Zero results search pages: the most honest page in your store is wasting the visit

Three causes account for nearly all of it. Vocabulary mismatch: the shopper writes trainers, the catalogue says sneakers, she writes settee, you wrote sofa. Spelling and phrasing: one transposed letter defeats an exact matcher that a human would not even notice. And genuine absence: the product or the size is not stocked, which is not a search failure at all but a range fact the page should admit and route around.

Split your zero results log by those three causes before buying anything. Synonym and stemming fixes are configuration, costing an afternoon. Absence is a buying conversation. Vendors sell search platforms on the whole problem, when most stores are looking at a synonym table nobody has maintained since launch and a matcher with the tolerance of a spelling bee.

Why searches return nothing

Never a blank and an apology. The page should do what a good shop assistant does with a no: name the closest thing. Relax the query in steps, drop the adjective, broaden the category, correct the spelling visibly so the shopper can object, and present those results labelled as close matches rather than pretending they are the answer. A shopper shown honest near misses browses. A shopper shown a void leaves, usually to a competitor whose search is dumber but less empty.

Below the near matches, earn the visit back: the category the query belongs to, the bestsellers in it, and a route to be told when the exact thing returns if absence was the cause. That waitlist signup is the only good outcome a genuine zero can produce, and most stores never offer it because the page was built as an error state rather than a selling surface.

What to show when there is no match

Read monthly, zero results queries are the cleanest demand data a store owns: unmet wants, stated in the shoppers own words, with counts. Recurring queries for a product you do not stock are a buying lead with the market research already done. Recurring queries for a product you do stock are a catalogue defect: the title, tags, or attributes never use the words shoppers use, and the product is invisible to the exact people asking for it.

Act on both in the same meeting. Add the missing synonyms and the shoppers own phrasing to product data, and put the persistent absences in front of whoever owns the range with the query counts attached. Search reports that stay inside the ecommerce team optimise the matcher. The money is in the two fixes that leave the team: better product words and a range that includes what people keep requesting.

The query log is a range document

The zero rate alone invites the wrong win: loosen the matcher until everything returns something and the rate collapses while relevance quietly rots. The metric that matters is recovery: of the shoppers who hit a zero or a relaxed result set, how many clicked, and how many bought, compared with a normal search session. A page that recovers a third of those visits at half the normal conversion is doing heavy commercial work.

Track the spelling correction acceptance too. If shoppers routinely reject the did you mean and retype, your corrector is confidently fixing words that were not wrong, usually brand names and technical terms from your own catalogue. Whitelist your own vocabulary first. It is embarrassing, and common, for a store search to correct its own product names into something else.

Measure recovery, not just zeros

Keep reading: How do default sort orders on sale pages bury the products that actually convert?, How should stores merchandise back in stock items without training shoppers to wait?.