Tuesday, March 19, 2019

Expedia's "sold out" labels and diversionary phone numbers for hotels lead to class certification


Buckeye Tree Lodge & Sequoia Village Inn, LLC v. Expedia, Inc., No. 16-cv-04721-VC, 2019 WL 1170489 (N.D. Cal. Mar. 13, 2019)

Really interesting false advertising class action, with a smaller-than-plaintiffs-wanted class of hotels certified against Expedia, which offers hotel bookings on its websites. The plaintiffs own hotels that are not available through Expedia. Apparently, the information of a hotel that hasn’t contracted with Expedia sometimes “finds its way onto one of the numerous websites Expedia operates, including Expedia.com, Hotels.com, Orbitz.com, and Travelocity.com.”  When customers search for their hotels on Google or within one of Expedia’s websites, Expedia allegedly falsely suggests that these hotels can generally be booked on the websites, but that they are “sold out” for the period that the customer wants to book them. Expedia then allegedly steers customers to similar hotels that are, in fact, available on its websites.  I’m interested in what Eric Goldman will think—he generally doesn’t like holding websites liable for design features, but “sold out” sure seems likely to mislead consumers. The Lanham Act is strict liability, and it may make sense to require advertisers to bear the costs of mistakes like this one--as this case indicates, they lack nonlegal incentives to fix the problem on their own.

An ad on Google may offer a customer the chance to book a room at the searched hotel through Expedia’s websites. When the customer clicks on the ad and navigates to the “infosite” page for that hotel on Expedia’s sites, they will see a picture of the hotel, a map of where it’s located, a star rating, and details about the property’s amenities, like parking, wifi, and breakfast. Sounds pretty standard for a website that can provide bookings, but … “Next to the hotel’s name and information, the website will communicate some variation of the message: ‘We are sold out.’” The same basic thing happens to a customer who starts searching on an Expedia site and picks a specific not-carried hotel from the list of search results.

Expedia also lists a phone number with each hotel, both on the infosite and search results pages. But “[t]he numbers simply connect to the Expedia call center,” and call center employees are trained on how to handle “customers [who] think they are calling the hotel directly,” “[b]ecause ... the placement of our phone numbers within the hotel search results or on the hotel details page of the website[ leads] some customers [to] think they are calling the hotel directly.” The employees are instructed to encourage callers to book at hotels that have booking agreements with Expedia.  (Okay, even I think an initial interest confusion claim passes the laugh test here.)

For named plaintiff Buckeye Tree Lodge, from January 2015 through August 2016, 149 people landed on the Buckeye infosite page on Expedia; four of those people subsequently made a reservation with another hotel during the same visit to the website.  Only one of the named plaintiffs was able to get itself removed from Expedia before joining the lawsuit; the others weren’t until they sued/joined the suit.

To certify a Rule 23(b)(2) class to pursue injunctive relief, the plaintiffs needed to show standing, including a likelihood of future injury for the class. Expedia argued that the named plaintiffs lacked standing because they weren’t listed on the websites anymore. But defendants can’t “rob a court of jurisdiction by taking strategic unilateral action to moot a plaintiff’s claims before the plaintiff has had the opportunity to seek class certification.” Expedia’s reform wasn’t “genuine, irrefutably demonstrated, and comprehensive”: there was no evidence that it had “made a meaningful attempt to ensure that its website will stop suggesting that hotels it cannot book are sold out.” Indeed, the court thought that the named plaintiffs themselves could end up on Expedia’s websites again because Expedia hadn’t taken measures to stop that.

Rule 23(a)’s numerosity, commonality, typicality, and adequacy of representation requirements were satisfied, as was Rule 23(b)(2)’s condition that the plaintiffs seek “uniform relief from a practice applicable” to the whole class, but only for a specific class: “owners of hotels that do not have booking agreements with Expedia and are not capable of being booked through Expedia, but appear on Expedia’s websites.”  Commonality was the only serious question, and proceeding as a class would generate common answers to the false advertising requirements given the facts above. Did the phone numbers mislead consumers? Did “We are sold out” mislead the reasonable consumer to think that the hotels are fully booked? If so, were the hotels likely to lose business due to Expedia’s conduct? Classwide evidence could be provided, including through surveys.

Expedia argued that it used different language during the class period, including “We have no rooms available for your selected dates ...”, “No rooms available on our site for the selected dates ...”, or “Sorry, the [hotel name] is not available on Hotels.com for your travel dates. You may choose alternative travel dates OR select from the hotels below.” It didn’t track which unavailability messages were displayed to which customers. However, three of the websites used the same unavailability message on their search results page throughout the class period: “We are sold out.” Further, the messages weren’t so different or so numerous that they couldn’t be evaluated through common proof. And because a class was being certified only for injunctive relief, the plaintiffs might only need to evaluate the current messages.

Expedia also argued that typicality was defeated by the myriad unique ways through which the plaintiffs’ information landed on the websites. For example, named plaintiff Buckeye had begun the onboarding process to create an account with Expedia, but ultimately did not go through with the agreement. Plaintiff Mansion’s information got to Expedia because the hotel had contracted to share its information with a third party, TravelClick, that itself disseminated the information to three distribution systems, which in turn gave the information to Expedia. When the Mansion ended its relationship with TravelClick, two of the distribution systems failed to tell Expedia that it would no longer be providing the Mansion’s data, and the other named plaintiffs had similarly complicated stories of data-sharing.

But the plaintiffs were all not capable of being booked through Expedia, which made the Lanham Act question of deception uniform.  Other common questions included: “Given the automated-but-disorganized nature of the online travel industry, does the Lanham Act impose on Expedia an affirmative obligation to institute controls to ensure that its systems are not causing customers to be misled about availability at unaffiliated hotels? And if so, is Expedia satisfying this affirmative obligation?”

And injunctive relief would uniformly benefit the class members. “Expedia could also be required to clearly indicate when listed phone numbers connect directly to its call centers. It could be enjoined to ensure that Google ads pertaining to unbookable hotels clearly say as much. And … Expedia could be ordered to take measures to better comb its system to ensure that unaffiliated hotels are not being listed in the first place.”

The court declined, however, to certify a Rule 23(b)(3) class to seek disgorgement Expedia’s profits for want of predominance. Plaintiffs “failed to proffer a model or a legitimate theory for how those damages would be estimated, let alone disseminated among class members.” They couldn’t explain how they’d prove sales related to Expedia’s false advertising. An earlier model estimated the percentage of all hotels in Expedia’s inventory that were actually unaffiliated and were marked, like the plaintiffs’ hotels were, as “sold out.” Plaintiffs’ counsel previously proposed that their disgorgement award should be that percentage of Expedia’s total revenue, but that percentage wasn’t logically connected to consumers’ reaction to the false advertising here.  Nor would the court certify a limited class action to decide common issues of liability under Rule 23(c)(4) class, because that wouldn’t advance the disposition of the litigation as a whole.

No comments: