Official problem of the 2020 Canadian Competition Law Moot

Competition Tribunal of Canada

Reference: The Commissioner of Competition v Nile.com, Inc.

IN THE MATTER OF an Application by the Commissioner of Competition for an order pursuant to section 79 of the Competition Act, RSC 1985, c C-34, as amended;

BETWEEN:

The Commissioner of Competition
(Applicant)

And

Nile.com, Inc.
(Respondent)

REASONS FOR ORDER DATED OCTOBER 15, 2019

Canadian competition law moot


Table of contents

Introduction and Overview

Overview

  1. The Commissioner of Competition (the “Commissioner”) has filed an application pursuant to section 79 of the Competition Act, R.S.C. 1985, c. C-34, as amended (the “Act”), seeking an order to remedy the substantial lessening or prevention of competition (“SLPC”) that she alleges has occurred as a result of Nile.com, Inc.’s (“Nile”) alleged abuse of its dominant position.
  2. Nile is the largest e-commerce company in Canada. Its primary business is an e-commerce marketplace through which Nile itself – as well as third-party sellers – can sell and ship all kinds of products to customers Canada.
  3. Through its platform, Nile collects data from consumers and third-party sellers, which it uses to make predictions about consumers’ likely future purchasing decisions. Nile holds by far the most comprehensive consumer insight data available in Canada, and likely anywhere in the world. Using this consumer insight data (the “Data”), Nile has developed algorithms capable of predicting what products consumers are likely to purchase, and when. Nile makes its predictive models available to third parties through an application programming interface (“API”).
  4. Nile offers third-party sellers access to its API and Data for a fee. However, as a matter of policy, it only offers access to third-party retailers that sell their products through the Nile platform.
  5. Oak Inc. (“Oak”) is a retailer whose business model involves the sale of monthly personalized subscription boxes. Each box contains an assortment of grooming and hygiene products that are selected by Oak. Oak’s business model is not compatible with Nile’s platform, which does not support the sale of monthly subscription boxes by third parties. Oak’s product faced significant difficulty because many customers found that the products included in each monthly box were not relevant to their needs or interests, and cancelled their monthly subscriptions. In an effort to save its business, Oak sought access to Nile’s API and Data in order to better predict what types of products would be relevant to its customers. However, Oak was denied access to the API – in accordance with Nile’s policy – because Oak’s products were not sold through Nile’s platform.
  6. While Nile denied Oak’s request for access to the API, after speaking with Oak, Nile realized that its Data and API could be used to sell more attractive subscription boxes Nile offered to acquire Oak and several other subscription box suppliers; while Oak refused, Nile was successful in acquiring three other small subscription box suppliers in Canada. After the Commissioner’s application was filed in this matter, Nile launched its own subscription box, leveraging these acquisitions. As a result of fierce competition from Nile and its business challenges, Oak has discontinued its subscription box product and now sells hygiene and grooming products as a regular, third-party seller on Nile’s platform.

The Commissioner’s Application

  1. The Commissioner’s application alleges that Nile’s conduct toward Oak contravened section 79 of the Act. Section 79 of the Act contemplates a three-part test. The Commissioner submits that Nile’s conduct satisfies each part of this test:
    1. Nile substantially or completely controls the market for subscription boxes for personal care products in Canada, as a result of its control of an “essential facility” into the supply of such boxes: consumer insight data and predictive purchasing data;
    2. Nile engaged in a practice of anti-competitive acts by refusing to grant access to its API and Data to Oak (as part of a policy to withhold access to the Data and API to any third party that does not sell its products through Nile’s platform); and
    3. the practice has had, is having and is likely to continue to have the effect of lessening or preventing competition in the market for subscription boxes for personal care products.
  2. In response, Nile argues that:
    1. It cannot “substantially or completely control” the market for personalized subscription boxes for personal care products because, at the time of the Commissioner’s application, it did not sell subscription boxes. The mere ownership of Data and a predictive API does not confer control over downstream markets (such as subscription box sales) that may use the data and API as inputs;
    2. It did not engage in any practice of anti-competitive acts. Its refusal to grant access to its Data and API to Oak constitutes a mere exercise of its intellectual property rights; and
    3. There has been no prevention of competition associated with its actions. To the contrary, Nile has introduced a new lower-priced subscription box service that offers higher quality products through its use of the Data and API.
  3. This matter raises a number of important issues about how the Act should be applied to the digital economy. In the Tribunal’s respectful view, many of Nile’s arguments attempt to unreasonably narrow the meaning of s. 79, to the point that it would be incapable of addressing market dominance in high-tech industries.
  4. The Tribunal finds that Nile substantially or completely controls the market for subscription boxes. While Nile did not participate in this market at the time of the Commissioner’s application, it controlled two essential inputs – consumer insight data and predictive algorithms – that are necessary for success, as Oak’s experience demonstrates.
  5. The Tribunal further finds that Nile’s policy of restricting API access to Nile marketplace sellers constitutes a practice of anticompetitive acts. In particular, the Tribunal is persuaded by the Commissioner’s argument that the predominant intent and effect of Nile’s policy was to disadvantage and exclude non-marketplace retailers, such as Oak, in order to encourage them to become Nile marketplace retailers. This constitutes an exclusionary practice that has had the effect of preventing and lessening competition and innovation in the market for subscription boxes for personal care products, and in several other e-commerce markets. The Tribunal is not persuaded that Nile’s policy constitutes a mere exercise of its intellectual property rights.
  6. The Tribunal also finds that Nile’s actions have resulted in a substantial lessening and prevention of competition in the market for subscription boxes for personal care products. The Tribunal recognizes that – subsequent to its refusal to grant API access to Oak and subsequent to the Commissioner’s application – Nile has chosen to introduce its own subscription box product. While the Tribunal agrees that Nile’s product is cheaper and more relevant to consumers than Oak’s product, Nile’s actions have made it nearly impossible for innovative new subscription box products to enter and compete with Nile, and have therefore prevented (and lessened) competition substantially.
  7. However, as to remedy, the Tribunal finds that the Commissioner’s proposed remedy – compelling the separation of Nile’s data and retail businesses – is too extreme, and would be appropriate only in a truly egregious case. Accordingly, the Tribunal rejects the Commissioner’s proposed remedy and instead orders Nile to make its API available to all retailers – regardless of whether they participate on Nile’s marketplace – on consistent and commercially reasonable terms.

The Parties

The Commissioner

  1. The Commissioner is the public official appointed by the Governor in Council under section 7 of the Act to be responsible for the administration and enforcement of the Act.

Nile

  1. Nile is the largest e-commerce firm and the most valuable retailer in the world. Nile’s primary business involves the operation of an e-commerce marketplace through which virtually every kind of product can be sold and shipped to customers around the world, including in Canada. Through its online platform, Nile engages in two distinct types of e-commerce activities:
    1. Online retail: Nile purchases, stocks, sells and ships products to consumers, earning a margin based on the difference between the cost of goods sold and the selling price on its platform. Increasingly, as part of this business, Nile chooses to work with third-party manufacturers to design and produce white- labelled products under its NileBasics brand. Like many other “store brand” or “control label” products, NileBasics products are often inspired by pre-existing products that perform well on Nile’s platform, and are generally offered at a lower price than the branded equivalent.
    2. Marketplace: Second, Nile offers listing and logistics services to third parties, allowing them to sell their products through the Nile platform in exchange for listing fees (and other fees). Such sales are identified to the customer as Nile marketplace products, with third-party sellers operating their own virtual storefronts on Nile’s platform. Nile marketplace sales may or may not be shipped through Nile’s logistics infrastructure. Regardless, Nile never takes title to the products and does not control pricing or stocking decisions with respect to such sales. Many third-party sellers on the Nile marketplace have their own retail distribution infrastructure separate from Nile.
  2. Nile’s e-commerce platform allows it to collect enormous quantities of highly valuable consumer insight data. For instance, Nile collects data regarding customers’ searching, browsing and purchasing patterns, as well as their product reviews and returns. Nile also allows third-party sellers to provide consumer purchasing data for purchases taking place off its service (e.g., on retailers’ own websites or in their retail stores). As such, Nile’s Data comprises not only purchases made through Nile, but also data from many other retailers and retail formats.
  3. Nile leverages its Data to make various types of predictions. In particular, Nile can anticipate which products a particular customer is likely to be interested in purchasing in the future. Nile uses the Data to make predictions for its own operational and commercial purposes (e.g., to assist in determining when and where to stock certain products, to promote its own products for sale and to develop new NileBasics products). Nile’s also uses Data obtained from outside its platform (such as from retailer websites and bricks and mortar retail stores) to identify the purchasing habits of customers at alternative retail formats, and to attempt to improve its platform to compete more effectively against other retailers.
  4. Nile makes its consumer insight predictions available, in anonymized form, to certain third parties through its API. For a fee, permitted third parties can access Nile’s API to receive Nile’s predictions regarding their customers’ interests and likely future purchases. The Nile predictive API is made available through agreements allowing retailers to use the API and the underlying Data. As a matter of policy, Nile enters into such agreements only with retailers who sell through the Nile marketplace. Retailers who sell solely through other channels are not permitted to obtain access Nile’s API. Retailers who sell through both the Nile marketplace and other retail channels are eligible for access to Nile’s API.
  5. Nile’s public disclosure explains that this policy is designed to help Nile recover its investment in its predictive algorithms by incenting retailers to participate in the Nile marketplace (thereby growing Nile’s Data and the accuracy of its predictions). Nile’s public disclosure also explains hat the policy is intended to ensure that Nile has an opportunity to verify and improve its predictions by tracking the accuracy of its API’s predictions across the Nile platform.

Relevant Third Parties

  1. Although Oak is not a party to this proceeding, it bears a brief description here. Oak is a retailer of grooming, cosmetic and hygiene products for women and men. Oak sources niche, sample, overstock and “factory second” products and, prior to the events described below, resold them through a monthly subscription box service called OakChest. Customers subscribed to OakChest, provided Oak with basic demographic and product preference information and received a curated assortment of products each month. Product selection was tailored slightly from customer to customer based on the customer’s stated preference, but product selection was mostly dictated by the availability of attractively-priced products in sufficient quantities in a given month. OakChest subscriptions were sold through Oak’s website.
  2. Although OakChests were initially extremely popular due to Oak’s effective sourcing strategies, Oak struggled with very high levels of “churn”, as customers tended to cancel their subscriptions after only a few months. Oak’s research suggested this was because customers found that OakChests contained too many products that were of no interest, or which the customer already had in sufficient quantity.

Factual Background

  1. In early 2018, Oak found itself struggling with high levels of churn for OakChest, which was putting the company in an unsustainable financial position. Oak’s business model for OakChest involved a substantial discount in the first three months of a customer’s subscription, that was recouped through higher fees in the ensuing months. However, customers could cancel their subscriptions at any time, making churn particularly damaging to Oak’s business.
  2. Oak tried various strategies to learn about what types of products would be of greater interest to its customers, such as offering customers discounts for taking surveys, and running contests to obtain additional data from consumers. Oak was not able to obtain any meaningful consumer insight data owing to extremely low customer participation. Eventually, Oak executives realized that Nile’s predictive data would be an invaluable tool for improving the curation of the OakChest products, thereby reducing churn. Although it knew that Nile made the Data and API available only to Nile marketplace retailers, Oak nonetheless contacted Nile’s data sales division and requested a meeting. The meeting occurred on May 6, 2019.
  3. At the meeting, Oak requested an opportunity to purchase access to the API and Data on Nile’s normal commercial terms. Oak was not willing or able to sell OakChest through Nile’s platform because, at the time, the platform did not support monthly subscription sales, and in any event, Nile’s commissions and fees would make the product unprofitable. However, Oak offered to agree to the same commercial terms as Nile marketplace retailers, including paying the same fees (or even a premium above the standard fees) and submitting to the same pre-screening and security audit process. Nile responded that granting access to Oak would not be consistent with its policy, but agreed to consider whether it might be possible to make an exception, potentially for a higher fee.
  4. Following its meeting with Oak, Nile considered internally and determined that Oak’s concept of using predictive data to curate subscription boxes would be an excellent way to increase sales of NileBasics products, particularly new ones. With its scale, platform and access to Data and predictive insights, Nile was confident that it could develop a range of NileBasics subscription boxes that would substantially outperform third-party competitors in quality and price. In order to give the NileBasics subscription box project “the best possible chance of success” (as described in a Nile document), however, Nile decided to offer to acquire the leading subscription box providers or, failing that, to attempt to hire their key employees.
  5. Nile met with Oak again only a few days later, on May 13, and confirmed that because of its policy, it could not allow Oak to access the predictive API. However, Nile admitted that it was impressed with Oak’s idea and offered to acquire Oak at a price significantly above a normal multiple, and to offer leadership roles within Nile to Oak’s key senior managers. Nile conducted no formal diligence into Oak or its employees before making this offer. At the meeting, Nile suggested that Oak should “consider this offer carefully”, because the likely alternative was to be outsold by NileBasics and “left with nothing”.
  6. After conferring with counsel, Oak contacted the Competition Bureau, which commenced an inquiry and eventually made an application to the Tribunal under section 79 of the Act. In the meantime, Nile launched its NileBasics subscription box (NileBoxes), which has been extremely successful. NileBoxes offer personal care products at a significantly lower price than at traditional retail, and its pricing is much lower than any other provider of subscription boxes (including Oak). By accurately predicting which types of products customers like, and need to replace in any given month, NileBoxes are also significantly more popular. Nile’s internal documents and data supporting that customers maintain their NileBox subscriptions for at least 7-10 months, compared to 2 months for OakChests.
  7. Nile very quickly became a dominant provider of subscription boxes for personal care products, while existing competitors in this market struggled to retain their existing customers and experienced dismal results attracting new ones. Owing to its low sales of personalized subscription boxes, particularly following Nile’s entry into the market, Oak no longer offers a subscription box and instead operates as a traditional reseller of grooming and hygiene products, operating through the Nile marketplace.
  8. NileBasics now represents an overwhelming share of the subscription box business, well in excess of 90%. Indeed, a prominent industry analyst observed that “no investor should expect any new subscription box suppliers to reach viable scale going forward, particularly outside of the Nile marketplace. There is no question that Nile is simply untouchable.”

Position of the parties

  1. The Commissioner’s application alleges that Nile’s conduct toward Oak contravened section 79 of the Act because:
    1. Nile substantially controls the market for subscription boxes for personal care products in Canada within the meaning of paragraph 79(1)(a) of the Act, as a result of its control of a necessary input, consumer insight data;
    2. Nile engaged in a practice of anti-competitive acts within the meaning of paragraph 79(1)(b) of the Act by refusing to grant access to its API and Data to Oak; and
    3. the practice has had, is having and is likely to continue to have the effect of preventing competition substantially in the market for subscription boxes for personal care products, within the meaning of paragraph 79(1)(c) of the Act.

Paragraph 79(1)(a) – Substantial or Complete Control of a Class or Species of Business

  1. This Commissioner alleges that Nile does – and, at the time of its refusal to grant API access to Oak, did – substantially or completely control the market for subscription box sales. The Commissioner acknowledges that Nile did not supply subscription boxes for personal care products when it refused to provide API access to Oak. However, the Commissioner alleges that, consistent with the Federal Court of Appeal’s decision in TREB, Nile nonetheless controlled the market by virtue of its ownership of a key input – an “essential facility” – into the supply of subscription boxes: namely, consumer insight data and predictive algorithms.
  2. In Nile’s submission, “it is patently ridiculous to suggest that Nile substantially or completely controlled a market in which it did not even compete.” While Nile concedes that its Data and API are the most robust collection of consumer insight data available to third parties, it fervently denies that they confer control over other markets that may benefit from their use. Nile argues that the Act does not create an “essential facilities doctrine”.
  3. Nile further submits that its Data and API are “novel, but really nothing new”. Nile points to other large retailers (including Mall-Cart Stores, Inc., Priceco Wholesale Corporation and Bullseye Corporation) and large manufacturers of cosmetics products (including The Practical Gambler Company, O Really? S.A. and OneLeaver Plc). Nile argues that each of these large firms could readily replicate its consumer preference data and algorithms, such that they can confer no control over any market.

Paragraph 79(1)(b) – Practice of Anti-Competitive Acts

  1. The Commissioner alleges that Nile engaged a practice of anti-competitive acts by refusing to provide Oak and other non-marketplace retailers with access to its predictive API and underlying Data.
  2. The Commissioner submits that Nile’s refusal was intended, or could have been reasonably foreseen, to prevent Oak (and other similarly situated non-Nile marketplace sellers) from introducing an innovation that would enable it to compete more effectively in the market for subscription boxes for personal care products, in order to block these competitors’ innovations and allow Nile to enter and achieve a dominant position. The Commissioner further argues that Nile’s policy of providing access to its API only to third-party sellers on its marketplace was intended to entrench and maintain the dominance of its marketplace platform, by leveraging its dominance in consumer insight data and predictive algorithms.
  3. The Commissioner submits that Nile’s business justifications do not provide any credible pro- competitive rationale for its actions, and therefore do not constitute legitimate business justifications under the Act. Specifically, the Commissioner submits that the desire to recover an investment, earn a financial return or improve one’s own product is not a sufficient justification for otherwise anti-competitive conduct. The Commissioner points to Nile’s practice of acquiring suppliers of subscription boxes as further evidence that its strategy was intended to eliminate competition in the supply of subscription boxes, and not to further any pro- competitive objective.
  4. While the Commissioner concedes that Nile has intellectual property rights in its API and Data, she alleges that Nile’s refusal was not a “mere exercise” of an intellectual property right because its granting of access to the Data and API was selective, and moreover, was conditioned on the use of another of its products (i.e., its marketplace platform), which amounts to tying of products. Accordingly, the Commissioner submits that subsection 79(5) of the Act does not apply.
  5. In response, Nile submits that its policy is a mere exercise of its intellectual property rights its Data and API: the right to exclude others from using its intellectual property. Consistent with Warner Music and Tele-Direct, Nile submits that selective licensing of intellectual property rights – regardless of the reasons for such selective licensing – is nothing more than the mere exercise of an intellectual property right. It does not matter if Nile’s lawful exercise of its intellectual property rights had the effect of tying one Nile product to the use of another. These are merely network effects, which are not proscribed under the Act.
  6. Nile argues that it owns and is entitled to profit from its copyright in its Data and API, which are not raw data (as in Nielsen), but instead the product of Nile’s own compilations and analysis of that data, and as such the Commissioner’s only recourse would have been to apply to the Federal Court under section 32 of the Act.
  7. Nile further argues that its policy does not represent a practice of anticompetitive acts as it was not undertaken for any anti-competitive purpose. Nile’s policy of granting access to its API was instead wholly motivated by various legitimate business justifications, including:
    1. its interest in recovering its substantial and ongoing investment in its predictive algorithms by incentivizing retailers to participate in Nile marketplace (and thus further bolstering its Data used by its API and making its predictions more robust); and
    2. its interest in ensuring that it has an opportunity to verify and improve its predictions by tracking the purchases that are actually made following the use of its API.

Paragraph 79(1)(c) – SLPC

  1. Finally, with respect to paragraph 79(1)(c), the Commissioner alleges that Nile’s conduct has had, and continues to have, the effect of preventing or lessening competition substantially in the market for subscription boxes.
  2. The Commissioner argues that Nile – by refusing to provide access to its Data and API to suppliers of subscription boxes – foreclosed Oak’s intended product innovation, which would have enabled it and other subscription box suppliers to compete more effectively. Additionally, the Commissioner points to evidence that Oak and other suppliers of personal care subscription boxes have exited the market due to their inability to compete with Nile’s superior data for personalization. The Commissioner argues this lack of ability to compete, even for existing suppliers, makes clear that new entry into the market for subscription boxes for personal care products, outside of Nile’s marketplace, is extremely unlikely due to the barriers created by Nile’s policy.
  3. The Commissioner argues that, while Nile’s introduction of its own product has given consumers a significantly improved subscription box offering at a lower price, the relevant legal question is not whether prices have declined and quality has improved since the commencement of the conduct in question, but rather whether the competitive situation would have been even better had the conduct never occurred. In other words, the Commissioner argues that the market was likely to have been even more competitive, but for Nile’s anticompetitive conduct.
  4. Nile argues that it is inappropriate to conclude that a substantial lessening or prevention of competition has occurred when quality has improved and prices have fallen. Nile argues that it leveraged its innovative spirit to build its market share (including the accumulation of Data), which is entirely legal and ought to be encouraged. Nile argues that – across its business – it won through superior competitive performance. Nile argues that “demonizing and victimizing innovative companies and good ideas” would serve only to stifle innovation in Canada.

Remedies

  1. The parties made voluminous and passionate submissions regarding remedies.
  2. The Commissioner takes the view that a “mere” order requiring Nile to grant access to its Data and API on usual commercial terms to third-parties who do not sell through Nile marketplace would be insufficient to remedy the anticompetitive effects of Nile’s conduct. The Commissioner argues that Nile’s “lethal combination” – its retail business combined with its access to Data – has allowed (and will continue to allow) it to quickly dominate entire markets and drive untold numbers of smaller competitors out of business.
  3. Even if the Tribunal were to order Nile to grant commercial API access, the Commissioner argues, it would be difficult or impossible for any third party to compete with NileBasics. First, Nile does not pay a per-transaction cost to access the API (as other third parties must). Second, through its platform, Nile has access to additional data and insights not made available through the API. Third, through its scale and its access to control-label NileBasics products, Nile can source components for boxes more cheaply than its competitors. Overall, the Commissioner argues that Nile is one of the most dynamic companies in the world, and it is virtually certain to continue to generate and maintain advantages for NileBasics , which other retailers (both on and off the Nile marketplace) cannot hope to match. The Commissioner argues that this situation allows Nile to take competitors’ (or would-be competitors’) innovative ideas, and misappropriate them for its own benefit. Simply put, the Commissioner’s submission is that Nile’s scale and its history of exclusionary practices, including this latest one, have rendered further competitive entry impossible.
  4. The Commissioner argues that an attempt to protect competition in e-commerce through piecemeal, cat-and-mouse prohibitive orders would deplete and misdirect the resources of the Bureau and the Tribunal and, in any event, is bound to fail.
  5. Instead of requiring Nile to grant open access to its API and Data, therefore, the Commissioner argues that the Tribunal must rely on subsection 79(2) of the Act, which gives it the authority to issue an order directing the respondent to take any such actions as are reasonable and necessary to overcome the effects of its practice of anti-competitive acts, including the divestiture of assets or shares. The Commissioner submits that the Tribunal should issue an order requiring Nile to divest and separate its third-party marketplace platform from its first-party retail business so that they operate as separate entities. According to the Commissioner, this is not only consistent with the well-established preference for structural over behavioural remedies, but is in fact the only way to remove the inherent and irreconcilable conflict of interest between Nile as a platform operator and Nile as a retailer, and give third party sellers a fair opportunity to compete with Nile on even ground.
  6. For its part, Nile argues that if the Tribunal finds that a remedy is required, an order to grant API access to non-marketplace retailers would be sufficient to restore competition to any affected market. Nile argues that the Commissioner’s proposed remedy is a radical one: it is both unnecessarily burdensome and intrusive and, moreover, it would harm consumers by increasing the price and decreasing the quality of Nile’s retail offerings. Nile argues that , pursuant to Southam, the Tribunal should order the most minimal remedy required to eliminate an SLPC.
  7. As a result, Nile argues that if a remedy is required, the Tribunal should reject the Commissioner’s extreme and drastic proposal and instead order Nile to grant open access to its API.

The Issues

  1. This application requires the Tribunal to decide certain highly novel issues and apply scant precedent to cutting-edge facts. How these questions are ultimately resolved will have significant implications for the direction of Canadian competition law in an increasingly digital economy.
  2. In particular, the Tribunal must decide the following issues:
    1. Does Nile substantially or completely control the market for personalized subscription boxes for personal care products? (Section 79(1)(a) of the Act.). In particular, can a non- competitor in a market control the market through its ownership of an upstream “essential facility” (i.e., the Data and the API)?
    2. Did Nile’s refusal to grant access to its API and Data to Oak constitute a practice of anti-competitive acts? (Section 79(1)(b) of the Act.) In particular, is Nile’s policy of selectively granting access to its API and Data solely to non-marketplace retailers more than the mere exercise of an intellectual property right?
    3. Did Nile’s refusal to grant access to its API and Data to Oak substantially prevent or lessen competition substantially in the market for subscription boxes for personal care products in Canada? (Section 79(1)(c) of the Act.) In particular, has there been an SLPC despite the fact that Nile has since introduced its own subscription boxes at a lower cost and higher quality than those previously offered by Oak and other retailers of personalized subscription boxes?
    4. If so, is the appropriate remedy:
      1. to order Nile to grant access to the Data and API to all third parties (including non-Nile marketplace sellers) on commercially reasonable terms; or
      2. to order Nile to divest its first-party retail business from its third-party marketplace business?

Tribunal’s Analysis

  1. The Tribunal has carefully considered the parties’ submissions, the relevant jurisprudence and the evidence available to it. For the reasons below, the Tribunal has concluded that Nile does control the market for personal care product subscription box sales; that its refusal to grant access to the API and Data to non-Nile marketplace sellers did constitute a practice of anti-competitive acts; and that this practice has substantially lessened and prevented competition.

Substantially or Completely Control a Class or Species of Business

  1. The Tribunal must first determine whether Nile substantially or completely controls a “class or species of business”. The Commissioner alleges that the “class or species of business” at issue is the market for subscription boxes for personal care products. The Commissioner has not advanced any evidence as to whether Nile may substantially or completely control other markets (e.g., online retail, consumer insight data, etc.). Accordingly, the Tribunal will constrain its analysis to the question of whether Nile substantially or completely controls the market for personalized personal care product subscription box sales.
  2. In dynamic markets such as this one, there is an important question of temporality, which bears a brief mention. At the time that Nile refused to grant API access to Oak (i.e., at the time of the alleged “practice of anti-competitive acts”, discussed below), Nile had no personalized subscription box offering at all. It did not participate in the market. However, after it learned of Oak’s idea and Oak’s proposed use of the Data and the API, and after Nile refused to grant Oak access to the API and Data, Nile introduced its own personal care subscription box. Notwithstanding that it only entered the market after the alleged practice of anti-competitive acts, for the reasons described below, the Tribunal finds that Nile substantially or completely controlled the market for subscription box sales even at the time it first refused to grant API access to Oak.
  3. The Tribunal has held that “substantially or completely control” requires a firm to have a substantial degree of “market power”, which the Supreme Court has defined as “the ability to ‘profitably influence price, quality, variety, service, advertising, innovation or other dimensions of competition.’”
  4. The Tribunal has defined a “substantial” degree of market power as “confer[ring] upon an entity considerable latitude to determine or influence price or non-price dimensions of competition in a market, including the terms upon which it or others carry on business in the market”. In TREB, the Tribunal found that market power also includes the “power to exclude”: the ability to restrict the output of existing or potential market participants.
  5. Through Oak’s experience, it is readily apparent that access to predictive consumer data is necessary for an entity to compete in the supply of subscription boxes. Without such data, it is impossible to offer subscription boxes of personal care products competitively. Consumer data is thus an “essential facility” into the supply of these personalized subscription boxes.
  6. Nile argues that it is “patently ridiculous” to suggest that it controls a market in which it does not participate. But a landowner who controls the only well in an area dominates the local agriculture market, regardless of whether she farms. An electric company that owns the only power generation equipment in an area dominates the local manufacturing market, regardless of whether it builds. The owner of an airport terminal controls the local air transportation market, regardless of whether it flies. A ubiquitous payment network controls the market for payments, regardless of whether it accepts payments. And so Nile dominates the subscription box market, even if it does not produce a subscription box.
  7. Nile argues that there are other sources of retail sales data and consumer purchasing history data. The Tribunal agrees, and further recognizes that other consumer insight and data firms are available to businesses wishing to better understand their customers. But the Tribunal finds that Nile’s combination of consumer purchasing data and predictive algorithms are unique in the market; no other firms make such a combination available to third parties (even if other firms may develop similar tools for internal use). Oak’s own experience trying to assemble consumer insight data from other sources was unsuccessful. Accordingly, data from other providers is an imperfect substitute for Nile’s Data and API, and the availability of such data is not sufficient to mitigate Nile’s substantial and complete control over the sale of personalized personal care product subscription boxes.

Refusal to Grant Access to Predictive Data and API

  1. Was Nile’s policy of restricting API and Data access a mere exercise of intellectual property rights?
  1. The Tribunal must determine whether Nile’s policy of making its Data and API available only to Nile marketplace retailers represents only the “mere exercise” of an intellectual property right. Pursuant to subsection 79(5), “an act engaged in pursuant only to the exercise of any right or enjoyment” of an intellectual property right “is not an anti-competitive act”.
  2. The parties are in agreement, as is the Tribunal, that Nile has copyright in both the Data and the API. As a result, if Nile’s selective licensing of its Data and API is merely the exercise of its copyright, then the Commissioner must apply to the Federal Court pursuant to section 32 of the Act should it wish to pursue these practices under the Act, and the Tribunal has no jurisdiction to make an order under section 79.
  3. The Tribunal has considered what constitutes the mere exercise of an intellectual property right in a number of past cases. In Tele-Direct, the Tribunal considered an application from the Director of Investigation and Research (the former title of the Commissioner) submitting that Tele-Direct's practice of selective licensing was an abuse of dominance. In its reasons rejecting the Director’s application, the Tribunal held that something more than the mere exercise of statutory rights, even if exclusionary in effect, must be present before there can be a finding of misuse of a trade- mark. The Tribunal noted that “[t]he respondents' motivation for their decision to refuse to license a competitor becomes irrelevant as the Trademarks Act does not prescribe any limit to the exercise of that right.”
  4. The Tribunal subsequently applied Tele-Direct to copyright in Warner Music, noting that “[t]he Copyright Act is similar to the Trade-marks Act, in that it allows the trade-mark owner to refuse to license and it places no limit on the sole and exclusive right to license.”
  5. The Tribunal in Warner Music held that the right to exclude is a fundamental aspect of intellectual property rights:

    [t]he right granted by Parliament to exclude others is fundamental to intellectual property rights and cannot be considered to be anticompetitive, and there is nothing in the legislative history of section 75 of the Act which would reveal an intention to have section 75 operate as a compulsory licensing provision for intellectual property.

  6. Though not binding on this Tribunal, the Competition Bureau’s own Intellectual Property Enforcement Guidelines note that unilaterally exercising an intellectual property right to exclude does not violate the general provisions of the Act, no matter the degree to which competition is affected. The Bureau’s guidance provides that to hold otherwise could effectively nullify intellectual property rights, and impair or remove the economic, cultural, social and educational benefits they create.
  7. At the same time, as this Tribunal noted in TREB, the Tribunal and Federal Court of Appeal have interpreted the provision in subsection 79(5) narrowly. Effectively, anything that goes beyond a refusal to license should be interpreted as outside the scope of s. 79(5). For instance, the Tribunal in TREB found that even if TREB had copyright in its Multiple Listing Service (MLS) data (which it did not), TREB imposed restrictions on its members’ use of the data, and these restrictions took the conduct beyond a “mere exercise” of any intellectual property rights that it could have had in the MLS data.
  8. Accordingly, the question for the Tribunal is whether Nile’s refusal to grant Oak (and other non- Nile marketplace sellers) access to its API represents a mere refusal to license, or whether there is something more. In this case, the Tribunal is of the view that Nile is not only engaging in a refusal to license its Data and API to competitors. Nile’s refusal to license to third party sellers who do not sell on Nile’s platform means that Nile is conditioning the supply of its Data and API (products in which it is the dominant or exclusive supplier) on its customers’ use of the Nile platform. Nile therefore uses its Data and API to compel retailers to move away from sales channels in which Nile is not dominant (i.e., bricks and mortar or independent online sales) and toward a channel where Nile is dominant (i.e., the Nile marketplace). This scheme allows Nile to leverage its monopoly in one market to preserve, enhance and entrench its dominance in another market. This amounts to tying the supply of one product to the customer’s purchase of another, and is more than the mere exercise of a copyright.
  9. In NutraSweet, this Tribunal considered that “in appropriate circumstances, a trademark might be the subject of a tying arrangement”. The Tribunal ultimately found that tied selling was not made out, as the allegedly tying product to the trademark (NutraSweet) was the same product sold under that trademark (aspartame). In this case, the tying product is the Data and API, while the tied product is a sales channel (i.e., the Nile marketplace) – thus, two separate markets. This confirms that Nile’s refusal to grant access to the API and Data in the present case represents more than the mere exercise of its intellectual property rights.
    1. Was Nile’s policy of restricting API and Data access to marketplace Retailers an anti- competitive act?
  10. Having determined that Nile’s conduct amounts to more than the mere exercise of a copyright, the Tribunal must now determine whether Nile’s policy of granting API and Data access to only retailers who sell through Nile marketplace (or, put another way, its policy of refusing access to retailers who do not) constitutes a practice of anti-competitive acts.
  11. It was uncontested by the parties, and appears to the Tribunal to be incontrovertible, that Nile’s restrictive access policy was applied generally, over many years, across all retailers. Accordingly, the Tribunal has no difficulty concluding that the selective licensing was a “practice”, and must only determine whether the selective licensing itself was an anticompetitive act.
  12. Section 78 of the Act sets out a non-exhaustive list of types of conduct that are deemed to be anticompetitive acts. Paragraph 78(1)(e) provides that “pre-emption of scarce facilities or resources required by a competitor for the operation of a business, with the object of withholding the facilities or resources from a market” constitutes an anticompetitive act. Although Nile’s policy bears some similarities to pre-emption, the Tribunal’s view is that “pre-emption” refers to a practice of gaining control of necessary resources that already exist and does not comfortably describe a situation in which a necessary resource is withheld by the party that itself developed or created this resource. In any event, the Tribunal finds that Nile’s primary object in creating or obtaining the predictive data was not to withhold it from other competitors in the market, but to use it for its own commercial and operational purposes and to earn revenue by making it available through its API.
  13. Similarly, paragraph 78(1)(g), “adoption of product specifications that are incompatible with products produced by any other person and are designed to prevent his entry into, or to eliminate him from, a market” does not apply to Nile’s licensing policy, as there is no evidence to suggest that Nile deliberately designed its marketplace platform in a manner that was incompatible with subscription boxes in order to exclude Oak or any other supplier. Accordingly, the Tribunal finds that Nile’s practices do not fall squarely within those acts enumerated in section 78 of the Act.
  14. To qualify as an anti-competitive act under section 79, the act must be committed with an anti- competitive purpose that is predatory, exclusionary or disciplinary. There must be evidence linking the impugned practice to the subjectively or objectively intended negative effect on a competitor in a market. The Federal Court of Appeal held in Canada Pipe that “an anti- competitive act is identified by reference to its purpose” and it went on to observe that the term “purpose” may be determined by reference to subjective intent, but may also be determined by reference to “the reasonably foreseeable or expected objective effects of the act”; indeed, “evidence of subjective intent, although certainly probative if available, is not required in order to find that a given act is anti-competitive”.
  15. The Tribunal finds that Nile’s licensing policy had an exclusionary purpose: namely, preventing non-marketplace retailers from innovating and competing with Nile (including by introducing new business models, such as subscription box sales) or with suppliers who offers such boxes through the Nile marketplace. The Tribunal does not accept Nile’s submission that any exclusionary effects of its policy were an “inadvertent” consequence of the fact that Oak’s business model was not supported by the Nile marketplace platform. It would have been obvious from the inception of the policy that non-marketplace retailers (regardless of their business models) would be excluded. It would similarly have been obvious that the effect of Nile’s policy would be to increase and entrench Nile’s market power in the operation of online retail marketplaces (which would notably also have the effect of continuously expanding the amount of data feeding into Nile’s predictive API in a vicious circle, and the market power of it or any of its marketplace suppliers in the market for personalized subscription boxes). The Tribunal takes these facts as evidence of Nile’s exclusionary intent.
  16. There is no evidence of Nile’s subjective intent for its exclusionary licensing policy, other than the consistent message in its public and confidential documents. The Tribunal accepts that these subjective motivations are:
    1. Nile’s interest in recovering its substantial and ongoing investment in its predictive algorithms by incenting retailers to participate in the Nile marketplace; and
    2. Nile’s interest in ensuring that it has an opportunity to verify and improve its predictions by tracking the purchases that are actually made following an API call.
  17. The existence of some legitimate business purposes underlying anticompetitive conduct is not sufficient to immunize the conduct from scrutiny under section 79. Instead, “a business justification must be a credible efficiency or pro-competitive rationale for the conduct in question, attributable to the respondent, which relates to and counterbalances the anti- competitive effects and/or subjective intent of the acts”. In TREB, the Federal Court of Appeal notes two facts must be established:

    First, there must be a credible efficiency or pro-competitive rationale for the practice. Second, the efficiencies or competitive advantages, whether on price or non-price issues, must accrue to the appellant. Put otherwise, the evidence must demonstrate how the practice generates benefits which allow it to better compete in the relevant market.

  18. While Nile’s licensing practice does generate benefits which accrue to Nile, the Tribunal finds that Nile’s interest in recovering its investment is not “a credible efficiency or pro-competitive rationale”. If a firm’s interest in self-enrichment were sufficient to immunize otherwise anti- competitive conduct from Tribunal review, the Tribunal would be powerless. With respect to Nile’s interest in improving its predictions, the Tribunal agrees that this is a credible efficiency justification, but finds that it is not sufficiently connected to the policy of refusing to provide non- marketplace retailers with access to Nile’s API. Nile has many other “levers” to incent retailers to join the marketplace platform, such as by reducing its referral and service fees. As an additional example, it could grant non-marketplace retailers access to its API in the event that they shared their customer data.
  19. Finally, the Commissioner led evidence that Nile has a history of attempting to acquire firms as they grow large or innovative enough to represent future threats to Nile (creating a so-called “kill zone” around Nile’s business). Indeed, Nile attempted to acquire Oak at the same moment it refused to provide Oak with access to its API. The Tribunal finds that this practice of acquisitions provides further support for Nile’s exclusionary intent: excluding its competitors by acquiring them.
  20. The Tribunal finds that Nile’s refusal to provide non-marketplace retailers with access to its predictive data API constituted a practice of anticompetitive acts, and was not a mere exercise of its intellectual property rights.

SLPC

  1. Finally, having concluded that Nile’s conduct represented a practice of anti-competitive acts, the Tribunal must determine whether the conduct in question has had, is having or is likely to have the effect of preventing or lessening competition substantially in a market.
  2. Generally speaking, this determination requires consideration of whether the conduct preserves or adds to a firm’s market power, for example by preventing new competition that would have materialized in the absence of the impugned practice. The Tribunal must also analyse the pro- competitive aspects of the conduct. These factors may be assessed, inter alia, by asking whether prices in the market have risen and/or quality has decreased.
  3. The degree to which the anti-competitive acts create entry barriers in the relevant market should also be considered. Consistent with Nielsen, having concluded that the impugned conduct was intended to exclude actual and potential competitors, the Tribunal must assess the degree to which Nile achieved this goal. As noted in Nielsen, the Tribunal “must establish what the conditions of entry would be without the [anti-competitive conduct] and, then, determine how the anti-competitive acts altered the prospects for economically feasible entry.”
  4. As a result of Nile’s policy, Oak and other third-party subscription box suppliers were unable to pursue product innovation that would have improved their products and, by reducing churn, allowed them to offer lower prices and compete more effectively in the subscription box market. At the same time, the Tribunal recognizes that Nile introduced innovations itself and, thanks to its scale and other advantages (including its ability to avoid fees for API calls and source lower cost products, through its existing NileBasics range), was able to offer a better product at a lower cost than the third-party subscription box suppliers would otherwise have been able to offer had they used the API and Data themselves.
  5. The Tribunal recognizes that many dynamic markets – including, potentially, subscription boxes – are “winner-take-all” in nature, meaning that even if Oak and the other third-party suppliers had had access to Nile’s API, the NileBasics boxes may have out-competed them in any event, leaving the state of competition exactly the same as it is now. The Tribunal recognized in Tervita that “it is not enough that a potential competitor must be likely to enter the market; this entry must be likely to have a substantial effect on the market”.
  6. However, simply because a giant like Nile can enter a market or launch a product and effortlessly reduce prices to a greater degree than anyone else, it should not receive a free pass for conduct that prevents existing competitors from trying to do better and reduces future incentives for others to enter and innovate in the market. The Tribunal has previously described innovation as “the most important form of competition.” Nile’s conduct has not only prevented innovation by others, but has also left no incentive for remaining participants to continue to innovate. Only Nile is left to innovate. Neither Nile nor the Tribunal can say what Oak might have done with the Data and the API, in the absence of Nile’s practice of anti-competitive effects. With respect to non- price dimensions of competition, such as quality, variety, service, advertising or innovation, the test to be applied, consistent with Tervita, CCS and TREB, is whether the level one or more of those dimensions of competition was, is or likely would be materially lower than in the absence of the impugned practice.
  7. The Tribunal finds that Nile’s conduct has raised the barriers to entry and served to entrench Nile’s position in the market, thus preventing future competition in the subscription box market. This reduced competition has manifested in the form of, at the very least, reduced variety and innovation. In addition, as in Nielsen, Nile’s conduct has foreclosed one of the essential inputs for competing in the relevant market, raising barriers to entry to a level that renders future entry very unlikely.
  8. Accordingly, the Tribunal finds that Nile’s refusal to grant access to its predictive API and underlying Data resulted in an SLPC and thus contravened section 79 of the Act.

Remedy

  1. The Commissioner has argued for an unusual and unprecedented remedy: an order requiring Nile to divest its first-party retail business, so that it is separate from Nile’s marketplace platform.
  2. The breaking up of large, dominant firms represents a heroic part of antitrust history and it has been the subject of a great deal of recent interest. But, it is a “nuclear option” and is not to be undertaken lightly. The Commissioner argues that, in this case, it is the only remedy that will reliably restore competition to the Canadian e-commerce market.
  3. The Tribunal prefers to impose the least intrusive remedy where possible. However, as the Supreme Court of Canada made clear in Southam (in the merger context, but equally applicable here), the most important consideration for a remedy is that it must be effective in fully eliminating the SLPC, even if the fully effective remedy is not the least intrusive one:

    The appropriate remedy for a substantial lessening of competition is to restore competition to the point at which it can no longer be said to be substantially less than it was before the merger ... If the choice is between a remedy that goes farther than is strictly necessary to restore competition to an acceptable level and a remedy that does not go far enough even to reach the acceptable level, then surely the former option must be preferred. At the very least, a remedy must be effective.

  4. The Commissioner submits that because of Nile’s history of abusive practices, including these latest ones, it is too late to eliminate the SLPC and restore competition simply by requiring Nile to grant open access to its Data or API. According to the Commissioner, not only have many would- be competitors already been driven out of the market as a result of Nile’s conduct, but more importantly, NileBasics subscription boxes (and Nile’s first-party products more generally) enjoy a “moat” that no competitor can cross because they benefit from unique advantages that only Nile, as platform operator, can obtain. The Tribunal agrees.
  5. For example, because of Nile’s status as platform operator, its first-party retail business benefits from access to predictive data that is not made available to any third-party retailers, greater visual prominence on the platform, the avoidance of both API fees and other platform fees and myriad other advantages, including those not yet devised, that other retailers cannot obtain. These advantages place competitors at a severe, insurmountable competitive disadvantage. By tying access to its Data and API to the use of its platform by retailers, Nile further entrenches its market power in several markets, effectively leveraging its monopoly in an online retail platform to obtain the most robust predictive consumer insight data possible, and subsequently leveraging its dominance in predictive consumer insight data to enhance and entrench its market power as an online retail platform. The Commissioner argues that simply allowing access to Nile’s API and Data will not “break the vicious cycle”.
  6. The Commissioner submits that, as a result of Nile’s dynamism, the Competition Bureau and the Tribunal lack the resources and the speed to effectively prevent such abuses and preserve competition in large and growing Canadian e-commerce markets. A “structural remedy”, as the Commissioner terms its proposed remedy, would address the problem at the root by eliminating Nile’s incentive to prefer its own retail operations to those of other retailers on and off the platform. The Tribunal also recognizes a general preference for structural remedies over behavioural ones, as behavioural remedies are generally more difficult to enforce.
  7. The Tribunal is sympathetic to the Commissioner’s concerns regarding Nile and other large, dynamic firms that sometimes appear to have grown “too big to regulate”, and whose dominance in one or more markets can be leveraged in other markets. It appears unlikely that this will be the last or the most egregious case of reviewable conduct that Nile will need to defend before this Tribunal.
  8. Nile has argued that, even in the case of clear monopolies, the Commissioner and the Tribunal have accepted that granting access to key inputs is a sufficient remedy, pointing to both Nielsen and, though not binding as it was by way of consent order agreed upon between the Commissioner and the respondents, Interac.
  9. Indeed, the Tribunal’s jurisdiction to make remedial orders under section 79 is circumscribed. Abuse of dominant position is not illegal in Canada unless and until this Tribunal, and any reviewing court, determine that a particular course of conduct contravenes the Act and issues an order. The Tribunal recognizes that it cannot impose remedies to correct or prevent conduct that has not yet occurred. It falls to the Commissioner to be vigilant in fulfilling her mandate and bringing each case to the Tribunal for relief as it arises. If either the Commissioner or the Tribunal require more resources or different legislation to perform their respective responsibilities, that is a matter for Parliament.
  10. While the Tribunal agrees with the Commissioner that there is a possibility that it may be too late for mandatory open access to the Data and API to restore competition to the subscription box business or to other affected e-commerce markets, in the Tribunal’s view, a break-up of Nile is more likely to create its own even greater harms. For example, a break-up would reduce the incentive of Nile’s platform business to invest in improvements and would likely increase the price consumers pay for goods sold by Nile’s retail business. Where the “fully effective” remedy would eliminate the SLPC only to create even more harm, the Tribunal finds that it may choose a remedy that may be less than fully effective but leads to a better outcome in aggregate. Accordingly, the Tribunal finds that it is appropriate and sufficient in this case to order Nile to grant open access to its predictive data API and not to order Nile to divest its retail operations.

Order

  1. For these reasons, the Tribunal will order Nile to make its Data and API available to all suppliers of subscription boxes of personal care products on commercially reasonable terms, as set out in the Order issued contemporaneously with this decision.

DATED at Ottawa, this 15th day of October 2019.
SIGNED on behalf of the Tribunal by the Panel Members.

Date modified: