Sportswear Brand

The Challenge

The client is a mid-size sportswear and lifestyle retailer running on their own Shopify store — strong repeat customer base, active paid social, decent organic presence, annual revenue in the $3–5M range when the engagement started. Most of their traffic was mobile.

They came to us frustrated. Paid ads were performing below industry benchmarks, they'd switched agencies twice in two years, and they couldn't figure out why. Traffic looked fine, the brand resonated with the audience, but somewhere between the ad click and the checkout they were losing people.

The diagnosis came from a single afternoon with Google's performance tools. On a typical mobile connection, their product pages were taking over six seconds to fully load — more than double Google's threshold for an acceptable mobile experience. At that load time, roughly one in two mobile visitors leaves before the page finishes rendering. For a brand that gets the bulk of its traffic from a phone, that isn't a margins problem or a creative problem. It's a plumbing problem.

The site had been built by a freelancer two years earlier on a popular Shopify theme. Over the months, apps had accumulated — reviews, upsells, a loyalty widget, a chat bubble, a size-guide overlay — and nobody had circled back to check what they were doing to load time. Each app added a few hundred milliseconds. Twelve apps later, the site was effectively unusable on a phone with a normal connection.

Meanwhile, Google was watching. Slow sites get less organic traffic — not as a penalty, but because Google ranks pages partly on the experience they deliver. The brand's rankings had been quietly softening for 18 months without anyone connecting the trend to the technical decay.

The Approach

Five months. No redesign, no rebrand, no new theme. The brief was to make the site work the way it was supposed to on a phone — and nothing else.

Product page templates rebuilt from scratch

The product templates were the highest-traffic, highest-leverage surface on the store. They were rebuilt from the ground up — same visual design, faster bones underneath. Render-blocking patterns were unwound, theme code was rewritten where the original Liquid was doing more work than the markup needed, and unnecessary DOM was stripped before the browser ever had to parse it.

Image pipeline replaced

The legacy lazy-loading setup was replaced with a proper next-generation format pipeline (WebP and AVIF with appropriate fallbacks, correctly sized for each viewport). Hero and gallery imagery now appeared in under a second on mobile rather than trickling in over four or five.

App audit: remove what's not earning its weight

Every installed app was audited against actual business impact. Four were removed outright — they had no measurable contribution to revenue, retention, or any other tracked outcome but were still injecting JavaScript and CSS into every page load. Two more were replaced with lighter-weight alternatives that did the same job without the overhead.

Font loading sequence fixed

The original setup left text invisible for nearly two seconds while web fonts loaded — the classic "flash of invisible text" pattern. The font loading was reordered so the system fallback rendered immediately, with the brand fonts swapping in once they finished downloading. Visually almost imperceptible; for Core Web Vitals (and for users on slower connections) it removed two seconds of dead air.

The Outcome

By the end of month 5, the brand's mobile pages were loading in under two seconds — down from the original six-plus.

Within 90 days of go-live, the business numbers had moved in lockstep:

• Mobile sales up 34% — same traffic, same ad spend, more of it converting.
• Cart abandonment on mobile down 22% — fewer people bouncing between the add-to-cart click and the order confirmation.
• Google organic clicks up 38% as rankings recovered across the top product category pages.
• Revenue per visitor up $9 on mobile versus the pre-rebuild baseline.

The paid social ROAS improved over the same period — not because anything was changed in the ad accounts, but because the same budget was now sending traffic to pages that actually worked.

The founder put it simply: "We'd been blaming our ads team for two years. Turns out we were sending people to a broken shop."

Tools & Stack