CRO · ANALYTICS · SEO · 16 DECEMBER 2025 · 8 MIN READ
Zero-result searches are a merchandising report
Customers typing words you have no answer for, in their own language, sorted by volume. Most stores never open it.
A zero-result search tells you a customer wanted something and your store had no answer, but "no answer" splits five ways and only one of them is a stock problem. The query may be for something you genuinely do not sell; something you sell under a different word; something you sell that is currently unavailable and hidden; a question rather than a product; or a typo. Each has a different and mostly cheap fix — a synonym group, a redirect, a stock policy change, a content page — and reading the report is what tells you which one you have. Shopify surfaces it as "Searches with no results" in the search reports.
IN SHORT
- Zero-result searches are the only place customers describe what they want in their own words, unprompted and at volume.
- Shopify reports it as "Searches with no results", alongside "Searches by search query" and "Searches with no clicks".
- Predictive search interactions are excluded from Shopify’s search performance metrics, so the dropdown most customers actually use is not in the report you are reading.
- Search reporting can lag by up to 72 hours, so do not judge a change you made yesterday.
- "Searches with no clicks" is the harsher report of the two: the store returned results and the customer rejected all of them.
- Most entries deserve a synonym group or a redirect, not a purchase order.
Where it lives, and what it does not include
Shopify’s search reporting gives you four things worth knowing about: "Searches by search query", "Searches with no results", "Searches with no clicks", and "Search conversions over time". The Search & Discovery app shows a recent window in-app; the full reports with custom date ranges sit in Analytics.
Two limitations matter before you read a single row. Search data can be delayed by up to 72 hours, so this is not a report to check the morning after a change. And — this is the one that catches people — predictive search interactions are excluded from the search performance metrics. The dropdown that appears as customers type, which on most stores is where the majority of searching happens and ends, is not represented here.
That does not make the report useless. It makes it a sample: the queries from customers who typed a full term and pressed enter. Treat the volumes as relative rather than absolute, and do not build a business case on the total.
The five things a zero-result query means
Go down the list and put every row into one of these. The exercise takes an hour and it is the whole job.
- You do not sell it. The honest case, and rarer than people expect. Genuine demand for something adjacent to your range is a buying signal, and it is the only entry on this list that costs money to act on.
- You sell it, under a different word. The most common case by a distance. Your copy says "trainers", the customer typed "sneakers"; your copy says "duvet", they typed "comforter". This is a synonym group and takes two minutes.
- You sell it and it is hidden. Unavailable products default to appearing last in search rather than being hidden, but if someone has set them to hide, a temporarily out-of-stock line produces zero results for its own name. That is a customer who knew exactly what they wanted and was told it does not exist.
- It is not a product. "Returns", "delivery times", "size guide", "gift card balance". Customers use the search box as a help desk because it is the most visible input on the page. Every one of these is a page that should have been findable, and search is telling you it was not.
- It is noise. Typos, competitor names, pasted URLs, the occasional test query from your own team. Ignore, but read a few first — a typo appearing forty times is not noise, it is a spelling your customers believe is correct.
The report nobody reads is the more damning one
"Searches with no results" is a failure of matching. "Searches with no clicks" is a failure of merchandising, and it is worse.
A no-click search means the store found things, showed them, and the customer looked at the set and rejected all of it. Zero results at least tells the customer something clearly. A page of twenty-four products the customer does not want tells them your range is wrong, which may be untrue and is far harder to recover from.
The usual causes are dull and fixable: the right product is on page two; the top results are accessories rather than the item; the thumbnails do not show the attribute the customer searched for; the query was broad and there is no filter rail to narrow it. A no-click query is worth opening yourself and looking at, because the answer is almost always visible in the first screen.
Sorting the list so you act on the right end of it
The report arrives sorted by volume, which is the right first cut and the wrong only cut. Two piles are worth separating.
The head is a handful of terms with meaningful volume. These deserve individual attention: a synonym group, a boost, or a deliberate redirect to the collection that answers the query. A term with real volume returning nothing is a bug with a revenue number attached.
The tail is hundreds of terms with one or two searches each, and it is where people waste an afternoon. Do not fix the tail one row at a time. Read it for patterns instead: if forty different rows are all material names, your customers shop by material and you have no material filter. If they are all part numbers, predictive search is not looking at variants.sku — it searches title, product_type, variants.title and vendor by default, and SKU has to be requested explicitly. One structural change clears a hundred rows.
A single query with one search is not a signal. A hundred queries that rhyme are.
The fixes, cheapest first
Deliberately ordered by cost, because the expensive answer is almost never the right one and it is always the one suggested first in the meeting.
- Synonym groups. Match the words customers use to the words in your catalogue. Free, fast, and it clears more rows than anything else on this list.
- Send the query somewhere. A handful of head terms should land on a specific collection rather than a result set. Do not reach for Shopify’s URL redirects to do it — the documentation warns that URLs containing query strings "might not work as expected" with redirects, and a search is a query string. This is a rule in the search template or a feature of a search app. Worth the effort for the head, never for the tail.
- Change the out-of-stock policy. If unavailable products are hidden, consider showing them last instead. A customer who finds the product and a back-in-stock form is a retained customer; one who is told it does not exist is not.
- Write the page. For "returns", "delivery", "size guide" — the content exists, it just is not in the search index or the result types. Check whether pages are enabled as a result type at all.
- Fix the field configuration. SKUs, tags and body copy are opt-in for predictive search. One change, whole categories of failure removed.
- Add a filter. If the tail says customers shop by an attribute you do not expose, expose it. Metafield-backed filters are the ones worth the effort.
- Buy stock. Last, and only when the first six have been ruled out and the demand is real and repeated.
Design the empty page, because you will keep having empty pages
However good the merchandising gets, some searches will return nothing, and the page that appears is usually the least-designed page on the store: a heading, a sentence, and nothing to do next.
Three things make it earn its place. Echo the query back so the customer can see their own typo. Offer a route onward that is not the home page — best sellers, the top-level collections, or a genuinely relevant set if you can produce one. And give them a way to tell you, because a customer who searched for something you do not stock is the cheapest research you will ever get.
The other failure state deserves the same attention: one or two results in a grid designed for four columns reads as "we do not have this", even though the customer found the product. A results page that changes layout when it is nearly empty is a small piece of work with a direct effect.
Make it a habit rather than a project
This is an hour a month, and it works because it compounds: the synonyms you add stay added, and the same rows stop appearing. The stores where search works are not the ones that bought the best search app. They are the ones where somebody reads this report regularly and the list gets shorter.
Put a name against it. Not a team — a person, with a recurring hour, in the same way somebody already owns the home page. Unowned recurring work does not happen, and this is the most easily-postponed task in ecommerce because nothing visibly breaks when it is skipped.
Measure it with what Shopify already reports — click rate and purchase rate for search, and "Search conversions over time" — and give changes a fortnight before judging them, since per-term volumes are low enough that a week is mostly noise. If you would rather it ran as part of a structured programme with the tests and the reporting attached, that is what our [conversion rate optimisation](/services/optimize/shopify-conversion-rate-optimization) work covers. The first month of it, though, is genuinely just reading a list.
Questions this raises
Where do I find zero-result searches in Shopify?
In the search reports as "Searches with no results", alongside "Searches by search query" and "Searches with no clicks". The Search & Discovery app shows a recent window inside the app, and Analytics holds the full reports with custom date ranges. Note that the data can be delayed by up to 72 hours.
Does the report include searches from the search dropdown?
No. Predictive search interactions are excluded from Shopify’s search performance metrics, so what you are reading is the subset of customers who typed a full query and submitted it. Since the dropdown is where a lot of searching begins and ends, treat the figures as a relative signal rather than a count of everyone who searched.
Should a zero-result search make us buy more stock?
Rarely, and never as the first response. Most zero-result queries are vocabulary mismatches that a synonym group fixes, questions that belong on a content page, or products that exist but are hidden because of an out-of-stock setting. Work through those first. A genuine, repeated demand for something adjacent to your range is a real buying signal, but it is the last conclusion to reach, not the first.
What is the difference between no results and no clicks?
No results means nothing matched. No clicks means the store returned results and the customer rejected every one of them. The second is usually the more serious signal, because the customer has now formed a view about your range rather than about your search box, and the cause is normally visible if you run the query yourself and look at the first screen.
How many zero-result searches is too many?
There is no benchmark worth quoting, because the honest comparison is with your own store last month. What matters is the shape: a head of high-volume terms returning nothing is a problem to fix this week; a long tail of one-off queries is normal on any store and should be read for patterns rather than cleared row by row.
How often should we review search reports?
Monthly is enough for most stores, with a check during a peak trading period when customers are searching for gift terms and seasonal language that does not appear in your catalogue the rest of the year. The work compounds, so the review gets shorter each time — which is how you know it is working.
NEXT STEP
Free store audit
A senior Shopify engineer reviews your storefront, theme performance and checkout, then sends a prioritised list of fixes.