Blog / Shopify & e-commerce /

How to test collection filters before publishing theme changes

The theme update looks great in the preview. Then you tap a filter on your phone, select “Black”, and the page shows zero products. Black products exist. Lots of them.

Filters help shoppers narrow choices, but a theme update can change their labels, order, URL behaviour or the way they respond to touch. Test them before publishing, so the collection stays understandable, mobile-friendly and consistent with the products it shows. This is one of my must-do checks on any Shopify theme change.

Map important filter tasks

List the choices customers need to make: size, material, compatibility, colour, price or availability. Confirm which attributes exist in the catalogue and how reliably they’re filled in. A filter with inconsistent data can hide relevant products or return empty results. Don’t add a filter just because the theme supports it.

Decide how selected filters should appear, whether multiple values combine, how a person clears them and what happens on back navigation. Review the URL changes and whether filtered routes are intended for sharing or search discovery. Many filter combinations may not need indexable landing pages.

Test on realistic collections

Use collections with enough products and a range of attribute values. Test one filter, combinations, no-result states and clearing all selections. Check that product counts, cards and selected chips update correctly. Confirm the page remains usable with long labels and translated terms.

Test mobile touch targets, keyboard navigation and screen reader announcements. A filter drawer should open, retain selection and close without losing the state. Ensure sticky buttons do not cover content. Check that controls work with slow loading or when a script fails.

Inspect technical output

Review rendered links, URL parameters, canonical output, robots directives and crawl behaviour. A theme change may alter how filters are encoded or applied. Check a crawl of representative states rather than assuming the default collection test covers them. Confirm analytics still measures filter actions if those events inform decisions.

Publish and monitor

Test staging and production after deployment, since apps, caching or market settings can differ. Keep screenshots, test URLs, expected behaviour and the owner in the release record. Watch support feedback and no-result patterns. If a filter is confusing or unreliable, I’d remove or fix it rather than leave a broken control on the page.

Include catalogue and app dependencies

Filters often rely on metafields, tags, product options or app data. Confirm which source controls each value and how changes propagate. If a theme update changes filter configuration, an apparently empty result may be a data mapping issue rather than a visual defect. Check representative products in the admin and on the storefront.

Test with the actual market and language settings. A filter label may be translated while the underlying value remains in another language. Check currency ranges and measurement units where those are used. A product should not disappear from a category because its attribute was formatted inconsistently.

Set release acceptance criteria

Write expected behaviour for each test: selected state, result set, count, URL, clear action and accessibility behaviour. Keep this short and relevant to the change. Require a review when filters alter indexable URLs or expose new landing pages. Confirm that no accidental crawl path is introduced.

After release, review errors and support questions. If a filter is rarely used, find out whether it solves a real decision before you remove it. It may serve a small but important product set.

Test the page without filters too

The default collection should remain useful before any filter is selected. Check the product order, category description and number of choices. If every visitor must filter immediately to see relevant products, the collection or navigation may need a different structure. Test reset behaviour so clearing the last filter restores a sensible state.

Use keyboard-only navigation and assistive technology for the filter controls. Confirm labels describe the attribute and state, that focus returns predictably when a drawer closes, and that result updates are announced where appropriate. A visually selected checkbox should not be the only indication that a filter is active.

If filters create crawlable URLs, document which combinations are meaningful and which aren’t. Make sure a theme change doesn’t suddenly expose a large number of near-duplicate pages.

Check performance and fallback behaviour

Filters can require scripts or data requests. Test whether the collection remains usable when a request is delayed or fails. A loading indicator should not obscure the results indefinitely, and the visitor should be able to clear the selection or return to the collection. Check that browser back and forward controls behave predictably.

Review how filter selection interacts with sorting and pagination. Changing a filter should not leave the visitor on an empty page number or preserve a sort that produces confusing results. Confirm counts and product cards update together. Test with products that have missing attribute values so gaps do not create silent exclusions.

Use test cases that reflect actual customer choices rather than every possible combination. Keep an owner for filter data and a process to remove values that no longer apply.

Check how the filter behaves after a visitor follows a product link and returns to the collection. Keep a useful selection when appropriate, but don’t trap the shopper in a stale or empty result set.

Run the filter test after content and product data updates as well as theme releases. A new attribute value can cause the same empty-result problem as a code change, so ask merchandising to verify representative products. Which filter on your store would hurt most if it returned nothing after the next update?