Certifying Your Best Listings Quietly Devalues All the Others

Based on the research ofKim, Wang, and Montauti, "When Certification Backfires: Negative Spillovers on Noncertified Offerings," Organization Science, 2026

The intuition behind every trust badge is that it lifts the whole platform: guests feel safer, certified hosts earn a premium, and the reputational halo spills over to everyone else. Heeyon Kim, Qian Wang, and Martina Montauti went looking for that halo around Airbnb Plus and found its opposite. When Airbnb began labeling a small set of inspected, design-vetted homes as "verified," the listings that did not get the label did not simply stay put. They lost revenue. Nothing about those homes had changed. What changed was the frame they were now seen in.

The mechanism is a contrast effect, and it is unsettling precisely because it requires no bad behavior from anyone. Certification works by drawing a bright line between the vetted and the unvetted. Once that line exists, buyers do not read the unvetted side as "presumably fine, just not premium." They read it as "not verified, so what is wrong with it?" The badge manufactures a new question in the shopper's mind that did not exist before the program launched, and the noncertified majority has to answer it with the same information it always had.

The contrast trap

The paper's design is what makes the finding hard to wave away. Airbnb rolled Airbnb Plus out in Los Angeles in early 2018 but not yet in San Diego, two markets close enough to share tourists, weather, and short-term-rental regulation. That staggering gave the authors a clean difference-in-differences comparison: track noncertified listings in the treated city against near-identical noncertified listings in the untreated one, before and after the badge appeared. The result was a negative effect on the revenue of noncertified Los Angeles listings, a drop that coverage of the study put at roughly 10 percent, with no corresponding change in the control city. The homes that were harmed were not competing for the Plus badge and did nothing differently. They were simply standing next to listings that now carried one.

What sharpens the effect is how narrow the certified set was. Airbnb Plus was invite-only, gated behind an in-person inspection against a 100-point checklist, and it never scaled: Airbnb targeted 75,000 listings and had reached only about 26,000 by early 2020. At its peak the program covered just over 3 percent of supply. So the contrast ran the wrong way for almost everyone. A tiny, glossy verified tier became the reference point against which the other 97 percent were silently re-graded. As one of the authors put it in Cornell's write-up, by explicitly labeling some homes "verified," Airbnb "unintentionally made the lack of verification stand out more sharply for everyone else." The platform did not add information about the uncertified listings. It added a spotlight that made their existing information look inadequate.

Certification does not just rank the certified above the uncertified. It rewrites what "uncertified" means, from ordinary to suspect.

Why the damage is not spread evenly

The negative spillover is not a fixed tax. Kim, Wang, and Montauti show it bends with the information environment, and this is where the finding turns from a warning into a design brief. The harm shrank on three margins. It was weaker when a noncertified listing already carried strong alternative quality signals, a deep bank of reviews, high ratings, because a buyer staring at four hundred five-star reviews does not need a badge to resolve the uncertainty the badge created. It was weaker when the noncertified listing was clearly not comparable to the certified ones, so the contrast never triggered. And it was weaker when the listing was hard to substitute off the platform, where the guest had nowhere better to defect to. The through-line is that the contrast effect bites hardest exactly where the shopper has nothing else to go on.

That is why the effect should worry any platform running a thin, editorial-style badge on top of a shallow information base. Amazon's Choice is the cautionary version. It is awarded algorithmically, not by human vetting, from a blend of price, availability, ratings, and conversion, and it has drawn sustained criticism for endorsing products that did not deserve it, from a widely inaccurate infant thermometer rated 3.6 stars to padlocks a reviewer could pick in two seconds, sometimes propped up by review manipulation the algorithm never caught. Set aside whether the badged product is good. The structural point from the anchor paper is the one operators miss: every time that badge appears next to a competing listing, it re-frames the unbadged one as the thing the platform declined to single out. On a category page where the badge is the loudest signal in view, that contrast does the most damage, precisely the low-information setting where the paper predicts the spillover is worst.

The programs that quietly learned the lesson

Watch what Airbnb did next, because it reads like the paper's managerial section enacted. Airbnb Plus was discontinued, stopping altogether on November 6, 2023. Its replacement was not another narrow verified tier but Guest Favorites: a listing-level distinction covering roughly two million homes, on the order of a quarter to a third of all active listings, awarded automatically from guest ratings, review history, and reliability data drawn across hundreds of millions of trips, with an average rating above 4.9 stars. Two design choices matter here. First, the honored set is broad, not a 3 percent elite, so the "everyone else" that the contrast punishes is a much smaller and lower-quality residual. Second, and more to the point of the paper, the badge is computed from the very signals, ratings and reviews, that the study identified as the brake on negative spillover. The quality information is not hidden behind an inspection the buyer never sees; it is the same public review data the shopper can already read. Airbnb rebuilt its top-of-market signal around the moderator that makes certification safe.

This is the real lesson, and it is not "never certify." Certification genuinely reduces uncertainty for the certified, and platforms need premium signals to compete. The lesson is that a badge is a two-sided instrument: it resolves uncertainty for the labeled and manufactures it for the unlabeled, and the second effect lands on the far larger population. A program that certifies a sliver, hides the underlying evidence, and lets the badge be the only visible quality cue is engineered to maximize the spillover harm. A program that honors a broad set on transparent, buyer-visible criteria is engineered to minimize it. Superhost, keyed to a host's actual track record and displayed alongside the reviews that justify it, sits closer to the safe end for the same reason.

What certification design owes the uncertified

The forward-looking question for any platform is no longer "should we add a trust badge" but "what happens to the majority we don't badge." That population is usually where most of the platform's supply, and most of its transactions, actually live. If certifying the top decile quietly shaves revenue off the bottom nine, a program that looks like a premium upsell is really a redistribution, and if it alienates the base that carries the marketplace, expect it to be scaled back, exactly as Airbnb Plus was. The instrument to watch is not the badge's adoption rate among the certified. It is the revenue trajectory of the noncertified, measured against a market where the badge never launched.

So the design rule falls out cleanly. Certify against criteria the buyer can see, so the badge summarizes evidence rather than replacing it. Keep the honored set wide enough that "uncertified" does not read as "rejected." And keep ratings, review depth, and verified history loud on every noncertified listing, because those are the signals that stop a shopper from filling the badge-shaped hole with worst-case assumptions. Do that, and certification informs. Skip it, and you have paid a small verified minority a premium by taxing the unlabeled majority who never knew they were entered in the contest.