working alone. The situation of a manager at a bustling urban centre simply cannot be answered in common, based on job title alone, with that of a manager at a quiet prairie location.

[42] Citing Navaratnarajah v. FSB Group Ltd., 2021 ONSC 5418, Plaintiff’s counsel submit that detailed differences in work arrangements are manageable as common issues, so long as the essence of the job in issue is capable of being addressed in common. While this is an accurate statement of the law, it must always be applied in context.

[43] Thus, in Navaratnarajah itself, the Court was careful to point out, at para 18, that any such detailed differences “appear to generally exist at the margins of the working relationship and not at its core.” In the case at bar, the evidence is that differences between the potential class members reflect the very factors that must be considered in assessing whether or not they are managers – e.g. the percentage of time spent on non-managerial tasks. It therefore cannot be said that the differences between potential class members’ work “exist at the margins of the working relationship”, as those differences go to the core of the Plaintiff’s claim.

[44] The evidence in the record before me demonstrates that the variability and mix of the work actually performed by BRMs, ABRMs and SMs across hundreds of branches makes the alleged misclassification impossible to resolve on a basis that is common to all employees in the proposed class. A class that includes personnel in small branches where the management is expected to be a jack of all job functions, and personnel in high-volume airport branches where the division of job functions between management and non-management is clear and hierarchical, is not a class that raises issues in the way envisioned by the CPA. Some of the putative class members may well be misclassified while others are not; but the factual basis for that determination is so varied that each member would have to bring their own claim and pursue their own findings of fact.

[45] For this reason, common issues 2, 4, and 5 do not meet the test for certification.

[46] As Defendant’s counsel point out, the misclassification issue is a threshold issue in this claim. If it cannot be certified, the balance of the issues by sheer logic must fall away. The other issues – i.e. whether the Defendant has been unjustly enriched, whether damages can be calculated on an aggregate basis, and whether the case is an appropriate one for punitive damages – are all derivative of the basic misclassification claim. Without being able to determine the threshold issue of liability, it stands to reason that a common issues trial court will not be able to determine the secondary issues that flow from the initial threshold.

[47] Having said that, I will pause to briefly address the aggregate damages issue. The parties have spent considerable effort on that question, and it merits some consideration in its own right.

[48] Section 24(1) of the CPA provides:

The court may determine the aggregate or a part of a defendant’s liability to class members and give judgment accordingly where,

(a) monetary relief is claimed on behalf of some or all class members;

(b) no questions of fact or law other than those relating to the assessment of monetary relief remain to be determined in order to establish the amount of the defendant’s monetary liability; and

(c) the aggregate or a part of the defendant’s liability to some or all class members can reasonably be determined without proof by individual class members.

[49] The key question for present purposes, and the one most vigorously debated by counsel for both sides, is the requirement of section 24(1)(c) that aggregate damages be determined “without proof by individual class members”. When it comes to certification of an aggregate damages question, the issue is one of methodology: Heller v. Uber Technologies Inc., 2021 ONSC 5518, at paras 197-8.

[50] The onus is on the Plaintiff to demonstrate that it has a methodology for assessing damages that has a “reasonable likelihood” of satisfying the section 24(1)(c) requirement, on the assumption that the Plaintiff would be otherwise successful at the common issues trial: Shah v LG Chem Ltd., 2018 ONCA 819, at para 104. In other words, the Plaintiff must present a methodology that offers a realistic prospect of establishing aggregate damages on a class-wide basis: Pro-Sys, at para 118.

[51] The seriousness with which this requirement has been taken is illustrated by the reasons for judgment in Uber, at paras 197-9. There the Court found that aggregate damages cannot be established where “individual questions of fact relating to the determination of each Class Member’s damages remain to be determined”. Although the Court observed that the defendant maintained detailed records of the activities of class members, it reasoned that the record lacked a means of obtaining the data needed to determine what individual class members were owed: Ibid. Justice Belobaba made the point as succinctly as possible in Omarali v. Just Energy, 2016 ONSC 4094, at para 94, where he explained that an aggregate damages question will not be certified where “losses cannot be determined without proof by individual class members.”

[52] In addressing the data collection and proof issue, the Plaintiff has presented two experts with two different proposed methodologies. The first is statistician and economist Stefan Boedeker. He proposes taking the information produced by employees’ computer log-ins and extrapolating from that the number of hours – specifically, the number of overtime hours – the Defendant’s employees work on average.

[53] Mr. Boedeker explains in his expert report that the log-in data is time stamped and is archived and accessible in the computers’ storage. He opines that, “log-on and log-off times and other time-stamped data are reliable approximations for the time worked of non-management employees who perform ‘administrative, clerical, technical, and supervisory job classifications’ because the nature of their work requires them to access a computer system to start their work.” Further, to the extent that there are gaps in this data, Mr. Boedeker proposes using statistical methods to fill in those gaps.

[54] The Plaintiff’s second expert is sociologist Graham Lowe. He proposes using crowdsourced data produced on the website Glassdoor.com as a window (or doorway) into the Defendant’s employment practices. More specifically, he proposes using reviews posted anonymously by persons identifying on this site as employees of the Defendant as a method of ascertaining the amount of overtime worked on average by the Defendant’s employees.

[55] Dr. Lowe explains in his expert report that this form of crowdsourced data is often very useful in various contexts. As he puts it, the information that it provides about an employer can “enable insights regarding company practices, pay and working conditions, and regional differences in the case of national or multinational employers. This information is used by employees when researching potential employers, and by employers to understand their employees’ perceptions and experiences.”

[56] With the greatest of respect, neither of these experts has presented a methodology that is acceptable as legal proof.

[57] Mr. Boedeker’s approach is one of statistical sampling, in which log-in data is used as a rough proxy for time spent at work. There is no guarantee, of course, that any given employee logged into a computer upon arrival at work in the morning or logged out upon leaving work at night; but the thinking is that if the sample is large enough there will be sufficient overlap between the data and the precise information sought that one can extrapolate the necessary information from the admittedly incomplete or faulty information at hand. As Mr. Boedeker explains it, the use of this data would also have to be supplemented by individualized information obtained at the branch level in order to properly understand the policies and variables that might impact hours worked at a given location at a given time.

[58] Dr. Lowe’s approach is also one of statistical sampling, but with an even more unreliable data set. In the first place, he has identified on Glassdoor.com only 104 unique reviews of the Defendant out of a class of 2,500 members. This small sample presents a problem in itself, as it may be unreliable even if one generally accepts the proposed statistical method. More important, however, is the fact that the Glassdoor.com postings are anonymous and are an inadmissible form of hearsay. They suffer from an obvious selection bias – complainers post on the internet more than people happy with the target of the post – and one cannot know whether the posts are authentic, whether multiple messages have been posted by a single reviewer, or whether the sentiments expressed in them are genuine.

[59] Without putting too fine a point on the argument, the fact that the data is found on a website is simply not enough to make it something that amounts to proof in the aggregate. If this kind of information and sourcing – effectively, “I read it on the internet” – can pose as an evidentiary record on which to base an expert opinion, then there is “evidence” available to support every conceivable opinion on every conceivable topic.

[60] Indeed, it takes only the most cursory familiarity with internet rants to know that they are used as “evidence” for even the most outlandish conspiracies and other theories. I say this not to compare Dr. Lowe’s approach to those kinds of extreme views, but rather to illustrate the frailty of internet sources and anonymous postings as evidence. As the Federal Court noted nearly two decades ago in ITV Technologies Inc. v. WIC Television Ltd., 2003 FC 1056, at para 18:

[T]he reliability of the information obtained from an unofficial web site will depend on various factors which include careful assessment of its sources, independent corroboration, consideration as to whether it might have been modified from what was originally available and assessment of the objectivity of the person placing the information on-line. When these factors cannot be ascertained, little or no weight should be given to the information obtained from an unofficial web site.

[61] In making these observations, I do understand and, in a sense, admire the efforts to which Dr. Lowe has gone to find a substitute source for individualized interviews with BRMs, ABRMs, and SMs. But it is important in the age of the internet that legal process limit itself to cogent and reliable standards of evidence. Students at University of Toronto are instructed not to use similarly anonymous internet sources as a basis for research because the information posted there “may be inadequate or incorrect”: Why can’t I use Wikipedia for my assignments?, University of Toronto Libraries, https://onesearch.library.utoronto.ca/faq/can-i-use-wikipedia-my-assignments. If such data is too unreliable for an undergraduate essay, it is certainly too unreliable for juridical purposes.

[62] In any case, the Court of Appeal in Fulawka – a case on which the Plaintiff heavily relies for his liability argument – has unreservedly rejected the type of evidence proposed here regardless of the reliability of the underlying data. In Fulawka, the plaintiff’s expert proposed extrapolating overtime hours from a random sampling of class members who would give out-of-court statements collected by the expert. The expert opined that the random sample of statements about overtime work could be reliably projected to the class as a whole.

[63] The Court was unwilling to accept that approach. Indeed, Winkler CJO, for a unanimous panel, took some umbrage at the notion that statistical sampling of this sort could substitute for proof. He stated that, “The plaintiff’s proposed procedure for arriving at a global damages figure is antithetical to the requirement in s. 24(1)(c) that the aggregate amount of the defendant’s liability ‘can reasonably be determined without proof by individual class members’”: Ibid., at para 137.

[64] Accordingly, proposed common issues questions 6,7, and 8 dealing with aggregate damages cannot be certified. The record contains no reliable methodology by which the aggregate damages assessment can be made.

d) Preferable procedure – section 5(1)(d)

[65] As already indicated, since the Plaintiff’s proposed common issues fail to meet the certification test the balance of the certification analysis becomes superfluous.

[66] In Brown, at para 199, Justice Strathy stated with respect to section 5(1)(d) that, “The insurmountable impediment in this case, and the reason why the preferable procedure requirement has not been met, is that the issue of CIBC’s liability to pay overtime to every class member is an individual issue.” The same can be said of the case at bar. The claim would not be fairly, efficiently and manageably advanced by a class proceeding because the lack of commonality makes it an inappropriate class proceeding in the first place: Fischer v IG Investment Management Ltd, 2013 SCC 69, [2013] 3 SCR 949, at para 48.

[67] Accordingly, a class proceeding is not the preferred procedure for this claim. The proposed class members have statutory enforcement mechanisms as well as individual lawsuits open to them as alternative procedures. Since neither of those require any claimant to show commonality with another, both are preferable to this action under the CPA.

III. Disposition

[68] The Plaintiff’s motion for certification is dismissed.

[69] The parties may make written submissions on costs. I would ask counsel for the Defendant to provide me with brief submissions by email to my assistant within 3 weeks of today. I would then ask counsel for the Plaintiff to provide me with equally brief submissions by email to my assistant within 3 weeks thereafter.

Morgan J.

Date: July 15, 2022