Related products that actually fit

The "you may also like" row under a product is some of the most valuable space on a store, and most of it is wasted. The widget shows whatever it guesses is similar: a competitor's brand next to your own, a part that does not fit the product the shopper is looking at, a clearance item you would rather not promote. And almost nobody measures whether those suggestions sell. This post covers how to set rules that make the suggestions right, and how to see what they are worth.

Why most related-product widgets disappoint

A typical recommendation app looks at your catalog and your traffic and picks items that seem similar or that other people viewed. That works for some businesses. It breaks down when:

• Fit matters. Parts, accessories and refills have to match the product. A brake pad for the wrong model is worse than no suggestion.
• Brand matters. If you carry several brands, a shopper looking at one often wants more of the same.
• Some products should never be suggested. Clearance, discontinued, special order, or items with thin margins.
• You already know what goes together. The best pairings are often ones you could name from memory.

Rules in plain words

In Orkie, related products on Shopify and BigCommerce stores follow rules you set, and you can set them by asking Claude or ChatGPT in plain words:

• "Only show the same brand on product pages."
• "Only suggest parts that fit the same model," matched on the product attributes you choose.
• "Never suggest anything tagged clearance."
• "For this jacket, always show these three items first."

Rules apply to a specific spot, like the product page or the cart, so a change to one does not quietly change the others.

You see it before it goes live

Every rule change starts as a draft. The assistant reads the rules you have today, shows you exactly what would change, and saves the new version only when you say so. Turning the new rules on for shoppers is a separate step, again with your approval.

That matters because recommendations touch every product page. You want to see the change before a shopper does, and you want to be able to go back.

Measure what they sell

This is where most setups stop. The suggestions go live and nobody checks.

Because the same Orkie pixel sees what shoppers do after they click a suggestion, you can ask what the related products actually sold, spot by spot. It also tells you honestly when there is not enough data yet to say, rather than calling an unmeasured rule a failure.

When the answer comes back, the lesson goes into your company brain: which rules lifted orders, and which ones did not. The next time someone asks to change the suggestions, it starts from what already happened.

Where this fits

Related products are one of the five connections on our How it works together page: the pixel and the store sharing one set of data. It works for Shopify and BigCommerce stores. See Orkie for Shopify or Orkie for BigCommerce, or talk to us about your catalog.

---

Want this for your business? Talk to us How the pixel works