Network Effect: The Growing Significance of Yelp as a Local Layer

The Yelp Syndication Network

You may have heard the recent news that Yelp has officially partnered with OpenAI to bring local listings content to ChatGPT, along with related features such as booking a restaurant reservation or requesting a quote from a service provider. This news is probably the strongest public acknowledgment yet on OpenAI’s part that local information is key to the success of ChatGPT as an all-purpose interface. 

For those of us in local marketing, that realization is not surprising. We know that a huge proportion (often estimated at 50%) of all search queries have some type of local intent. And we know that Google, the dominant name in search for more than two decades, has made local a central use case, to the extent of building a world-class local layer in the form of Google Maps.

In fact, Google’s dominance in local is likely one of the reasons a company like OpenAI would pursue a partnership with Yelp. Google is, of course, OpenAI’s chief competitor in the race for dominance in consumer AI. Given the popularity of local queries, one of Google’s key advantages in the AI race is owning the world’s most highly regarded and utilized local directory. As a competitor, OpenAI is unable to license local data directly from Google, so the company turns to Yelp, a comparatively neutral party who can offer a strong alternative. 

This is a story that played out in very similar ways in the pre-AI era, when companies like Bing and Apple made deals to license Yelp reviews, photos, and listing details in order to build mapping services that could compete with Google. 

In fact, many of these pre-AI relationships persist today; and as for AI companies, OpenAI Is far from the only one to license Yelp content. In fact, Yelp’s position as a neutral player that happens to offer world-class local reviews and listings data has quietly transformed the company over the last several years into the central node in a powerful distribution network.

Yelp data, comprising some 330M reviews and 500M photos for 14M businesses, now officially feeds into 6 of the top non-Google AI platforms, and can also be found in Apple Maps and search engines like Bing, Yahoo, and DuckDuckGo (which licenses Apple Maps including its Yelp layer). In addition, Yelp licensed data appears in automotive navigation systems as well as vertical apps like DoorDash. 

Official partnership is one thing, but if you’ve been using ChatGPT for a while, you probably know that Yelp content has been showing up in local recommendations for quite some time. This is, essentially, because of Yelp’s dominance as a site indexed in search. Even Google features Yelp content prominently and has done so for years in search. Now, Google’s AI interfaces, such as AI Mode and AI Overviews, cite Yelp more than any third party when it comes to local answers. 

Some brands express skepticism about the relevance of Yelp to their customer base, imagining, for instance, that Yelp is only popular for restaurants. This has always been a mischaracterization of the broadness of the platform, which has a strong representation of businesses in all categories. It’s even more of a problem to hold this limited view of Yelp when you consider its network reach. Consumers encounter Yelp star ratings, photos, and other content across a range of interfaces that have no industry-specific bias.

Quality of Review Content Differentiates Yelp from the Crowd

A recent article in Forbes entitled “How Yelp Became AI’s Ground Truth” lays out another key reason why Yelp’s content is so widely trusted: Yelp’s local reviews have maintained a level of quality over the years that is probably unmatched elsewhere. Yelp’s algorithm for promoting reviews as trustworthy is among the most choosy and complex of any platform. Users who have never published a review before, or who only post negative or only positive reviews, are unlikely to see their reviews promoted to “recommended” status. Only recommended reviews are featured in the main profile page for a business, used in its star rating calculation, and syndicated to other platforms. 

So too, Yelp has been strict in its policy of only allowing reviews that include text. Google, by contrast, publishes far more reviews, but a proportion of them include nothing but a star rating. 

Both companies try to fight spam in reviews, but Google’s policies for inclusion are inarguably more lax. Yelp’s hard-won reputation as a trustworthy source of consumer feedback makes it especially attractive to AI platforms.

The Long-Term Prognosis

This isn’t to say partnering with Yelp has no drawbacks. For a company like OpenAI, the clear advantage of a Yelp partnership is the ability to quickly scale up a competitive local offering in the United States. But Yelp’s global reach is less robust, so in order to grow outside the U.S., OpenAI will be forced to make another series of build–or-buy decisions. 

OpenAI and other platforms may recognize, sooner or later, that they must build a native architecture for local business interactions — a means for businesses to provide authorized data directly into the platform, and a means for consumers to register their feedback. So far, any such architecture goes outside the paradigm of generative AI, but it may still be an inevitable consequence of AI’s growth. 

But if the past can help us read the future, Yelp’s value is likely secure for some time to come. Like Google Maps, Reddit, and Wikipedia, Yelp represents a knowledge graph upon which sophisticated services like generative AI can be successfully anchored. Such knowledge graphs are critical to the operation of AI services and will likely remain so. They are not easily replaced.

And brand marketers who want to be visible across the spectrum of search and AI services, now and in the future, must treat Yelp as a critical platform.