Twelve months inside Ardant
Ardant is an 8-figure jewelry brand. This is what an embedded CRO program actually looks like over a year: not a highlight reel, the real sequence, losses where they happened, and how single tests compounded into an offer-level rebuild of the main funnel.
Months 1 to 2
The teardown and the first levers
The program opened the way every program does: a full-store teardown and a ranked backlog built from patterns proven across our portfolio. The first offer-level swing was a free-shipping threshold, up +9.5% revenue per visitor. The same weeks also produced an early landing-page loss that set a rule we still use: moving product content up the page is not automatically better.
Months 3 to 5
The Q4 offer season
Into BFCM we pushed the offer surface hard: collection-page banners carrying the promo, custom card badges, gift wrap in the cart, and a stock widget above the fold. Four winners. The most valuable result of the quarter, though, was a loss: cutting the free-shipping threshold for BFCM looked obvious and would have cost $833K a year. The test caught it; the control stayed.
Months 6 to 8
Compounding the product page
With the offer structure holding, testing moved to the buy zone: UGC videos under the add-to-cart button, a cleaner cart with restructured upsells, and information dropdowns answering sizing and quality questions at the decision point, the biggest single lift of the program at +12% revenue per visitor. A bestseller price increase also went live here, lost, and was rolled back within the test window.
Months 9 to 12
Offer-level testing
The mature phase of a program is offer work. We tested the shipping price itself, $10 against $8 against $7, worth +7.1% conversion, added trust badges inside checkout, and replaced the top-spending ad funnel outright with a custom-coded two-page stack builder: +15.2% conversion, +10% AOV, +26.7% revenue per visitor, roughly $100K a month. That funnel test also carried a new offer, retiring the email-popup discount entirely.
From the Landing Page Lab
The stack builder funnel: the offer-level rebuild from months 9 to 12, with the full pages and the story.
Read the case ↗Numbers pulled from Ardant's live performance summary in the shared Notion workspace: 37 completed tests, 16 winners, +$10.7M added annualized revenue, 222:1 return on fees. Every linked test below shows its own full data.
Want a program like this on your store?
Free teardown: we walk your store live and show you the three highest-impact tests we would run first.