Start learning free

Case study · ecommerce · conversion

Nothing found: How zero-result searches send ready buyers away

An interactive pattern case study: why site searches return nothing for products the shop sells, how to find them, and what a helpful no-results page does.

WXM 1 · Wonder (Look)WXM 2 · Architect (Plan)WXM 3 · Validate (Test)WXM 4 · Expand (Grow)
Pattern analysis
By Phil the CRO Guy

Phil the CRO Guy

Your CRO guide at Mentorele

Phil walks you through every lesson and case study. He asks the questions you might be too polite to ask, and he's happy to admit when a test didn't win. Every lesson he presents is checked by a real CRO expert.

Learn more about MentoreleMeet Phil

Expert-reviewed by Filip Doneski

Filip Doneski

CRO Expert and Strategist · Partner and Lead CRO Strategist at Wavesy

10+ years in conversion optimization and experimentation for DTC, retail and SaaS. Reviews every Mentorele lesson for accuracy before it’s published.

A/B testingConversion rate optimizationDigital psychology and persuasion

Filip’s expertise at WavesyLinkedIn profile

Updated

Slide 1 of 9

The pattern: searchers who can't find what you sell

People who use site search usually know what they want. That makes a dead-end search expensive: a ready buyer told "no results" for something the shop actually sells.

Phil uses a composite shop, Brightside Home, which sells lighting and furniture.

A search box is a customer telling you exactly what they want, in their own words. Best data on the site.

Slide 2 of 9

WXM 1 · Wonder (Look)

Wonder: finding zero-result searches

Site search can be tracked with GA4's enhanced measurement. [Source](#src-ga4-enhanced-measurement) The team adds a result-count parameter so zero-result searches can be filtered.

  1. A noticeable share of searches return zero results.
  2. Top zero-result terms: "lampshade" (shop uses 'shade'), "bedside lamp" ('table lamp'), "sofa" ('couch'), and typos like "pendent".
  3. Sessions with a zero-result search often end on that page.
Top zero-result searches (composite example) Original wireframe, not a screenshot.
  1. We sell these. Most zero-result searches are for products in stock, described in different words.

Slide 3 of 9

WXM 1 · Wonder (Look)

Wonder: the no-results page today

A blank page saying 0 results.

Control (A)
Search results page with no matches, current version Original wireframe, not a screenshot.
  1. Dead end. No suggestion, no alternative, no way forward except leaving.

Slide 4 of 9

WXM 2 · Architect (Plan)

Architect: two fixes, two hypotheses

The research points to two separate problems:

  1. The search engine: synonyms and typos. Adding synonyms ('lampshade' → shades) and typo tolerance is a fix, not a test. Ship it.
  2. The page: some searches will always fail. Because sessions often end on the zero-result page, we believe a helpful no-results page (query shown, corrections, popular categories, help link) will increase orders per searching visitor, because it gives a dead-end search a way forward.

Not everything needs an A/B test. If 'sofa' finds nothing in a sofa shop, just fix it.

Slide 5 of 9

WXM 2 · Architect (Plan)

Architect: how does the page change score on PXL?

Noticeable and adds elements, backed by analytics, quick to build. But it only affects visitors whose search fails, so it isn't high-traffic.

PXL score: 9 / 14

Is the change above the fold?Yes1
Is it noticeable within 5 seconds?Yes2
Does it add or remove an element?Yes2
Is it designed to increase user motivation?No0
Does it run on high-traffic page(s)?No0
Does it address an issue found in user testing?No0
Does it address an issue found in qualitative feedback (surveys, polls, interviews)?No0
Does it address insights found in digital analytics?Yes1
Is it supported by mouse-tracking heatmaps or eye tracking?No0
Ease of implementationUnder 4 hours3

PXL by CXL. A high score ranks an idea; it does not predict a win.

Slide 6 of 9

WXM 3 · Validate (Test)

Validate: the variation

The page says clearly that nothing matched, repeats the query, and offers ways forward, in line with usability guidance for no-results pages. [Source](#src-nng-no-results)

Control (A)
A: blank 0 results Original wireframe, not a screenshot.
Variation (B)
B: a helpful dead end Original wireframe, not a screenshot.
  1. A way forward. Correction, categories and help instead of a blank page.

Slide 7 of 9

WXM 3 · Validate (Test)

Validate: the test plan, written before launch

Fixed in advance:

  1. Primary metric: orders per visitor who sees a zero-result page.
  2. Secondary: clicks on suggestions, exits from the page.
  3. Sample: this audience is small; check the sample size first. If it would take months, ship it as a best-practice fix and monitor.
  4. Decision rule: ship if orders improve, or if the test can't reach its sample and nothing gets worse.

Slide 8 of 9

WXM 4 · Expand (Grow)

Expand: what the pattern usually leads to

This is a pattern, not a report of one company's test. Typical next steps:

  1. Review zero-result searches every month and add synonyms.
  2. Feed search terms into product naming and category labels: customers are telling you their words.
  3. Look at searches with results but no clicks. Those are the next dead ends.

Slide 9 of 9

Your turn

Search your own shop for three products using words a customer might use rather than your product names. Then check whether site search is tracked with a result count. Read GA4 for CRO for the setup.

What this pattern teaches

  • Wonder used search terms as voice-of-customer data.
  • Architect separated a straightforward fix (synonyms) from a testable change (the page).
  • Validate checked whether the audience was big enough to test at all.
  • Expand made zero-result review a monthly habit.

Brightside Home is a composite store used to illustrate a common pattern.

Frequently asked questions

What should a no-results search page show?

Make it clear nothing matched, repeat the query, offer spelling corrections and related searches, show popular categories or products, and give a way to get help. Never show a blank page.

Is this based on a real company?

No. It describes a pattern seen across many online shops, illustrated with a composite store. It doesn't attribute results to any company.

Sources (4)

  1. 3 Guidelines for Search Engine "No Results" Pages. Nielsen Norman Group. Accessed 14 November 2026.
  2. Enhanced measurement events. Google Analytics Help. Accessed 14 November 2026.
  3. WXM: the Wave Experimentation Model. Wavesy. Accessed 4 November 2026.
  4. PXL: A Better Way to Prioritize Your A/B Tests. CXL. Accessed 4 November 2026.

Learn the principles behind this case

Search lessons, guides, case studies and templates. Press Esc to close.