Ky. Off. of the Att'y Gen., Complaint, Commonwealth v. RealPage, Inc. (filed July 2, 2025)

Citation: Ky. Off. of the Att'y Gen., Complaint, Commonwealth v. RealPage, Inc. (filed July 2, 2025)
Jurisdiction: Kentucky (state)
Source: Official source
Original Source: https://www.ag.ky.gov/Press%20Release%20Attachments/2025.07.02%20Realpage%20C... ↗

Related Parts of This Source

1,775 chars

RealPage. These guardrails have even spurred multiple landlords to tell RealPage that AIRM and

YieldStar are not dropping recommended rents as much as their individual conditions, or even

market conditions, would warrant.

119. Hard Floor. AIRM and YieldStar will not recommend a floor plan price that falls

below the smoothed market minimum effective rent. The market minimum is a hard floor. AIRM

and YieldStar thus explicitly constrain floor plan price recommendations based on the prices of

competitors, using shared nonpublic information.

120. Revenue Protection Mode. RealPage created a “revenue protection” mode that

effectively lowers output to increase revenues. Revenue protection activates when AIRM or

YieldStar predict—using calculations incorporating competitors’ data—that demand is too low

for a landlord to meet its target occupancy. Rather than lowering the price to stimulate demand,

the algorithm reduces the target number of leases. AIRM and YieldStar then maximizes revenue

for the reduced occupancy level, which tends to reduce price decreases or increase rental prices.

39
121. RealPage acknowledges that revenue protection “may seem counterintuitive to

leasing needs.” In June 2023, a landlord complained to RealPage that “something in your model

is broken” because “the pricing model is not lowering rents dramatically” despite the client’s

high exposure during a busy summer leasing season. RealPage explained that, with revenue

protection, “the model still sees the way to make more revenue is to lease fewer units at higher

prices.” In other words, the model seeks to “raise rates to get the highest dollar value possible for

the leases we can statistically achieve” and ignore those leases that the client wants but the model