Impixo/Case studies/D2C e-commerce
A D2C brand doing solid paid-social volume with a bloated theme, a 6.2-second product page and no structured product data. We replatformed, rebuilt the PDP around objections, and made the catalogue readable by both Google Shopping and AI assistants.
Same traffic mix, same ad spend. The lift came entirely from page speed, PDP structure and a shorter checkout.
6.2s down to 1.8s LCP on 4G mobile. Nineteen apps audited, eleven removed, the rest deferred.
Fewer steps, visible total including shipping earlier, and express wallets above the fold.
01 / The problem
The brand was buying attention and then losing it. A premium theme with nineteen installed apps meant 3.1MB of JavaScript before a single product image rendered. On mobile — 78% of their traffic — the product page took 6.2 seconds to show anything meaningful. That cost them on conversion and again on ad quality scores.
Underneath the speed problem was a data problem. Product pages had no Product or Offer schema, variants weren't exposed, and reviews lived inside a third-party widget that rendered client-side. To a crawler — or an AI assistant asked to recommend a product in that category — the catalogue was effectively blank.
02 / What we built
We rebuilt the theme from a minimal base, cut eleven apps outright and replaced three more with native code. Images moved to responsive AVIF with explicit dimensions, third-party scripts were deferred behind consent, and the critical CSS for above-the-fold content ships inline.
The product page was restructured around the actual buying objections, taken from support tickets and post-purchase surveys: sizing, delivery timing, returns and material sourcing. Each got a dedicated block rather than being buried in a tab. Reviews were pulled server-side so they exist in the HTML.
On the data side, every product now emits full Product / Offer / AggregateRating markup with variant-level availability and price, plus category-level buying guides written to answer comparison questions directly. That combination is what gets a brand named when someone asks an assistant for a recommendation instead of typing into Google.
03 / What moved
The commercial lift showed up within the first fortnight — revenue per session climbed 38% on flat traffic, and cart abandonment fell 24%. The slower-burning win was visibility: within two months the brand started appearing by name in assistant answers for category comparison prompts, driven by the buying guides and the structured product data behind them.
Paid traffic pays once. A catalogue an assistant can actually read keeps paying.
04 / The retainer
The engagement runs on a Scale retainer — one structured CRO test per month, new buying-guide content, schema updates as the catalogue changes, Core Web Vitals monitoring, and a monthly AI visibility re-score against a fixed prompt set so we can see whether the content is landing.