Brunt Workwear's 'Shop by Toe' Test: When Guided Nav Beats a Clean Homepage

Brunt Workwear is running a homepage test that comes down to a single, deceptively simple question: does telling boot shoppers to pick their toe type before they even scroll past the hero help them, or just get in the way?

We caught it via same-run variance — same day, same traffic mix, two different homepages depending on browser. Visitors on Safari and Firefox (variant BW137-A) see a "Shop by Toe" section with three tiles — Soft Toe, Comp Toe, Brunt Toe — sitting right between the "Gear Up & Save" banner and "Shop Our Lineup." Visitors on Chrome and Edge (variant BW137-C) never see it. The section just doesn't exist for them; the page jumps straight from the promo banner to the lineup grid. Our analyst flagged this at 92% confidence, and it's a clean one — no ambiguity about layout differences once you know the safety-toe categories are the whole test.

Why toe type is the highest-leverage filter on this site

If you've never bought safety footwear, this test looks arbitrary. If you have, it's the first thing you ask. Soft toe, composite toe, and steel toe (Brunt's branded "Brunt Toe") aren't style preferences — they're compliance requirements tied to a jobsite. A warehouse worker with an OSHA mandate for comp toe isn't browsing for aesthetics; they're filtering out 60% of the catalog before they care about anything else.

That's what makes this a smarter test than most "add a section to the homepage" experiments I've seen teams run. It's not decoration — it's pre-qualification. The hypothesis Brunt is almost certainly testing: does surfacing the toe-type decision early reduce the number of people who bounce off a generic lineup because nothing obviously matches their requirement?

I've shipped this exact pattern before, just with a different constraint variable — company size instead of toe type, region instead of comp toe, tier instead of steel toe. The lesson was consistent: when the segmenting variable is something the user already knows about themselves and it's non-negotiable, putting it above the fold isn't clutter, it's triage. When the segmenting variable is soft (style, "who are you shopping for") it tends to add friction instead of removing it, because most users don't have a firm answer yet.

What the omission variant is really testing

It's tempting to read variant C (no section) as "the control" and A as "the challenger," but I'd push back on that framing a little. Homepage real estate above "Shop Our Lineup" is prime territory, and every section you add there pushes the lineup — arguably your actual money-maker — further down. Brunt's design and growth teams are weighing a real tradeoff: guided segmentation vs. speed to the general catalog.

This lines up with what Brunt has apparently learned from prior site work. Their case study with Anatta on the US site launch reported a 20% conversion lift within 11 days and a 40.7% lift in mobile conversion — gains that came from tightening the path to product, not from adding more homepage real estate. That history tells me the team is disciplined about page weight, which makes the "Shop by Toe" addition more interesting, not less. They wouldn't test adding a section unless they had a real signal that toe-type confusion was costing them clicks somewhere in the funnel — likely from search queries, support tickets, or return reasons tied to wrong-toe-type purchases.

The part most teams get wrong with category selectors

I've watched category selector tests go sideways in two predictable ways. First, teams build the selector, ship it, and measure homepage engagement instead of downstream conversion — clicks on the tiles go up, but nobody checks whether those visitors actually convert at a higher rate than the ones who used the main lineup. A tile that gets clicked a lot but leads to comparison paralysis is a vanity win. Second, teams add the selector but don't remove anything to compensate, so the page just gets longer and slower, and the lift from better segmentation gets eaten by higher bounce from load time or scroll fatigue.

If I were running this test at Brunt, the metric that matters isn't clicks on Soft Toe / Comp Toe / Brunt Toe — it's whether the segment of users who land on a filtered category page from that tile convert at a materially higher rate than users who arrive at the same category via the general lineup or search. If the answer is yes, the homepage cost (one more section, more scroll) is worth it. If the lift is mostly people who would've found their way there anyway, it's just added weight.

What a two-variant browser split tells us about test maturity

One detail worth flagging for other teams running experiments: splitting by browser (Safari/Firefox vs. Chrome/Edge) rather than a randomized cookie-based split is unusual and worth a gut-check on Brunt's end. It's possible this is an artifact of how their testing tool is bucketing sessions rather than an intentional targeting rule — but if it's real, it introduces a confound, since browser choice correlates with device type, OS, and even income bracket in some categories. A "Safari-heavy" variant population skews more iOS/mobile and often more affluent than a Chrome-heavy one. If Brunt is reading this test as a clean A/B on layout, they should double check the assignment logic isn't accidentally also testing "mobile Safari users" against "desktop Chrome users." That's a classic way to get a false read on a legitimate UX hypothesis — the frameworks in The Good's testing playbook are a solid gut-check for teams who want to map customer journey segments before trusting a lift number.

What I would actually do next

If I were on Brunt's growth team, I wouldn't just watch homepage CTR on this test — I'd instrument the next two steps in the funnel (category page engagement and add-to-cart rate) segmented by entry point, and I'd manually verify the variant assignment isn't accidentally confounded with device or browser cohort before reading any lift as causal. And if you're running a similar "guided filter vs. clean page" test on your own homepage this sprint, ship it behind a randomized session split, not a browser-based one — you want to know if the filter works, not which browser your best customers happen to use.