Algorithmic Pricing and Competition Law in Canada
Algorithmic pricing can make markets more efficient, but it can also facilitate collusion, exclusion and misleading pricing. The Competition Bureau published a discussion paper in June 2025 and consulted until August 2025. This note explains how the Competition Act applies to pricing algorithms, shared software and personalized pricing, and the controls businesses should adopt.

Key takeaways
- Agreements implemented through pricing algorithms can be criminal under s. 45.
- Shared software using competitors' non-public data can create hub-and-spoke risk.
- Tacit coordination is generally not an offence but matters in merger review.
- Dominant firms' algorithmic conduct is reviewable under s. 79.
- The Bureau's algorithmic pricing consultation ran from 10 June to 4 August 2025.
Why algorithmic pricing raises competition questions
Algorithmic pricing uses software, often powered by machine learning, to set or recommend prices based on data about demand, inventory, costs, competitors' prices and customer behaviour. It is now common in retail, travel, hospitality, rental housing, ride-hailing, online marketplaces and many business-to-business markets. Most algorithmic pricing is lawful and can make markets more efficient by responding quickly to changes in supply and demand. But it can also create competition risks: competitors using the same software or data pool may end up coordinating prices, a dominant firm may use algorithms to exclude rivals, and personalized pricing may mislead consumers. In Canada, these risks are assessed under the general provisions of the Competition Act.

The Competition Bureau has signalled its interest. On 10 June 2025 it published Algorithmic pricing and competition: Discussion paper and held a public consultation that ran until 4 August 2025. The paper describes how algorithmic pricing works, the data it uses, and potential issues under the Act, including algorithmic price-fixing and competitor collaborations, practices of anti-competitive acts and deceptive marketing, and how algorithms factor into competition analysis such as entry, personalized prices and consumer switching. This note explains those issues and the compliance controls businesses should consider.
Algorithmic collusion and the conspiracy offence
The criminal conspiracy offence in s. 45 requires an agreement or arrangement between competitors to fix prices, allocate markets or restrict output. An algorithm cannot itself agree, but people can use algorithms to implement or monitor an agreement. Where competitors agree to use a common pricing tool to align prices, or agree on rules to program into their own tools, s. 45 can apply in the same way as to any other price-fixing scheme. The court may infer an agreement from circumstantial evidence (s. 45(3)), so communications about adopting a tool, configuring it or following its recommendations can matter.
Harder questions arise where competitors independently adopt the same third-party software that uses pooled, non-public data from multiple clients to recommend prices. There may be no direct communication between competitors, but the software can act as a hub through which competitively sensitive information flows. Whether such a hub-and-spoke arrangement amounts to an agreement under s. 45 depends on the facts, including what each client knew about others' participation and whether they committed to follow recommendations. Even where s. 45 does not apply, the arrangement may be reviewable under the civil provision in s. 90.1 if it substantially lessens competition.
| Scenario | Main provision | Risk level |
|---|---|---|
| Competitors agree to use a tool to align prices | s. 45 (criminal) | High |
| Vendor pools non-public competitor data and recommends prices to all | s. 45 (hub-and-spoke) or s. 90.1 | Medium to high, fact-dependent |
| Each firm uses its own algorithm on public data | Generally lawful; possible s. 79 if dominant | Low to medium |
| Dominant firm uses algorithm to target rivals' customers below cost | s. 79 (abuse of dominance) | Medium to high |
| Personalized prices with misleading "best price" claims | ss. 52, 74.01 (deceptive marketing) | Medium |
Tacit coordination and the limits of the law
Some algorithms can learn to avoid price competition without any agreement, by reacting predictably to rivals' price changes. This "tacit" coordination is generally not unlawful under s. 45, which requires an agreement. It can, however, be relevant in merger review, where s. 93(g.5) allows the Tribunal to consider the likelihood that a merger will result in express or tacit coordination, and in market studies, where the Bureau can examine whether market features make coordination more likely. Businesses cannot rely on the absence of an express agreement if their conduct around the algorithm, such as sharing configuration details with competitors, provides evidence of one.
Merger review is another context in which algorithmic pricing matters. When the Bureau assesses a transaction in a market where firms rely heavily on pricing algorithms, it may consider whether the merger makes coordination more likely, for example by reducing the number of independent pricing systems, combining data sets that make rivals' behaviour more predictable, or bringing competing clients of the same pricing vendor under common ownership. Parties to such deals should be prepared to explain how pricing decisions are made and why the merged firm will continue to price independently of its rivals.
Abuse of dominance and personalized pricing
A dominant firm could use algorithmic pricing in exclusionary ways: targeting rivals' customers with selective low prices, adjusting prices to punish suppliers that deal with competitors, or using data advantages to entrench its position. Section 79 now allows orders based on substantial anti-competitive effects in a market where the firm has a plausible competitive interest, and penalties reach the greater of CAD 25 million and three times the benefit, or 3% of worldwide revenues.
Personalized pricing, where different customers see different prices based on their data, is not unlawful as such. The deceptive marketing provisions apply, however, if a business represents that a price is the lowest available, a regular price or a discount when that is not true for the customer, or if the general impression of a pricing practice is misleading. Drip pricing rules apply equally to algorithmically generated prices. Privacy laws, including federal and Quebec private-sector privacy legislation, govern the collection and use of the personal information that personalization relies on.
Before licensing a pricing tool, ask the vendor whether it uses data from other clients, whether those clients include your competitors, how data is aggregated and whether recommendations reflect other clients' confidential prices. Record the answers. If the tool pools competitors' current, non-public data, get advice before using it.
Compliance controls for algorithmic pricing
- Inventory tools. Identify every system that sets or recommends prices, including third-party software and marketplace tools.
- Understand inputs. Confirm what data each tool uses, especially whether it includes competitors' non-public information.
- Keep pricing decisions independent. Retain the ability to override recommendations and document independent decision-making.
- Restrict communications. Do not discuss pricing tools, settings or results with competitors, including at vendor user groups.
- Review vendor contracts. Require confidentiality of your data and prohibit its use to set competitors' prices.
- Check consumer-facing claims. Ensure price representations remain accurate when prices change dynamically or vary by customer.
- Audit periodically. Test outputs for unexplained parallel pricing with competitors and review exclusionary effects if the firm is large.
The Bureau's Competitor Collaboration Guidelines and its guidance on information sharing apply to data pools and benchmarking services that feed pricing algorithms. See our note on information sharing between competitors.
Algorithmic pricing across sectors
The risks look different from sector to sector. In residential rental housing, revenue management software that recommends rents for competing landlords using pooled lease data has attracted scrutiny internationally, because it can reduce the independent pricing that keeps rents competitive in a local market. Canadian landlords using such tools should understand whether the vendor pools data across competing buildings and whether recommendations are presented as defaults that are rarely overridden. In travel and hospitality, dynamic pricing of fares and room rates is long established; the main issues are the source of competitor data, the use of shared channel managers and the accuracy of advertised prices once fees are added.
In retail and e-commerce, repricing tools that track competitors' public prices are widespread and generally lawful, but sellers on the same marketplace should avoid any understanding to set minimum prices or match each other's increases. In business-to-business markets, such as industrial inputs and logistics, pricing tools that incorporate bid histories or customer-specific data need careful access controls, because those data can be highly sensitive if shared across competitors. In each sector, the questions are the same: whose data goes into the tool, who sees the outputs, and do competitors' decisions remain independent.
Governance, documentation and board oversight
Algorithmic pricing should be governed like any other material risk. Businesses should assign ownership of each pricing system to a named manager, keep records of how it was configured and why, log significant overrides, and require legal review of changes to data sources or objectives. Where machine-learning models are used, the business should understand what the model is optimizing for; a model rewarded for maximizing margin in a concentrated market may learn behaviour that looks like coordination even if no one intended it. Periodic testing, for example comparing price movements with those of key competitors, can reveal patterns worth investigating.
Boards and senior management should receive periodic reports on the use of pricing technology, the compliance controls in place and any issues identified. If a problem emerges, such as evidence that a vendor shared confidential data across clients, the business should take advice promptly on its options, which may include stopping use of the tool, preserving evidence and considering the Bureau's Immunity and Leniency Programs if an agreement may exist. Clear documentation of independent decision-making is the best protection if the Bureau or a private plaintiff later questions pricing outcomes.
Responsibilities of software vendors and platforms
Vendors of algorithmic pricing software and platforms that set prices for sellers are not immune. A vendor that knowingly facilitates an agreement among competitors may itself be party to the offence, and a platform's pricing tools may raise abuse of dominance issues if the platform is dominant. Vendors should design products so that each client's confidential data is not used to recommend prices to its competitors, offer clear documentation of data use, and avoid marketing their products as a way to "stabilize" prices or avoid competition. Platforms that offer pricing tools to sellers should consider whether default settings or parity requirements affect competition.
What to expect next
The Bureau's consultation closed in August 2025. Its discussion paper sets out how algorithmic pricing may be assessed under the existing provisions, and the Bureau also has market study powers, introduced in 2023, that allow court-ordered information requests in a formal inquiry into the state of competition in a market. Given the expanded civil provisions, private access to the Tribunal and international enforcement activity, businesses that rely heavily on algorithmic pricing should treat compliance as an ongoing governance issue rather than a one-time review. Tools, data sources and vendors all change frequently, and a configuration that was safe when adopted may become risky after a software update or a change in the vendor's client base.
How TRW & Co helps
We advise retailers, platforms, software vendors and businesses that use pricing tools on algorithmic pricing under Canadian competition and consumer law. Our work includes vendor due diligence, contract terms, compliance programs and audits, product design advice for vendors, and responses to Bureau inquiries. See our pages on cartels and criminal enforcement and artificial intelligence, and our note on digital platforms competition. This is general information, not legal advice. To discuss your pricing tools, contact us.
Frequently asked questions
Is algorithmic pricing legal in Canada?
Yes, in general. Businesses may use algorithms to set prices independently. Risk arises when algorithms implement an agreement between competitors, when shared tools pool competitors' confidential data, when a dominant firm uses them to exclude rivals, or when price representations to consumers are misleading.
Can using the same pricing software as competitors be illegal?
It can be, depending on the facts. If the software pools competitors' non-public data and recommends prices to each, and the competitors know others participate and follow recommendations, the arrangement may support an inference of agreement under s. 45 or be reviewable under s. 90.1. Understanding the tool's data inputs is essential.
What has the Competition Bureau said about algorithmic pricing?
On 10 June 2025, the Bureau published a discussion paper on algorithmic pricing and competition and consulted until 4 August 2025. The paper covers how algorithmic pricing works and potential issues under the Act, including price-fixing, anti-competitive acts and deceptive marketing.
Is personalized pricing allowed?
It is not unlawful as such, but representations about prices must not be false or misleading, drip pricing rules apply, and privacy laws govern the use of personal data. A dominant firm's personalized pricing may also be examined under the abuse of dominance provisions if it excludes rivals.
Sources
General information only, not legal advice. References to statutes, courts and authorities describe published law and practice at the date shown.


