Search results merchandising: sorting rules that help shoppers find the right product

Updated October 21, 2026 · DeftMerch answers

Short answer: Customers who use site search have already decided what they want; the store's job is simply not to fumble the answer. Yet internal search is often left on default settings while teams polish category pages for hours. The result is a strange store where browsing is curated and searching, the highest intent behaviour on the site, returns results in warehouse order.

Relevance first, then commercial sense

A good default sort blends textual relevance with signals the shopper cannot see: stock health, margin, return rate, and how well the product converts for that query. Pure best seller order fails on narrow queries, where the third most relevant product outsells its way above the exact match. Pure relevance fails on broad queries, where everything matches and nothing guides the eye.

Synonyms and language do the quiet work underneath. Shoppers search in their own words: trainers, sneakers, couches, sofas, plus misspellings of the store's own brand names. A synonym map built from real search logs, reviewed monthly, is the highest return hour a merchandising team can spend.

Boosts, buries, and pinned results

Manual controls earn their place when they are few and purposeful. Pinning the hero product for a campaign query, burying a discontinued line that still ranks, or boosting in stock sizes over broken ranges are legitimate merchandising. A results page carrying forty manual rules nobody remembers setting is a maintenance debt that produces stranger results every month.

Every boost should have an owner and an expiry date. Campaigns end, stock situations change, and the pin that made sense in November quietly taxes March. Stores that review their search rules on the same calendar as their promotions find that most rules can retire on schedule.

The zero results page is a sales page too

A meaningful share of searches find nothing, usually because of wording rather than absence. The store often stocks the product under another name. A dead end page that just says no results wastes the highest intent click on the site; a page that offers the closest category, spelling suggestions, and popular alternatives recovers a real share of those shoppers.

Zero result queries are also the best product research a store gets for free. Repeated searches for items the store does not carry are demand signals in the customers' own words, and repeated searches for items it does carry under another name are naming problems the store can fix in titles and synonyms.

Measure search like a channel

Search deserves its own dashboard: searches per session, results click through, search exit rate, and revenue per search compared with browse. When a change to sorting rules ships, those numbers show within days whether it helped. Without them, search debates run on opinion and the loudest anecdote wins.

Segment by device while you are at it. Mobile searchers type less, tap suggestions more, and abandon faster. Autocomplete quality, which barely matters on desktop, can be the whole mobile search experience, and it is built from the same query data the merchandising team already owns.

Keep reading: How should stores merchandise preorders on product and category pages?, How should stores merchandise limited edition products without discounting them?, and How should stores merchandise a new arrivals section that actually sells? cover the nearby cases in this series.