Lipstick sold? The lip liner, gloss, and setting spray complete the look.
Beauty doesn't shop SKU-by-SKU. It shops routines: lipstick + matching lip liner + setting spray, foundation + concealer + setting powder, serum + moisturizer + SPF. MarketBasketAnalysis mines per-cohort patterns (skin type, color family, routine slot) and proposes the bundle that completes the look without cheapening the brand.
(Shopify, WooCommerce, Magento, plus subscription tooling)
Why this fit
Beauty AOV depends on routine completion: serum plus moisturizer plus SPF, primer plus foundation plus concealer plus setting spray, lipstick plus matching liner plus gloss. Generic recommendation widgets pair by category similarity (more lipstick when the customer bought lipstick) instead of by routine completion (the liner that finishes the lip look, the setting spray that locks in the foundation). Our mining surfaces routine-completion patterns specific to skin-type and color-family cohorts: the bundle that converts for the oily-skin customer is different from the one that converts for the dry-skin customer, and we segment accordingly.
The right primitives for the job
Skin-type cohort mining
Mining segments by skin-type cohort (oily, dry, combination, sensitive) using SKU metadata + customer signal. The hydrating moisturizer that pairs with the dry-skin serum is not the mattifying primer the oily-skin customer wants.
Color-family bundles
Lipstick + matching lip liner + complementary gloss. Foundation + matched concealer. Eyeshadow palette + the eyeliner shade that finishes the look. Color-family pairing is mined from real customer behavior, not chosen by a stylist's gut.
Routine-slot ranking
HUI engine ranks bundles by margin contribution AND routine completion. The serum-plus-moisturizer-plus-SPF bundle that completes the AM routine beats the random-sample-pack bundle that crosses category lines for no reason.
Brand-voice AI bundle copy
AI-generated bundle names + descriptions trained on your existing PDP copy. A bundle named 'The Glow Routine' reads like your brand wrote it, not like a recommendation engine.
Replenishment for finishers
Setting sprays empty in 6 weeks. Mascaras dry out in 3 months. SPF gets used up daily. predict_reorder predicts the actual cadence per product category and per customer, so your subscription tool (Recharge, Loop, Smartrr) plugs into the right reorder window automatically.
Compliance-aware copy
AI bundle copy generator respects FDA structure / function rules on skincare claims. No 'cures acne,' no 'treats wrinkles.' Pulls language patterns from your existing PDP copy so brand voice + compliance stay aligned.
What you skip
The friction we're explicitly cutting out.
- Per-impression billing on your high-traffic PDP and collection pages
- Revenue share on every bundled finisher product
- Stylist-hand-curated bundles that go stale with every collection drop
- Generic ML that pairs more lipsticks when the customer wants a setting spray
- Vendor lock-in (your bundles are native catalog products that persist at uninstall)
Want the full buyer's guide?
Three free guides cover the math + the vendor landscape: agentic-commerce in 2026, MBA vs Bloomreach, and the pricing math vs the Glood / Rebuy / PickyStory cluster.
Browse the resource hubThe lipstick is the hero. The liner, gloss, and setting spray are the routine.
$49/month flat. Skin-type and color-family cohort mining, brand-voice AI bundle copy, native Shopify Bundle products, predict_reorder for replenishment. Plugs into Recharge / Loop / Smartrr / Shopify Subscriptions on day one.
Get started