How to use Search Page Optimization
Shoppers who use on-site search convert 2-4x the site average. This guide covers how to optimize the search results page — relevance, synonyms, merchandising, and the metrics that matter.
Search Page Optimization
The practice of tuning the on-site search results page to convert high-intent searchers — through better relevance, synonyms, typo handling, and merchandising.
Search Page Optimization is the work of making the internal search results page on your store actually convert. That means returning relevant products for the query, handling typos and synonyms, surfacing in-stock and high-margin SKUs first, and giving the shopper enough filters and sort controls to narrow down without bouncing.
It sits inside the broader discipline of page optimization, but it deserves its own playbook because the audience is different. A shopper who types into your search bar has already declared intent — they know roughly what they want. Getting them from query to PDP in two clicks is one of the highest-leverage CRO projects on the site.
Most stores treat search as a feature they shipped years ago and never touched again. The default Shopify or WooCommerce search returns title matches, sorts by something arbitrary, and gives up on a typo. That's leaving real money on the table — searchers are typically the 2-4x converting cohort on any e-commerce site.
The good news: search is one of the few CRO surfaces where a single well-scoped project can move sitewide revenue 3-8%. The bad news: it's also one of the most ignored. This guide walks through the four levers — relevance, merchandising, measurement, and edge cases — with concrete numbers and what to ship first.
Lever 1: Relevance and query understanding
Relevance is the floor. If a shopper searches "linen shirt" and the first row is a linen tablecloth, nothing else you do matters. The work splits into three parts: matching, typo tolerance, and synonyms.
Matching means searching across title, description, tags, and ideally attribute fields like color and material. Most native platform search only hits the title, which is why "red dress" and "crimson dress" return wildly different result sets even when the same SKU should match both.
Typo tolerance and synonyms are where the wins compound. Around 10-15% of search queries contain a typo, and an empty-results page on a typo converts at roughly zero. Map your top 200 queries against your synonym dictionary monthly — "sneakers/trainers", "jumper/sweater", "pants/trousers" — and add the misses.
Zero-results pages are conversion killers
Any query that returns zero results is a 100% bounce risk. Audit your zero-results report weekly — usually 60-80% of the volume comes from 20-30 fixable queries (typos, plurals, brand spellings, discontinued SKUs). Adding a fallback ("No exact matches — try these popular items") recovers 15-25% of that traffic instead of losing it entirely.
Lever 2: Merchandising the results page
Once relevance is solid, the question shifts: of the 200 products that match "black boots", which 12 go in the top row? Pure relevance sorting ignores margin, stock depth, return rate, and conversion history. Smart merchandising blends all of them.
A typical scoring model weights relevance ~60%, recent CVR ~20%, stock availability ~10%, and margin or strategic boost ~10%. Tune the weights against revenue per search, not click-through — a high-CTR result that drives no checkout is worse than a slightly less clicky one that closes.
Conversion rate: search users vs. browse users
The pattern above holds across most apparel, beauty, and home stores: each layer of search engagement raises CVR. That's why merchandising the results page is so leveraged — you're optimizing the highest-converting moment of the funnel, not the lowest.
Lever 3: Measurement — the KPIs that matter
Most teams measure search wrong. They look at search usage (% of sessions that used search) and stop there. That tells you nothing about whether the page is good. The metrics that move revenue are search exit rate, click-through to PDP, and revenue per search.
Benchmark your current state, then set a 90-day target per metric. The table below shows the ranges we see across Shopify and WooCommerce stores in the €1M-€15M revenue band — use them as a sanity check, not a goal in themselves.
On-site search KPI benchmarks for mid-market e-commerce
| Metric | Underperforming | Typical | Best in class |
|---|---|---|---|
| Search usage rate | <5% | 8-15% | 20-30% |
| Search exit rate | >40% | 20-30% | <15% |
| Click-through to PDP | <35% | 45-60% | >70% |
| Zero-results rate | >10% | 3-7% | <2% |
| Revenue per search | <€0.50 | €1.20-€2.50 | >€4.00 |
| Searcher CVR uplift vs site avg | 1.2-1.5x | 2-3x | 3-4x+ |
Revenue per search is the metric to anchor on — it folds CTR, CVR, and AOV into one number and is directly comparable across A/B tests. If you can only watch one KPI, watch that one.
Lever 4: Edge cases and mobile UX
The edges are where conversions leak. Mobile autocomplete that hides behind the keyboard, filters that reset on back-button, sort dropdowns that don't persist — each of these costs 1-3% of search revenue and almost nobody tests for them.
Run a search journey heuristic review every quarter: open your site on a real mid-range Android phone, search a typo, apply two filters, hit back, hit a result, return to search. Note every friction point. Most teams find 8-15 fixable issues per audit, and the top three usually pay for the whole project.
Ship order: relevance first, merchandising second
The temptation is to start with the shiny merchandising rules ("boost high-margin SKUs!"). Don't. Fix zero-results and typo handling first — those are conversion floors. Then layer in synonyms, then merchandising weights, then personalization. Each layer compounds only if the previous one works.
Frequently asked questions
Stores starting from default platform search typically see 3-8% sitewide revenue lift from a 4-6 week project. Most of that comes from fixing zero-results and typo handling — the merchandising layer adds another 1-3% on top.
Not always. If your catalog is under ~2,000 SKUs and search usage is under 10%, native Shopify or WooCommerce search plus a synonym plugin can be enough. Above that, a dedicated tool pays for itself within a quarter — mostly through better typo handling and merchandising controls.
Page optimization is the parent discipline covering every template — PDP, PLP, cart, checkout. Search page optimization is the slice focused on the search results template specifically, where the user has expressed an explicit query. The audience is more intent-loaded so the levers and KPIs differ.
Most search tools expose this in a report. In GA4 you can pull it from the `view_search_results` event filtered to queries with zero result impressions, or instrument it directly with a custom event. Sort by volume — the top 20-30 queries usually account for 60-80% of zero-results traffic.
No. Block internal search results pages in robots.txt or with noindex. They generate near-infinite URL variations and dilute crawl budget. Google's guidelines explicitly call them out as low-quality for the index.
Monthly is a good cadence for most stores. Pull the top 200 queries, cross-reference against zero-results and low-CTR queries, and add new synonyms or product mappings. Seasonal launches and product renames need ad-hoc updates.
Indirectly. The search page itself shouldn't be indexed, but the queries shoppers type are a goldmine for SEO — they reveal exact-phrase demand you can use to refine product titles, category pages, and content. Treat your search logs as a free keyword research tool.
Test at the algorithm level (variant A = current relevance weights, variant B = new weights), not per-query. Use revenue per search as the primary metric and run for at least 2 full weeks to capture weekly seasonality. Most search tools have a built-in split-test mode.
8-15% of sessions is typical for mid-market apparel and home stores. Beauty and electronics tend to skew higher (15-25%) because shoppers arrive with specific SKUs in mind. If you're below 5%, the search bar is probably hard to find — start by making it more prominent.
Both, in a hybrid. The left column shows suggested queries (text only, fast to scan), the right shows 3-6 product thumbnails with price. This hybrid pattern outperforms either alone — query-only feels like Google, product-only skips the long-tail. Test it on mobile first; that's where the layout breaks.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.