Ecommerce Site Search: What Shoppers Type, What Your Store Returns, And How To Close The Gap

A shopper who types into your search box has told you what they want, in their own words. Ecommerce site search is the part of your store that answers them. It’s one lever in ecommerce conversion rate optimization, alongside product pages, checkout and email, and it’s the one where the shopper writes the request for you.

When search lets a shopper down, it does it in a few predictable ways. A query returns nothing. A query returns products, but the shopper’s word isn’t the word on your listing, so the right item sits low or doesn’t show. Or the filters on the results page don’t match the way shoppers sort your products in their heads.

Each of those problems shows up in data Shopify already collects, and most of them get fixed in your product data rather than in a new app. This post covers where to find that data, what to fix first, and when Shopify’s free Search & Discovery app stops being enough.

Baymard Institute, a UX research firm, combined category navigation and search in its usability testing of 19 major ecommerce sites (opens in new tab), including Amazon, Best Buy, IKEA, REI and Zappos. Participants either couldn’t find what they were looking for or got frustrated enough to give up 31% of the time, and 65% of the time they needed more than one search attempt. The sessions turned up more than 750 usability issues specific to ecommerce search, which Baymard distilled into 49 guidelines.

Those were large retailers. A Shopify store runs on Shopify’s built-in search unless it installs something else, and the person deciding how that search behaves is usually the store owner.

The failures fall into three groups. A zero-result search is a query that returns nothing at all, because you don’t carry the product, your catalog uses a different word, or the misspelling is too far off to correct. A no-click search returns results, but nothing the shopper wanted to open. A filter mismatch happens on the results page, when the shopper wants to narrow by fit, material or use and your filters only offer vendor and price.

Shopify records what shoppers type into your ecommerce site search and what happens next, and the reports are free. In Analytics > Reports you’ll find four search reports (opens in new tab): Searches by search query, Searches with no results, Searches with no clicks, and Search conversions over time. The first three list each search term with its number of searches. The fourth shows, day by day, the sessions that searched and the share of them that clicked a result, added to cart and purchased.

The Search & Discovery app puts a summary of the same data (opens in new tab) on one screen: click rate, purchase rate, top search terms, searches with no results and searches with no clicks. That summary always covers the last 30 days. For any other date range, go back to Analytics > Reports.

Three details decide whether you can trust the numbers. The search reports run up to 72 hours behind (opens in new tab), so yesterday’s searches may not be there yet. Shopify filters identified bots out automatically, and search reports don’t let you change that filter. And searches are only tracked if your theme uses the url property returned from search.results, which every Shopify-built theme does.

If you’re on a custom or third-party theme, confirm the reports show searches at all before you read anything into them. Run a few searches on your own storefront and look for those terms in Searches by search query once the delay has passed.

The reports also cover the search results page only. A shopper who types into the predictive search dropdown (opens in new tab) and clicks a product straight from it isn’t counted, so the reports undercount search on a store where most shoppers pick from the dropdown.

Shopify changed how it measures sessions between September 21 and 23, 2026 (opens in new tab), and results from the sessions and searches datasets can shift across that date. A before-and-after comparison of search data that spans it is comparing two measurement methods, so keep both periods on the same side of the change.

If you’d rather have this data read alongside the rest of your funnel, a CRO audit maps where visitors drop off from landing pages through checkout and ranks each fix by the revenue it’s projected to recover.

Shopify’s search form sends the query to /search in a parameter called q (opens in new tab). GA4’s enhanced measurement fires a view_search_results event (opens in new tab) whenever a page loads with one of five default query parameters in its URL: q, s, search, query or keyword. So with enhanced measurement on and your GA4 tag firing on the search page, GA4 records Shopify storefront searches with no extra setup.

The term the shopper typed lands in Search term (opens in new tab), a built-in GA4 dimension, so you don’t need to create a custom dimension for it. You can also set the event to look for other URL query parameters, which is the setting to change if a search app moves the query somewhere other than q.

Use GA4 for what it’s good at here: tying a search to the traffic source, the device and the rest of the session. Use Shopify’s Searches with no results report for the terms that came back empty.

Reports list what shoppers typed, not how they got there. Session recordings show a shopper typing, reading the dropdown, deleting a word and trying again, and they cover the predictive dropdown that Shopify’s search reports leave out. Pull recordings of sessions that reached /search, then watch the ones that ended without a product view.

Start with Searches with no results, sorted by number of searches. Each term on that list falls into one of three buckets: you carry the product under a different word, you carry a close substitute, or you don’t carry anything like it. The third bucket is a note for whoever does your buying, not a search fix.

Next, read Searches with no clicks. These terms returned results but got no clicks, which usually means the right product is there but buried, or the results match the words and miss the intent. Search a few of those terms yourself and look at what comes back.

Then compare the purchase rate in Search conversions over time with your store’s overall conversion rate for the same dates. Read that gap carefully. Shoppers who search often arrive knowing what they want, so they can convert above your average even when search is mediocre, and the gap alone doesn’t prove search is working. The comparison that tells you something is your search purchase rate against itself, before and after a fix.

Changes to the results page layout or the predictive dropdown are different, because they change the experience for every searcher at once. Run those through A/B testing instead of a before-and-after read, and size the test on the sessions that actually search, which is always fewer than your total traffic.

Every Shopify online store gets predictive search and typo tolerance by default (opens in new tab), and results draw from products, blog posts and pages. Typo tolerance (opens in new tab) catches a word that’s one letter off from a word in your store, or that has two letters swapped. Search also matches other forms of a word, such as singular and plural, in each language it supports.

Semantic understanding is the newer layer. It’s included in online store search (opens in new tab) and uses your product title, description, product type, product category and tags to work out what a shopper means, not only which words they typed. A search for christmas party shoes can return red pumps whose product data contains none of those words. Product images aren’t part of it, so anything a photo shows and the text doesn’t say is invisible to search.

Ranking is the part you don’t control. You can’t change the order results appear in (opens in new tab), and the levers you do have are your product data and the Search & Discovery app. Typo tolerance and semantic understanding both switch off when a shopper uses search syntax. Out-of-stock products sit at the end of results by default, and the app lets you show them, hide them or keep them last (opens in new tab).

Shopify’s free Search & Discovery app (opens in new tab) adds filters, product boosts, product recommendations and search analytics on top of that. It only works with Shopify Network Intelligence (opens in new tab) turned on in your privacy settings, so check that first if the app’s settings don’t take effect.

Synonym groups let you tell search that two words mean the same thing. The app’s App Store listing and Shopify’s Search & Discovery page (opens in new tab) still advertise them. Merchants began reporting in mid-September 2026 that the synonyms option had disappeared from the app (opens in new tab).

Open Search & Discovery in your own admin and check whether a Synonyms section is still there. If it isn’t, the fix for a vocabulary gap is the product data that semantic understanding reads. When shoppers search a term your catalog never uses, such as a nickname for a product line, add that word to the product’s tags or description where it fits.

Because you can’t reorder results directly, most ecommerce site search fixes on Shopify are product data fixes. Take each term from Searches with no results and Searches with no clicks, find the product the shopper should have reached, and check whether the shopper’s word appears in its title, description, product type, category or tags. Semantic understanding reads those fields, so adding the word where it reads naturally is the first fix to try.

Skip misspellings and plurals. Typo tolerance already handles one-letter errors and swapped letters, and search already matches singular and plural forms, so your time goes further on words the catalog doesn’t use at all.

Most ecommerce site search best practices come back to the same habit: describe each product in the words shoppers use for it, including what it is, what it does and what it’s for. A listing titled with a style name alone gives search nothing to match a shopper who types what the product is.

A product boost (opens in new tab) pushes a product higher for the search terms you choose, with a maximum of 10 search terms per boosted product. Boosts apply only to products available for sale, they don’t apply when the shopper uses search syntax, and they don’t affect third-party search engines. Shopify recommends boosting one or a few products per term, because boosting many can push other relevant products lower.

Boosts sync from metafield edits, including edits made in the Bulk Editor or Shopify Flow, and changes reach search results the same day. Misspellings and singular or plural variants of a term don’t need boosts of their own.

A boost is a merchandising rule that applies to everyone who types the same term. Showing different results to different shoppers is personalization on Shopify, which is a separate set of tools with its own traffic requirements.

Standard filters (opens in new tab) for Availability, Category, Price, Product type, Tags and Vendor are available to every store. Custom filters come from a product option, a product metafield, a category metafield, a variant metafield or a standard product attribute. If shoppers narrow by fit, material or activity and that information lives only in your descriptions, there’s nothing for a filter to read. That’s why structured metafields come before the filter setup.

The free app’s limits are fixed. A store can have up to 25 filters (opens in new tab), and each filter source can be used only once. A filter displays at most 100 values, a filter group holds at most 200 unique values, and a store can have at most 1,000 filter groups across its filter settings.

Catalog size brings its own limits. Collections with more than 5,000 products don’t display filters (opens in new tab), filters are hidden when a search returns more than 100,000 products, and the filters and value counts on a search results page are based on the first 1,000 results. A store with a 6,000-product collection can’t filter that collection with Shopify’s app.

Two more limits apply to stores selling in several languages or currencies. The tag filter only displays to shoppers browsing in your default language, and the price filter doesn’t display in currencies other than your default. Your theme also has to support filters: you can create them in the app on a theme that doesn’t, but customers won’t see them.

Search gets the shopper to a product page, and the page has to finish the sale. The app’s product recommendations (opens in new tab) come in two kinds. Related products are similar to the one being viewed and are generated automatically, from purchase history, from the product description on English storefronts, or from related collections as a fallback. Complementary products are ones often bought alongside it, and they need a theme with a complementary products section.

You can’t edit the generated recommendations, but you can add your own: up to 10 related and up to 10 complementary products per product. Only one complementary block can sit on a product page at a time. What the shopper sees once they land is its own set of decisions, and our product page tests walk through five of them, from above-the-fold clarity to bundles and cross-sells.

Shopify’s free app is enough when your ecommerce site search problems come down to words missing from product data, a handful of boosts and a filter set that fits inside 25 filters. Look at a paid app when you hit a wall Shopify doesn’t let you move: you need control over the order results appear in, more filters than the free app allows, filters on a collection over 5,000 products, or synonyms you build by hand. Ask each vendor whether its app removes the specific limit you’ve hit before you compare anything else.

The paid apps don’t price on the same basis, so the cheapest app for a 500-product store may not be the cheapest at 50,000 products or at heavy search volume. The table shows what each price scales with.

Searchanise and Globo.io’s Smart Product Filter & Search both price by product count (opens in new tab), so the bill rises with the catalog. Searchanise’s free plan covers development stores and stores with fewer than 25 products. Globo.io charges Shopify Plus stores more (opens in new tab) for the same plan, and both offer a 14-day free trial, enough time to run your top no-results terms through each one before you pay.

Algolia bills on usage. Its Grow plan (opens in new tab) includes manual synonyms and keeps analytics for 30 days, Grow Plus adds AI Synonyms and keeps analytics for 90 days, and the Free plan keeps them for 7 days. On its Shopify listing (opens in new tab), Algolia says the app significantly reduces searches that return no results and saves merchants from building synonyms by hand.

Klevu’s app listing (opens in new tab) is free to install and lists Klevu as the developer, now Athos Commerce, with charges that may be billed separately from your Shopify invoice. Its three tiers (opens in new tab) are priced on request, and two of them are standalone annual plans. Klevu claims its platform significantly improves conversion rate and average order value.

Whichever app you trial, judge it on your own data, not a demo catalog. Take the top 20 terms from Searches with no results and Searches with no clicks, run each one through Shopify’s search and the app’s, and compare what comes back. After you install, recheck that Shopify’s search reports and GA4 still record searches, since Shopify’s tracking depends on how the theme builds the search page and GA4’s depends on the q parameter.

Shopify’s Search conversions over time report gives you the starting line for measuring ecommerce site search. Record the click rate, add-to-cart rate and purchase rate for search sessions before you change anything, fix the top zero-result terms in your product data, and read the same report again once the 72-hour delay has passed and a comparable number of sessions has built up.

Search data belongs in the same review as the rest of your funnel, and reading it there is part of the conversion rate optimization work BLKDG does, from finding where visitors drop off to testing the fixes. If your catalog has outgrown its navigation and you want to know what search is costing you, that’s the place to start.