The context
The brand was spending well on paid acquisition and pulling strong top-of-funnel traffic. But the numbers told a clear story: people were adding to cart and then vanishing. Store-wide conversion sat at 1.9% — below where a brand with this traffic quality and AOV should be.
The previous developer had treated the theme as “done.” Nothing was being measured. No tests were running. The team was guessing.
The problem
Three issues surfaced in the first session-recording review:
- The cart drawer hid shipping cost until the final checkout step — buyers hit an unexpected number and bailed.
- The mobile add-to-cart sat below a heavy block of marketing copy, so on the device 71% of traffic used, the primary action was off-screen.
- There was no urgency or reassurance at the decision point — no stock signals, no guarantee, no social proof near the buy box.
The intervention
I ran a structured CRO programme rather than a redesign. Each change was a hypothesis, instrumented and measured against a control:
- Surfaced shipping thresholds in the cart drawer with a free-shipping progress bar.
- Rebuilt the mobile PDP so the buy box sits above the fold, copy below.
- Added a trust cluster at the buy box — guarantee, verified-review count, low-stock signal where genuine.
- Simplified the checkout to a single accelerated path for returning customers.
The process
Weekly cadence: review the data, ship the next test, read the result, decide. No test ran without a control and a pre-registered success metric. Losers were reverted without ceremony — roughly a third of hypotheses didn’t win, which is exactly how a real programme behaves.
The result
Over eight weeks, store-wide conversion moved from 1.9% to 2.6% — a 38% relative lift — with checkout abandonment cut nearly in half. Because the gains came from existing traffic, the brand’s blended acquisition cost effectively dropped without touching ad spend.
The work that mattered wasn’t a redesign. It was discipline: measure, test, keep what wins, kill what doesn’t.