Yelp adds personalized search results based on users' diet, lifestyle, accessibility, and other preferences, to its iOS app
For good or ill, it's common for internet services to track what you do and offer “personalized” suggestions. But Yelp, that stalwart of internet reviews …
Context & Ripple Effects
Yelp's move to rank results by diet, lifestyle, and accessibility preferences extends a line it had already been walking: by 2023 it shipped AI-powered search suggestions and review highlights alongside short-video uploads, turning its review corpus into structured signals rather than just pages of text. The personalization layer is what converts those signals into per-user rankings.
The competitive frame was set by Google, whose activity cards for recipes, jobs, and shopping showed where consumer search was heading a year after this launch. And the payoff for building rich preference profiles became visible years later, when Perplexity began licensing Yelp's maps, reviews, and business details for AI-generated restaurant recommendations — preference-tagged data is exactly what makes Yelp's corpus valuable to assistants.
First-order effects
- iOS users with dietary restrictions, accessibility needs, or lifestyle constraints get results filtered to their profile without re-stating them each search, while businesses matching those attributes gain visibility they couldn't buy through generic ranking.
Second-order effects
- Google's parallel investment in personalized activity cards means the two largest local-search players are now competing on who knows the user better, pushing both to deepen preference capture rather than compete on review volume alone.
Third-order effects
- If preference profiles keep compounding, local-search platforms shift from selling placement to licensing structured user-context data — the Perplexity deal previews a market where assistants pay for the preference graph behind the reviews.
The trend: Local discovery is moving from one-size-fits-all rankings to preference-profile-driven results, with the underlying user-context data becoming licensable inventory for AI assistants.