From Search to Discovery Commerce

Two different jobs

Search and discovery do different jobs. Search retrieves: it takes a known intent and returns matching results as efficiently as possible. Discovery surfaces: it takes an unformed or latent interest and presents things the shopper did not ask for but might want. One is about precision against a query; the other is about relevance against a person.

Most real journeys use both. A shopper might be inspired in a feed, then search to confirm a price or find their size. The mistake is treating search as the whole system when it is increasingly the closing step of a journey that began somewhere less structured.

The two jobs also reward different kinds of investment. Search rewards a clean index, fast retrieval and precise relevance scoring against explicit terms. Discovery rewards behavioural modelling, rich visual content and the ability to infer what a person might like from what they have engaged with. A business that treats these as the same capability tends to under-serve one of them — building a fast search box that surfaces nothing unexpected, or a lively feed that collapses the moment a shopper wants a specific answer. Recognising them as distinct is the first step toward serving both well.

Why the shift is happening

Several forces push commerce toward discovery. Catalogues have grown too large to browse and too varied to query precisely — no shopper can type their way through millions of products. Feeds and recommendation systems have become the default way people consume everything else, so they naturally became a way to shop. And mobile, visual-first behaviour rewards surfaces you scroll over boxes you fill in.

Fashion accelerates all of this because it is bought on taste, not specification. A shopper often cannot articulate what they want in words, which is exactly the situation where a query fails and a well-tuned feed succeeds. That is why the category has been an early and clear example of the broader move, as our social shopping analysis shows.

The shift is reinforced by a simple economics of attention. Query-driven demand is capped by how often people think to search; discovery-driven demand is capped only by how much time people spend scrolling, which is far larger. A business that depends solely on being found is fishing in a smaller pond than one that can also form demand in the feed. That asymmetry is a large part of why platforms have invested so heavily in recommendation surfaces, and why merchants who once optimised only for the query are being pulled toward discovery whether they intended to or not.

What changes for merchants

For merchants, the shift reframes visibility. In a search-led world, being findable for the right keywords was close to sufficient. In a discovery-led world, a product also has to be surfaceable — good enough imagery, clean attributes, and signals that recommendation systems can act on. Optimising a title for a query is no longer the whole game.

None of this removes the need for clean product data — if anything it raises it, because a compelling surfaced product still fails if the underlying offer is stale. That foundation is the same one described in product-discovery platforms.

What changes for platforms

For platforms, the shift changes what a good experience even means. A search engine is judged on precision and speed against a query. A discovery platform is judged on how often it surfaces something the shopper is glad to have seen without asking. That is a harder, fuzzier target — it depends on modelling taste, not matching strings — and it changes what the platform must invest in.

It also changes the data requirements. Discovery leans on behavioural signals, rich attributes and fresh availability far more heavily than a keyword index does. The distinction between the two models is worth holding clearly, which is why we separate the search engine and the recommendation-led surface rather than treating them as one thing.

Search is not going away

It would be a mistake to read this as the end of search. Shoppers with a specific intent still want a fast, precise query, and verification — checking a price, confirming a size — remains a search behaviour. The point is not replacement but rebalancing: discovery now owns a larger share of where demand forms, and search increasingly plays a supporting, closing role.

The businesses that adapt best treat the two as a system rather than rivals. They let discovery form demand and let search resolve it, and they invest in the shared foundation — clean, fresh, well-matched product data — that both depend on. That systemic view is the through-line of this library and the frontier we track in AI fashion discovery.