Where to Put Product Recommendations: The Page-by-Page Ecommerce Placement Guide

The direct answer

Place recommendations where the shopper has context and room to act

The same product set can feel useful on one page and irrelevant on another. A belt beside trousers has an obvious relationship on a product page. The same belt beneath an empty search for “waterproof hiking jacket” is not an answer. Placement changes the shopper’s question, so it changes the recommendation’s job.

Use five tests for every placement: context, relationship, timing, friction, and measurement. The shopper should know what page they are on, why the products are grouped, when the module appears, how to ignore or use it, and which outcome will tell you whether it helped.

Placement 1

Use the product page for comparison and completion

The product page has the strongest anchor because the shopper is already looking at a specific product. Shopify supports related and complementary product sections on product pages through compatible themes and Search & Discovery configuration. The two modules should still have different jobs.

Related products support comparison. Show credible alternatives with a meaningful difference: another fit, material, color family, price point, or use case. Complementary products support completion. Show something the shopper may need to use, protect, install, or maintain the selected item.

Place the module after the shopper has seen the main title, images, price, options, shipping, and add-to-cart action. If the recommendation is a required compatibility item, it may deserve a more prominent location, but the relationship should be explicit and the primary product should remain understandable.

Baymard’s product-page research treats the product page as a central place where shoppers decide whether to purchase. That makes the recommendation card part of the evaluation experience. A wrong price, missing option, or unavailable candidate can damage the main product’s credibility.

Placement 2

Use collection pages to open an adjacent discovery path

Collections give shoppers a category anchor rather than a single product anchor. A recommendation module can help them move sideways into a related category, a project, or a complementary group. It should not make the main collection harder to scan.

For a home office collection, a small “Complete the workspace” module may lead to monitor arms or cable management. For a running collection, “Build your training kit” may lead to socks, hydration, and recovery products. The content should reflect the category’s shopping mission, not simply repeat the best sellers already visible in the grid.

Baymard’s product-list research emphasizes that list design, filtering, sorting, and product presentation work as one finding path. Treat a recommendation row as part of that path. Test whether it helps the shopper move forward or simply adds another block to scroll past.

Placement 3

Use search recommendations for recovery, not camouflage

Search is explicit intent. If the shopper searches “blue linen shirt,” a recommendation row should not replace the exact results or quietly mix unrelated products into the result list. Keep the primary answer clear. Use recommendations when the query is empty, the exact result is unavailable, or the shopper needs a labeled adjacent path.

The recovery heading matters. “You may also like” is vague after a zero-result query. “Try travel bags” gives the shopper a category path. “Similar products” gives a comparison path. “Complete your camera kit” gives a mission path. Write the heading after you decide the relationship.

For search analytics, keep recovery clicks distinct from result clicks. Otherwise, a high click rate can hide a high zero-result rate. The shopper may be clicking because the primary search failed, not because the recommendation experience is healthy.

Placement 4

Use the cart for small, optional add-ons

Cart recommendations can help when the item has a direct relationship to what the shopper already chose. They should be easy to understand and easy to ignore. A compact add-on with a direct add action is safer than a large carousel that sends the shopper back into browsing.

Be especially careful with checkout. Baymard’s cross-sell research describes strong frustration when users encounter a separate cross-sell step during checkout. That evidence supports a conservative rule: never require the shopper to reject an offer before continuing, and measure checkout completion rather than offer clicks alone.

If an add-on requires a size, color, or compatibility decision, send the shopper to a focused product page with the cart preserved. Do not silently add an arbitrary variant because it makes the implementation easier.

Placement 5

Use post-purchase recommendations for the next use case

After purchase, the relationship changes. The shopper no longer needs the product they just bought. They may need a replacement, a compatible item, care supplies, an accessory, or the next product in a project. The order gives you stronger context than a generic home page, so use it.

Avoid recommending the same item again unless replenishment is the explicit job. A skincare store can recommend a refill after the likely usage period. A parts store can recommend maintenance kits. An office furniture store can recommend a cable tray or monitor arm after a desk purchase. Label the reason and respect timing.

Placement 6

Design mobile placement around the thumb and the primary action

Mobile changes the cost of every recommendation. The viewport is smaller, the keyboard and sticky purchase controls compete for space, and a horizontal carousel can hide the heading or price after one swipe. Test mobile as its own placement, not as a desktop row that happens to shrink.

At 320px, check that the section heading wraps cleanly, product names do not disappear, prices remain visible, touch targets are large enough, and the primary product action stays easy to find. If the recommendation appears below a long product description, consider whether it is too late to help. If it appears above the main action, consider whether it is too early.

Audit

Run the 30-minute placement audit

Use one anchor product, one collection, one search query, one cart, and one post-purchase example. For each surface, write the job, heading, candidate relationship, placement boundary, and success event.

The output should be a placement matrix with one repair for each surface. If the product page needs better card data, fix that first. If search recommendations are mixed into primary results, separate the modules before judging candidate quality. If cart placement hurts completion, remove the detour.

Frequently asked questions