Marketplaces Prioritize Supply Over Search
Andrew Yates, CEO of Promoted.ai, on when marketplaces should layer on ads
The real point is that most marketplaces win by owning supply, demand, and operations, not by inventing frontier search science. For a furniture marketplace, a better ranking model matters only if it helps a buyer find a couch faster, raises seller sales, and lifts ARPU. Once search becomes good enough, the harder and more valuable work is inventory quality, fulfillment, trust, pricing, and seller economics.
-
The underlying job is not just search. It is measuring the full path from impression to purchase, then using that data to rank listings, personalize feeds, and decide when paid placement helps or hurts the marketplace. That is why Promoted frames ads as one configuration of a broader optimization system.
-
The exception is companies where discovery is the product. Netflix is the clean example. If the core promise is helping users find the next thing they will love from a huge catalog, recommendation quality is the moat. In a marketplace, the moat is usually better supply, lower friction, or stronger merchant economics.
-
This is also why marketplaces often buy rather than build. Wayfair already had a meaningful ads business, but even there the immediate opportunity was improving monetization on existing surfaces, not turning the company into a research lab for information retrieval. The economic test is simple, does the system raise revenue and conversion in A/B tests.
Over time, search and recommendation infrastructure is likely to become more modular, like cloud infrastructure did. The winners in marketplaces will use outside systems for ranking, targeting, and measurement, then spend internal talent on the pieces only they can own, better supply density, better logistics, and stronger repeat purchase loops.