Insight

Mobile Filters Disappeared During Product Browsing

Blend’s CRO audit found that Little Bible Stories’ filter and sort controls sat at the top of mobile collection pages and disappeared as shoppers scrolled. On smaller screens and across different resolutions, returning to those controls required extra effort just when customers were trying to narrow the product range. This created friction in the product-discovery journey: shoppers had to interrupt browsing and scroll back before refining results. Accessible filtering is especially important on mobile collection pages, where limited screen space makes it harder to compare options quickly. The recommendation was therefore to keep these controls within reach throughout the session. This evidence-led approach reflects how Blend uses Shopify CRO audits to turn observed journey friction into prioritised, measurable experiments.

Jump to results

A/B Test

Sticky Filter and Sort Controls on Collection Pages

We ran an A/B test on mobile collection pages using Intelligems. The control kept the existing filter and sort controls at the top of the page. Variant 1 introduced a sticky bar that remained available while users scrolled. We hypothesised that persistent access would help shoppers refine listings faster, improve product discovery and increase conversion potential. The primary goal was conversion rate, supported by filter and sort clicks, PDP visits, revenue per visitor, average order value and add-to-cart rate.

The Results

Variant 1 increased clicks on Filter and Sort by 299.58%, showing that persistent controls materially changed how shoppers used the mobile collection page. The commercial measures did not produce a statistically significant win: overall conversion rate and revenue per visitor moved directionally down by 2.23% and 1.80%, while average order value was flat. Segmenting the data revealed a more useful pattern. New visitors showed small directional declines, whereas returning visitors recorded a 2% increase in product-page views and 1% increases in revenue per visitor and average order value; the variant had a 93% probability of being best for returning-visitor product-page views. Blend therefore treated the result as a low-risk UX improvement, not a revenue uplift claim, and implemented Variant 1 permanently. A useful follow-up A/B test would examine whether clearer filter labels or default states help first-time visitors turn the extra interaction into faster product discovery, while monitoring conversion and revenue by visitor type.

  • 299.58%

    increase in filter and sort clicks

  • 3.1%

    increase in returning visitor net revenue

  • 2%

    increase in returning visitor product page views
Book a CRO discovery call

See More A/B Tests