DTC teams that stopped guessing what to show
Early-access pilot partners across apparel, home goods, and beauty. 8 brands, 90-day test windows, internal attribution.
Doveridge Goods: 12,000 SKUs, 90 days, +22% add-to-cart
The problem
Doveridge Goods had grown its catalog from 2,000 to 12,000 SKUs over two years. What worked for a curated 2K-product catalog, where a small merchandising team could manually surface the right items, broke at 12K. Their homepage carousel was showing the same 8-10 products to every visitor, all manually picked on Monday mornings by a merchandiser who couldn't possibly review 12,000 options.
The result was predictable: flat add-to-cart rate despite catalog growth. The new products weren't getting discovery. The seasonal depth they'd built was invisible to most visitors. The catalog was growing but the storefront wasn't keeping up.
What we tested
Doveridge integrated Aislegleam on their Shopify storefront in two days. The integration added the JavaScript tag for behavioral signal collection, and replaced the homepage carousel and PDP related-items module with Aislegleam API calls.
We ran a 90-day A/B test: 50% of sessions saw the personalized recommendations, 50% saw the original manually-curated carousel. Attribution was measured by the Aislegleam dashboard with a 1-day attribution window for add-to-cart events, matched to Shopify order data for revenue.
The catalog was not reorganized. No new products were added. The test was purely about what order the existing catalog was surfaced in, per session.
The result
The personalized sessions showed a +22% add-to-cart rate on the homepage carousel over the 90-day window. The effect was strongest in weeks 3-12, after the model had accumulated enough session signal to sharpen intent prediction beyond cold-start heuristics.
Secondary finding: catalog depth increased. Products outside the top-200 manually-surfaced items appeared in 34% of the personalized sessions' add-to-cart events, compared to 8% in the control group. The catalog was working for the first time.
Source: 90-day A/B pilot, Doveridge Goods, internal attribution via Aislegleam dashboard. Beta-access cohort, early-access pricing terms.
Across 8 pilot partners, 90-day windows
What pilot partners said
"We grew our catalog to 12,000 SKUs thinking that's what would drive revenue. What we didn't have was a way to show the right 12,000 to the right person. Aislegleam solved that without asking us to replatform or hire a data team."
"The part that surprised me was how quickly the model got useful. By week two we were already seeing session-level differentiation that made sense to our merchandising team. It wasn't a black box, it was matching what we'd expect good buying judgment to do."
"We were skeptical about a personalization layer that didn't need login data. Our 85% anonymous session rate made most vendor pitches sound like assumptions. The session-based model actually handled that correctly, which the A/B data confirmed."
Join the next cohort
We're onboarding a second cohort now. Standard pilot: 90 days, A/B test window, internal attribution. Book a call to discuss your catalog size and integration stack.