Customer service was fielding more complaints about product filtering than almost anything else on the site. Over five a week, every week, for a year straight. Customers couldn’t find what they wanted, filters reset the moment they clicked back from a product page, and the mobile experience buried style filtering entirely. It was costing us conversions, and nobody had connected the dots on how much.

I owned this end-to-end: identifying the problem before it was assigned to anyone, auditing the competitive landscape, and designing the fixes, which I prioritized by development effort so engineering could ship the highest-impact changes first.

This project didn’t start as a brief. While reviewing post-purchase survey feedback and customer service logs, I noticed the same frustration surfacing again and again: filters that reset without warning. Customers described the site as “extremely hard to navigate,” complained that hitting back wiped out everything they’d selected, and said they’d abandoned carts entirely because they couldn’t find their way back to an item. One customer put it simply: they’d enjoyed the shoes, but didn’t think they’d be back.

That pattern, repeated across dozens of survey responses and support tickets, became the case for a real audit rather than a bandaid.

I audited how four category leaders (Macy’s, Converse, Crocs, and New Balance) handled filtering, and mapped it against TOMS’ own experience. Macy’s, rated best-in-class by the Baymard Institute, got the fundamentals right: multi-select filters, saved state across sessions, truncated lists that didn’t overwhelm the user.

Rather than hand engineering a wish list, I organized every recommendation by development effort, low, medium, and high, so the roadmap could ship value immediately while building toward the bigger fixes. The single highest-priority item: saving a user’s filter selections for the duration of their session, the exact complaint driving most of the feedback.

The updated filter system introduced multi-select checkboxes with live result counts, collapsible filter groups to reduce scanning fatigue, and persistent state so a filter selection survived a trip to a product page and back.

Once shipped, filter-related complaints dropped by roughly 70%, and mobile conversion on filtered product listing pages rose about 8%.

What started as a pattern in survey data became one of the most measurable UX wins on the site. It’s also a reminder that some of the highest-impact work isn’t handed to you in a brief. It’s sitting in the feedback nobody’s read yet.