How should collection pages and site search sort and filter products?
Our answer
Build filters from the attributes shoppers choose by, and show applied filters and a visible search field on mobile. Test a sticky filter bar and a sales-based default sort against your current setup.
- Confidence
- Medium
- Strongest proof
- 9 controlled tests
- 5 official rules, 1 dataset, 7 practitioners
- Sources
- 7 sources
- 22 statements
- BC
BI
CCDG
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentMost large sites get filtering wrong. Baymard (ecommerce UX research firm) finds 79% of large sites use categories where filters belong, and 66% of mobile sites show no overview of applied filters.
A visible search field sells more on mobile. In tests by Blend Commerce (Shopify CRO agency), showing the field instead of an icon raised conversion 9.97% to 19.62% at three stores, at 80% to 97% significance.
Sorting may matter, though the proof is weak. Dynasort, which sells automated sorting, found its sorted collections convert at 1.48% against 0.74% on Shopify's default order. It says the gap partly reflects selection.
Our judgment
Make the search field visible on mobile, since Baymard's research and three of Blend Commerce's search tests point the same way. Test a sticky filter and sort bar rather than adding it by default, since Blend Commerce's two tests of it split. Test a sales-based default order, such as Shopify's Most relevant or best selling, against your current order before keeping it, going by Shopify's sort options and Dynasort's comparison, which Dynasort says is observational. Test collection card changes by device before rolling them out, going by Blend Commerce's second image test.
Our judgment, built from Baymard Institute, Blend Commerce, Shopify and Dynasort: our reading of these sources, not a finding from any one of them.
What the sources say
22 statements from 7 sourcesShopify's Search & Discovery app allows up to 25 filters per store and 100 values per filter, and collections with over 5,000 products show no filters. Your theme must support filters for shoppers to see them.
“Collections that contain more than 5,000 products don't display filters.”
Shopify's help page says new collections sort by Most relevant by default, based on sales performance. Other options include best selling, title, price, date or manual order.
“Products are sorted by Most relevant by default when you create a new collection.”
Shopify says its store search has built-in semantic understanding from product titles, descriptions, types, categories and tags, not images. You can boost products for set terms and place sold-out items last.
“Semantic understanding is included in online store search. You don't need to turn it on”
Shopify's help page says all merchants with an online store get predictive search and typo tolerance by default, while the Search and Discovery app adds filters, product boosts and recommendations.
“Core search features are available by default to all Shopify merchants that have an online store.”
Baymard's usability research on 180+ US and European sites found 79% use categories where filters should be used, which stops shoppers refining by everything they care about.
“79% of Sites Use Categories When Filters Should Be Used”
On the Scalability School podcast, a plain headline stating the product and its benefit beat persona-specific or clever headlines on collection pages by about 3 to 5% more revenue per visitor.
“that plain USP headline crushed everything... by like three to five percent more revenue per visitor.”
Blend Commerce, which sells A/B testing, reports one store's test where category pills lifted collection page conversion 51%, revenue per visitor 79% and add-to-cart 45%. No sample size or significance given.
“Conversion rate increased 51%, revenue per visitor rose 79%, and add-to-cart rate increased 45%.”
In one Blend Commerce A/B test on Stone Creek Coffee, a two-image gallery on collection cards raised mobile conversion 20% and revenue per visitor 11%, but cut desktop conversion 12%. Significance not given.
“On mobile, conversion rate increased 20%, revenue per visitor rose 11%”
In Blend Commerce's test on one fragrance store, showing trending searches when search opened lifted revenue per visitor 1.92% and order value 3.30% but cut conversion 1.33%; not significant.
“The test did not reach statistical significance”
In one Shopify store's A/B test, a sticky mobile filter bar raised filter clicks by 299.58%. Conversion rate and revenue per visitor moved down 2.23% and 1.80%, not statistically significant.
“The commercial measures did not produce a statistically significant win”
In one sports nutrition store's test by Blend Commerce, showing price per serving on collection pages raised revenue per visitor 8.25% and conversion 5.96%. No confidence level was given.
“No confidence level was supplied, so the uplifts should be treated as directional.”
In one store's Blend Commerce test, showing a search field under the mobile header instead of an icon raised conversion 9.97% and revenue per visitor 6.10%, at 87.31% confidence.
“only 1.42% of visitors used site search, yet visitors with search intent converted at 11.47%, compared with a 4.44% site average.”
Blend Commerce, which sells Shopify CRO, reports one client's one-week test: a visible mobile search bar raised conversion rate 19.62% and revenue per visitor 31.30%, at about 97% significance.
“After one week, the exposed search bar increased conversion rate by 19.62%, revenue per visitor by 31.30%”
Blend Commerce says a visible mobile search bar lifted one store's conversion 13% and revenue per visitor 12% over two weeks, at 80% statistical significance. Only 0.76% of shoppers had used search.
“The exposed-search treatment increased conversion rate by 13% and revenue per visitor by 12%”
Dynasort, which sells collection sorting, says sorted collections converted at 1.48% versus 0.74% by default across 8.1 million sessions in July 2026. It says part of that gap is selection, not a controlled test.
“These are observational fleet numbers, not a controlled experiment”
Clean Commit, a Shopify developer, relays an undated test where removing star ratings from category pages improved add to cart, since identical ratings added noise. Stars help there only if ratings vary.
“Stars on a collection page only help if they vary. Check the spread across your catalogue before assuming it's a win.”
Baymard's mobile benchmark lists common gaps: 42% of sites lack a View All option at each catalog level, 66% don't show applied filters in an overview, and 73% hide color swatches in product lists.
“Display “Applied Filters” in an Overview (66% Don’t)”
In one barbecue store's A/B test by Blend Commerce, a sticky filter and sort bar on mobile raised conversion rate 21.39% and revenue per visitor 95.34%. Single-store result, significance not shown.
“We introduced a sticky filter and sort bar on mobile collection pages”
Baymard Institute found 28% of the sites in its UX benchmarks show no overview of applied filters. Users who missed even one applied filter tended to misread what range of products the site carried.
“28% of sites across our UX benchmarks don’t display an overview at all.”
Blend Commerce says exposing mobile search cut conversion 11.76% at one coffee store where 20% of searchers saw no results, so fix search quality before making search more visible.
“Twenty percent of search users saw a no-results page during the experiment.”
Baymard says about half of shoppers prefer site search to find products, yet 46% of desktop and 58% of mobile sites have mediocre or worse search, and shoppers assume poor results are the full range.
“participants often simply assume that poor or limited results were the site’s full selection”
Google says faceted navigation URLs can create endless URL spaces and slow discovery of new pages; if you don't need them indexed, block them in robots.txt, since canonical and nofollow work less well.
“Use robots.txt to disallow crawling of faceted navigation URLs.”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer
Read next
More in Product and collection pages- Should we run upsells and cross-sells on the product page, in the cart, or after purchase?
- How many product photos and videos does a product page need, and which kinds?
- Should we show user-generated photos and videos on product pages?
- How long should product descriptions be, and how should they be laid out?