Marketing scenario
What this scenario gives you
More clicks into relevant products, better cross-sell, and higher order value from post-exit communication.
For recommendations after category browsing, after purchase, after newsletter signup, with abandoned carts, and for returning shoppers.
When it is worth using
- Stores with broad catalogs and choice complexity.
- Brands growing cross-sell and up-sell.
- Teams personalizing email without content overload.
When to skip it
Do not send recommendations without intent data or right after a purchase when similar products would feel like replacement pressure.
Implementation plan
How to set it up
Choose the recommendation source
Recommendations can come from viewed category, cart, last purchase, bestsellers, or season. One email should use one logic.
Order products by intent
Show the product closest to shopper behavior first, then alternatives, and only then add-ons.
Measure clicks by position
Check which email positions actually sell. This helps shorten the message and remove weak blocks.
Personalization
Relevance matters more than product count
Three relevant products can sell better than a long shelf. The logic should be visible: similar to the viewed model, matches the last purchase, often chosen with this category. Otherwise recommendations feel like leftover catalog.
- Do not mix too many categories.
- Show the reason for the recommendation.
- Use images and prices that support comparison.
Cross-sell
Complementary products must fit the moment
The contact moment changes the recommendation. After purchase, do not push a similar product if the shopper just bought one. Show an accessory, refill, care item, or something that solves the next problem after first use.
- After purchase, use add-ons and replenishable products.
- Before purchase, show alternatives and bestsellers.
- In cart context, watch the free-shipping threshold.
Quality
Recommendations should not look like catalog leftovers
An automated shelf needs quality filters. A product without a proper image, a variant with broken size availability, or an offer unrelated to history can ruin the email even if the algorithm technically matched something.
- Exclude unavailable products.
- Maintain minimum image and price quality.
- Do not recommend an already purchased product as the first option.
Measurement
What to measure after launch
Evaluate the scenario by shopper behavior and cart impact, not by impressions alone. These metrics help you see whether the campaign supports revenue or only creates activity.
- clicks by product position
- add-to-cart after email
- AOV from recommendations
- cross-sell and up-sell sales
