🧠 Food for thought

Implications, context, and why it matters.

Yelp doesn’t disclose how AI features affect profitability or unit economics

– Yelp posted record Q2 2025 net revenue of $370 million and adjusted EBITDA of $100 million 1. Adjusted EBITDA means earnings before interest and taxes. It excludes depreciation. It also excludes amortization and is adjusted for one-time items. Management did not say how Assistant or Menu Vision or natural language search affect conversion, CPC beyond the 11% rise, or spend per location 1.
– The company has not shared inference costs or margin effects from running large language models at scale 1. AI search API calls grew 20x year over year, with related revenue now above a $10 million run rate 1. Investors still do not know if these tools lift or squeeze the 27% adjusted EBITDA margin 1. Leaders also target stock-based compensation below 6% by 2027 and aim to reverse a 3% slide in paying advertising locations 1.

AI-native Yelp optimization services are emerging as a new business opportunity

– Discovery now leans on AI summaries across Popular Offerings, Menu Vision, and Assistant 2. Agencies and SaaS firms can build services that help businesses influence these surfaces.
– Opportunities include menu digitization and structuring, since Menu Vision leans on text recognition with photo databases 2. Teams can automate photo uploads with captions to feed that surface. The company has not published guidance on how Popular Offerings ranks items or which inputs matter most, so operators should watch API docs and public statements for clues 2.