No-Results Page UX: Design for Recovery

A shopper searches, the grid comes back empty, and the page just sits there — usually a single line of gray text and nothing to click. That page is where a lot of stores lose the sale, not because the shopper didn't want anything, but because the page gave them nothing to do next.

This is about the page itself: what should actually be on screen the moment a search returns zero results, separate from the backend work of recovering the query in the first place.

What Is a No-Results Page?

A no-results page is what a shopper lands on after a zero-result search — a query that matched nothing in the catalog, even if relevant products exist. Left as the platform default, it's usually a static message like "No products found" with no path forward. Zero-result searches explains why these queries happen; this article is about what the page shows once one does.

The distinction matters because a store can fix the matching problem — synonym expansion, typo tolerance, semantic fallback — and still ship a weak no-results page, or the reverse: a well-designed page with nothing useful behind it to show. Both halves have to work together.

Why the Default "No Results Found" Message Fails

The stock empty state fails for a few consistent reasons:

• It's a dead end, not a next step. A generic apology gives the shopper nothing to click except the browser back button.
• It doesn't acknowledge what they typed. Without echoing the query, the shopper can't tell if the store misread it, or if the product genuinely isn't carried.
• It offers no fallback content. No related products, no categories, no suggestions — just white space.
• The search bar itself gets buried or hidden, so re-searching means starting the interaction over.

The result is search abandonment — the shopper leaves the site rather than tries again. Since a search query is an explicit statement of intent, that's a more expensive exit than most other bounces on a store.

What to Put on a No-Results Page

Echo the query back

Show the exact term the shopper searched, in plain text: "No results for 'trainers'." It confirms the system read the input correctly and sets up any correction or suggestion that follows.

Offer a corrected or related query

If the query is a likely typo or synonym mismatch, suggest the closer term directly — "Did you mean 'sneakers'?" — rather than making the shopper retype it. This is the presentation layer on top of query reformulation: the backend proposes the correction, the page has to make it a single click.

Show fallback products instead of a blank grid

An empty results grid is the single biggest miss. Fall back to best-sellers, trending items, or products from the closest matching category rather than showing nothing. It's not the same as answering the query, but it keeps the shopper browsing instead of leaving.

Surface related categories

If "leather backpack" returns nothing, a link into the Bags category is a smaller ask than a new search. Category and facet links work as a manual fallback when there's no confident product match to show.

Keep the search bar visible and focused

The shopper's next move is often to try a different phrase. Keep the search input in view, ideally already focused, instead of requiring a scroll or a click back to the header.

Give a low-friction way to ask for help

For a genuinely unavailable product, a visible link to support or a contact form beats a dead end — especially for high-consideration or B2B catalogs where a human answer closes more sales than a product suggestion would.

Design Patterns by Store Type

• Fashion: Query-echo plus a "Did you mean" correction for synonym gaps ("trainers" → "sneakers"), backed by trending items in the closest category.
• Electronics: Near-miss SKU or model-number matching matters more than category fallback — a shopper who mistypes a model number wants the closest real match, not a generic "top sellers" list.
• Beauty: Ingredient and use-case queries ("moisturizer for sensitive skin") rarely match a product title directly, so fallback needs to work off attributes, not just keywords.
• Grocery: High query volume and fast-changing inventory mean the fallback list has to reflect current stock — a suggested product that's out of stock defeats the point.
• Multilingual stores: The empty state, the correction copy, and the fallback logic all need to work in every storefront language, not just the default one — a synonym or typo fix built for English says nothing about how the same gap shows up in French or German.

Common No-Results Page Mistakes

• Apology-only pages with no fallback content or next step.
• Auto-redirecting to the homepage, which discards the query entirely and forces the shopper to start over.
• Hiding or shrinking the search bar right when the shopper most needs to try again.
• Showing the same generic fallback regardless of query — a "check out our best-sellers" block that never changes gives no sense that the search was heard.
• No distinction between a true zero-result query and a weak-result one — a page that returns three barely-relevant items still needs the same recovery thinking as a page that returns none.

How Page Design Relates to Backend Recovery

The page can only show what the backend finds. Query rewriting, synonym expansion, typo tolerance, and semantic fallback — covered in Search Recovery: Tactics for Fixing Failed Searches — decide what candidate products or corrected queries exist for a given failed search. Vocabulary mismatches specifically are common enough to warrant their own fix; see ecommerce search synonyms for how that piece works. The page design decides how confidently and clearly those candidates get presented, and what to fall back to when there's genuinely nothing close.

Get either half wrong and the other doesn't matter: strong backend recovery wasted on a page that still says "No results found" instead of showing what was found, or a well-designed page with nothing behind it to populate.

Measuring Whether Your No-Results Page Works

A few metrics tell you whether the redesign is actually helping, not just looking better:

• Exit rate on the no-results page. If shoppers who land there consistently leave the site next, the page still isn't giving them anywhere to go. See search exit rate.
• Re-search rate. How often a shopper tries a second query after landing on the page, versus giving up. A working correction or suggestion should lift this.
• Click-through on fallback products or categories. If the fallback content never gets clicked, it's not the right fallback for that query.
• Zero-result rate over time. Run your own numbers with the zero-result calculator — the page redesign and the backend matching fix should move this number together.

At Wine House, roughly 39% of searches were returning zero results before a recovery layer was added; the recovered queries went on to drive an estimated $5,700 in cart value in one month, about 26.5% of revenue in that period — see the Wine House case study. At Lisa Leonard, around 10% of zero-result queries are recovered monthly, contributing roughly +1% in revenue — see the Lisa Leonard case study. Both numbers depend on the recovery layer and the page actually surfacing what it finds.

Mobile Considerations

On mobile, screen space forces priority calls the desktop page can defer. Put the corrected-query suggestion and one row of fallback products above the fold — a shopper who has to scroll past an apology message before seeing anything useful has usually already left. Keep the search input reachable with the thumb rather than requiring a scroll back to a header bar.

The Bottom Line

A no-results page isn't a dead end by default — it's designed into one, usually by omission. Echo the query, offer a correction, show a fallback that isn't random, and keep the search bar in reach. None of it replaces fixing the underlying match; it's what makes that fix visible to the shopper who hit the gap.

Want to see what your own no-results page is losing? Book a demo or see how Search Saver recovers failed searches on the same page shoppers already landed on.

Frequently asked questions

What should a no-results page show instead of "No results found"?

At minimum: the query echoed back, a corrected or related query suggestion if one exists, fallback products (best-sellers or closest-category items), and a visible, focused search bar so the shopper can try again without extra clicks.

How is a no-results page different from search recovery?

Search recovery is the backend work — query rewriting, synonym expansion, typo tolerance, semantic fallback — that finds candidate matches for a failed query. The no-results page is where those candidates (or the lack of any) actually get shown to the shopper.

Does redesigning the no-results page require changing my main search bar?

No. The redesign is scoped to the empty state itself — what renders when a query returns zero results — and can ship without touching the search input, autocomplete, or the results page for successful searches.

What products should I show when nothing matches the query?

Best-sellers or trending items from the closest matching category are the safest default. In-stock, currently available products only — a fallback suggestion that's out of stock undermines the point of showing one.

Can a no-results page work well without semantic search?

Partially. A manual synonym list and typo correction can cover a meaningful share of failures without semantic matching, but descriptive or paraphrased queries ("something for a dinner party") need semantic or hybrid matching to produce a confident fallback rather than a generic one.

How do I know if my no-results page needs a redesign?

Check the exit rate on that specific page against your site's overall exit rate, and look at your zero-result rate with the zero-result calculator. A high exit rate on an already-elevated zero-result rate usually means the page isn't giving shoppers anywhere to go.