Cross-selling accounts for roughly 21% of ecommerce revenue on average, but the majority of Shopify stores barely get a fraction of that. The usual problem is execution. Products that don't match what's in the cart, widgets crammed into the wrong page, suggestions that show up after the customer has already lost interest.
This guide covers 10 specific cross-selling tactics you can put in place this week. If you need the fundamentals first, our guide to what cross-selling is covers definitions, psychology, and types. And if you're choosing between tools, our breakdown of Shopify cross-sell apps compares what's available.
1. Pair by use case, not by collection
Most default recommendation setups pull products from the same collection. A customer looking at a floral midi dress sees four more dresses. She might swap her original pick for one of those, but she's unlikely to add any of them. Alternatives compete with the original purchase. Cross-sells complement it.
Pair products by how they'll be used together. That midi dress should surface a cropped denim jacket for layering and a pair of block-heel sandals. The question to ask for every cross-sell: "Would someone buy this alongside the original, or instead of it?" If the answer is "instead," it doesn't belong in a cross-sell widget.
2. Keep cross-sell pricing under 25% of the cart value
A $16 hair clip suggested alongside a $90 silk blouse feels effortless. A $75 handbag next to the same blouse triggers an entirely separate purchase evaluation. Cross-sells convert best when they feel like easy additions to a decision the customer has already made.
For most stores, keeping suggested items below 25% of the current cart value hits the sweet spot. Anything above that and the customer pauses to deliberate, which increases the chance she removes items or abandons the cart altogether.
3. Put your highest-converting widget in the cart drawer
Product pages get more traffic, but the cart drawer gets more buying intent. A customer who has opened the cart drawer is closer to checkout than someone still browsing product pages. One well-chosen cross-sell suggestion here converts at 2-3x the rate of the same suggestion on a product page.
Keep it to one or two items, not a scrollable carousel. The cart drawer is small real estate, and cluttering it with six suggestions turns a buying moment into a browsing moment.
4. Use "Complete the Look" instead of "You may also like"
"You may also like" is vague and customers have learned to scroll past it. "Complete the Look" tells the shopper exactly what this section does and triggers a visual gap: she sees the full outfit, notices her cart only has one piece, and feels the pull to finish it.
For a customer viewing a fitted blazer, "Complete the Look" might show a silk cami, tailored trousers, and pointed-toe mules. The framing matters as much as the product selection. PersonalizerAI's recommendation engine builds these pairing models from your store's actual co-purchase and browsing data, so the combinations reflect how your customers actually shop rather than generic tags.
5. Cross-sell consumables and care products after purchase
Post-purchase is the lowest-risk placement for cross-sells because the original order is already locked in. Use it for items that protect or extend the product the customer just bought: leather conditioner after a handbag purchase, fabric shaver after a cashmere sweater, stain repellent after white sneakers.
These convert well because the timing matches the customer's mental state. She just invested in something. A $12 care product that makes it last longer feels like a smart addition, not an upsell.
6. Set different cross-sell strategies by page type
A single recommendation rule across your entire store wastes most of its placements. Product pages, collection pages, cart pages, and post-purchase pages each catch the customer at different points in the decision process.
On product pages, show complementary items that complete an outfit or set. In the cart, show low-friction add-ons priced to feel effortless. Post-purchase, show care products or accessories the customer won't want to wait for a second shipment to get. Each surface has its own dynamics, and stores that treat them as separate cross-sell contexts, each with its own product selection and price ceiling, outperform stores that run the same logic everywhere.
7. Let co-purchase data pick the products, not your instincts
Merchandising intuition is useful for 30 SKUs. At 300+, it breaks down. A merchandiser might pair a navy blazer with a white oxford because it makes visual sense. But the data might show that customers who buy that blazer actually add a specific grey knit tee at 4x the rate of the oxford.
AI-powered cross-sell engines trained on your store's order history surface these non-obvious patterns automatically. PersonalizerAI's models learn from actual purchasing and browsing behavior across your catalog, catching seasonal shifts and product-to-product relationships that static rules miss entirely.
8. Cap product page cross-sells at three items
Six or eight recommendation slots on a product page is clutter that creates decision fatigue. The customer has to evaluate each suggestion, and when the cognitive load gets too high, she evaluates none of them.
Three targeted suggestions are enough. A complementary product in the same use case, an accessory, and a care item. When those three are relevant, the take rate per suggestion goes up because the customer actually evaluates each one instead of scanning and moving on. More slots with weaker relevance produces worse results than fewer slots with strong matches.
9. Time email cross-sells to the delivery window
Most merchants send cross-sell emails immediately after purchase or not at all. The highest-converting window is 2-5 days after delivery, when the customer has the product in hand and is actively using it.
A customer who received a linen jumpsuit two days ago is more receptive to a suggestion for a woven belt and espadrilles than she was the moment she clicked "Place Order." She can see the outfit taking shape. If you're running post-purchase email flows, tie the cross-sell trigger to your average shipping time, not the order date.
10. Track cross-sell take rate by placement, not just overall
An overall cross-sell conversion rate of 4% tells you almost nothing useful. If your product page widgets convert at 1.5% and your cart drawer converts at 8%, you need to fix product page relevance, not celebrate the blended average.
Break out take rate, incremental AOV, and items per order by each placement. This is the only way to know which surfaces are working and which are adding noise. Click-based attribution matters here too. Make sure your analytics track which specific widget a customer clicked before purchasing, not just which pages had recommendation widgets on them. For more on the metric that ties it all together, see our guide on revenue per visitor.
What to tackle first
If you're starting from scratch, tips 1, 4, and 3 give you the fastest path to measurable results. Get product-page pairings right by matching use cases instead of collections, frame them as "Complete the Look," and add a single well-targeted suggestion to the cart drawer. Most stores see AOV movement within two to three weeks from these three changes alone.
If you already have cross-sells running but they're underperforming, tips 7 and 8 address the most common root cause: too many suggestions based on merchandiser guesses instead of actual customer behavior. Switching to data-driven product matching and reducing the number of suggestions per placement tends to lift take rates immediately.
For the full strategy behind these tips, our cross-selling guide covers the psychology, types, and measurement framework. And our upselling tips covers the companion set of tactics for moving customers up the price ladder.