Proposed DOJ settlement provides guidance on use of competitive information in algorithmic pricing tools

DOJ allegations against RealPage

As discussed in our prior alerts regarding algorithmic pricing and information-sharing, DOJ brought a civil complaint charging RealPage and six multifamily residential lessors with violating Sections 1 and 2 of the Sherman Act based on allegations that RealPage’s revenue management software collects nonpublic competitively sensitive data from landlords to generate rental pricing recommendations, and the landlords shared this information both directly with each other and indirectly through the RealPage software.¹ DOJ also alleged that RealPage leverages landlords’ data to maintain a monopoly in the commercial revenue management software market, and has used “exclusionary conduct” to “obstruct rival software providers from competing on the merits via revenue management products that do not harm the competitive process.”² According to DOJ, RealPage’s “data and scale advantage is significant and creates a feedback loop that further increases barriers to competition for commercial revenue management software.”³

Terms of the proposed settlement agreement

The proposed settlement would allow RealPage to continue offering its revenue management tools subject to certain restrictions on the use of nonpublic competitively sensitive information in those tools.

The agreement provides a broad definition of “Competitively Sensitive Information,” (CSI) and includes under this umbrella information that: (1) “could be reasonably used to determine current or future rental supply, demand, or pricing for a property or a property’s units . . .” (2) relates to a property owner or manager’s use of settings or other “user-specific parameters;” or (3) relates to the “rental pricing amount, formula, or strategy, including rental price concessions of discounts” with respect to the property owner or manager’s property or properties.

The extent of the agreement’s prohibitions on RealPage’s use of nonpublic CSI depends on whether the data is being used in RealPage’s runtime operations⁴ or to train the models that run the company’s revenue management software. The agreement prohibits RealPage from using any nonpublic CSI—whether current or historical—in the runtime operation of any of its revenue management products. RealPage must also notify licensees that the company is not permitted to seek such information.

With respect to model training, however, the agreement provides that RealPage may train its models on nonpublic data collected from competitors if such data is at least 12 months old. In addition, RealPage will be prohibited from training certain revenue management models on nonpublic data that is at a geographic “granularity more specific than nationwide.”⁵ RealPage would also will be prohibited from using nonpublic data related to rental prices to train a model⁶ for the purpose of calculating market rent or market rank.

The agreement would also prohibit RealPage from:

• Soliciting, utilizing or sharing any information collected through market surveys for use in revenue management products or for purposes of recommending pricing or occupancy levels to users;
• Using the same modeled supply or demand curve across competitors in the same geographic area;
• Implementing certain design features in its revenue management products, including auto-accept features or any function that serves as a disincentive to reject RealPage’s price recommendations. Revenue management products also may not generate increased rental price recommendations when a floor plan reaches target occupancy (unless that function is based on a property’s own information); and
• Facilitating any discussions about market analysis or trends based on Nonpublic Data, owner-inputted data, or pricing strategies. In addition, pricing advisors will be prohibited from disclosing or disseminating Nonpublic Data or data inputted by property owners.

Finally, the settlement imposes compliance and monitoring requirements, including:

• Designating an antitrust compliance officer who will be tasked with establishing an antitrust compliance and annual training program for the company;
• Permitting compliance inspection upon request by the Assistant Attorney General for the Antitrust Division; and
• Submitting to the appointment of a Monitor selected by DOJ and approved by the court. The Monitor will have the power and authority to monitor RealPage’s compliance with the terms of the settlement agreement.

Takeaways

Although the proposed settlement does not have the force of law, it offers a glimpse into where DOJ is drawing lines between permissible and impermissible practices in the context of information exchanges and algorithmic pricing, which can help firms think about how to analyze risk in these areas. Among the insights from the settlement:

• The agreement does not entirely enjoin the practices that DOJ challenged and permits RealPage to continue offering revenue management tools and making pricing recommendations to competing landlords;
• The focus of the prohibitions appears to be more on how a third-party vendor like RealPage utilizes the information it collects from competitor firms, as opposed to prohibiting RealPage from collecting certain categories or types of information from competitors;
• DOJ can, and is willing to, engage in detailed and technical line-drawing around the specific information-exchange and algorithmic pricing conduct; and
• When it comes to determining which data is permissible to share between competitors, DOJ will likely focus on the age and nature of the data and will scrutinize those factors in the context of market realities, rather than applying universal benchmarks.

The last point may be among the most notable because of the contrast with the prior DOJ/FTC policy statements, withdrawn in 2023, which provided a “safe harbor” for competitor information exchanges if the data involved were sufficiently aggregated, anonymized by a third party, and sufficiently historical. The agencies took the position in those statements that competitively sensitive data older than 3 months old was unlikely to raise anticompetitive concerns, and this benchmark was regularly applied across firms and industries. By contrast, the proposed settlement prohibits RealPage from using any competitively sensitive data that is not at least 12 months old, and even then only in certain circumstances. This suggests that DOJ will be looking at the age of data in relation to the markets at issue and the specific context in which the data is used, rather than relying on a one-size-fits-all approach.

Authored by Ben Holt, Liam Phibbs, Holden Steinhauer, and Jill Ottenberg.