The cart is missing the thing that goes with it.

Connected

Current cart contents and category.

Inferred

Affinity: customers buying this camera frequently add a compatible SD card before paying.

Example eligibility rule

Show a complement only when its attach rate is at least twice the category baseline in a sufficiently large set of completed orders, it is compatible with the exact cart item, and the customer does not already own or have it in the cart.

A tent without stakes. A camera without a card. A shirt without the usual care kit. Basket analysis plus this cart says something is missing. The page still shows "you may also like" scented candles.

Avoid this assumption

Cross-sell is only personalization when it is about the job in the cart, not leftover merchandising slots.

Keep the offer inline, useful, and easy to test. A true complement is not decoration. It helps the customer complete the job already represented by the cart.

Reach

How many eligible people can you reach in the chosen channel and time period?

Cost

What will content, delivery, QA and ongoing maintenance cost?

Return

What incremental outcome would cover that cost, and what is it worth?

Planning inputs, not measured results. Estimate reach × incremental outcome rate × value per outcome, then subtract cost. Use a range and a consistent time period.

Example format: Recommendations

Inline add with a reason ("most people add a spare").

Compare lighter and deeper responses

Choose the lightest response that solves the need.

Assist: Offer useful help

One well-named complement under the cart line item. Easy to ignore.

Adapt: Change a meaningful part

Inline add with a reason ("most people add a spare").

Transform: Tailor the journey

A "complete the job" bundle, still optional, with a one-click add.

What to watch out for

• Three carousels. One complement.
• Recommending a second of the same item they already have.
• Inserting the module above the checkout button on mobile.

Control

Default cart, no job-aware complement.

Treatment

Inline complement based on cart contents.

Success measure

Complement attach rate and average order value, alongside checkout conversion.

Guardrail

Checkout drop-off. If the module adds friction, kill it.

Among carts eligible for the same complement rule, compare the inline offer with a no-offer holdout. Measure attach rate and order value alongside checkout completion, returns, and removal of the suggested item.

Win or lose, record what worked, for whom and why. What you learn becomes the next signal.

• Dynamic Yield Personalization and recommendations platform with a merchandising engine and Experience APIs for running recs and server-side experiments on any channel.
• Bloomreach Ecommerce Search AI-driven ecommerce search and discovery platform (Loomi AI) that personalizes ranking, recommendations, and merchandising from first-party shopper data.
• Monetate Retail-focused "Experience Optimization Platform" combining AI-driven personalization/recommendations, merchandising, and no-code/server-side A/B testing under one engine (MONET), backed by a managed-services "Concierge" team.
• AB Tasty content personalization Audience- and intent-based content targeting for anonymous visitors, built on live signals (referral source, device, location, scroll depth) rather than login or cookie data. As of September 2026, AB Tasty and VWO merged their brands under the new parent brand Wingify, with each platform still running independently during the transition.
• Rebuy Shopify-native merchandising widgets and bundles across every storefront page, cart through post-purchase.
• LimeSpot On-site recommendations plus adaptive headlines, banners, and collections for Shopify, BigCommerce, and SHOPLINE stores.

They're ready to buy something new. You keep selling last time. Let this visit lead. Don't trap them in what they already bought.

You know they'll need it again. You still make them hunt. Remind them when they're actually due.