What Is Cross Selling in Ecommerce? Pairing, Placement and Measurement

What is cross selling in ecommerce?

Cross selling is the practice of recommending a product that complements what a customer is viewing or buying. For example, you suggest a memory card to someone buying a camera, or filters to someone buying a coffee machine. The goal is to meet the full need and, as a result, raise revenue per order.

In short, cross selling turns "people who buy this also need that" into a system. It does the same job as the shelf next to the till in a physical shop. However, online you can measure which product you showed, at what moment and to whom. I have worked on ecommerce projects since 2012, and cross selling is one of the most neglected levers I see, even though it often pays back fastest.

This guide goes beyond the definition. I cover which products make good pairs, where recommendation blocks belong on your site, how to keep cross selling going after checkout and, above all, how to measure its real effect.

How does cross selling differ from upselling?

Many teams mix the two up, so let me keep the distinction short. Upselling moves the customer to a higher version of the same product: a 256 GB phone instead of a 128 GB one. Cross selling leaves the chosen product alone and adds a different one next to it: a case and a screen protector for that phone.

Put simply, upselling increases the value of one decision, while cross selling increases the number of decisions. Both lift basket size, but the psychology differs. Trading up means giving something up; adding a complementary item confirms the first choice. Therefore cross-sell offers usually meet less resistance.

I explain how basket size grows in general, including the AOV formula and threshold offers, in my guide to average order value. Here I focus only on the mechanics of cross selling itself.

Why does cross selling matter so much for online stores?

Because acquiring a new customer costs a lot, while growing the basket of a visitor you already paid for costs little. You pay the acquisition cost once per order. Every complementary item added to that order spreads the same cost over more revenue. As a result, ROAS and profit improve together.

There is also a service side. A toy without batteries, a curtain without hooks or a device without the right cable pushes the customer toward a second order or a return. A good recommendation prevents that disappointment up front. In other words, good cross selling is a customer experience decision before it is a sales tactic.

Moreover, customers who buy accessories, consumables and care products tend to come back. That creates a repeat purchase loop. If you want to see this effect over time, include cross-sell data in your customer lifetime value calculation.

Which products make a strong cross selling pair?

A strong pair answers one question: does the main product feel incomplete without the suggested one? If the answer is yes, you have a good match. In practice, these are the pair types I see perform best:

• Required companions: batteries, cables, adapters and mounting kits that the product needs to work.
• Consumables: filters, cartridges, replacement heads and refills that run out on a schedule.
• Protection: cases, screen protectors, bags and care sprays.
• Style companions: a belt with a dress, cushions with a sofa, items that complete a look.
• Usability helpers: accessories or guides that help the customer get more from the product.

On the other hand, a random "popular products" row is not cross selling. Showing a food processor to someone buying shoes fills space but creates no link. Also, keep the suggested price well below the main item. A suggestion that costs almost as much as the main product creates a second big decision, and most shoppers simply skip it.

Why should you separate similar and complementary recommendations?

Because they do different jobs. Similar items offer alternatives; complementary items add to the basket. Shopify draws the same line in its product recommendations documentation, describing related products as potential substitutes and complementary products as items customers often buy in addition.

Stores that ignore this distinction put similar products on the cart page. That reopens a decision the customer has already made and brings hesitation back. Similar items belong lower on the product page, where the shopper is still comparing. Complementary items belong after the decision: at add to cart, in the cart and after the order.

So a "You may also like" block and a "Complete your purchase" block should never compete for the same spot. Give each one its own location, its own purpose and its own tracking label.

What data should you use to build cross selling pairs?

Your own order history is the most reliable source. Export six to twelve months of orders and count the product pairs that appear in the same order. Even this simple count gives you a far better starting list than gut feeling.

That said, raw co-occurrence can mislead you. A bestseller appears in almost every order, so it looks "related" to everything. Therefore ask a second question: among people who buy product A, is the rate of buying B clearly higher than the overall rate of buying B? If yes, the link is real. If not, B is simply something everyone buys.

New stores with little order data can start from category logic instead. Knowledge from your product team, customer service and return reasons is gold here. For example, if customers keep asking "was the cable not included?", that cable is your strongest cross-sell candidate. As data builds up, compare your manual pairs with real buying behaviour and update them.

Should you use rule-based or algorithmic recommendations?

Both have a place, and the right answer is usually a mix. With rules, you choose the pair yourself: "Show this memory card and this bag next to this camera." With an algorithm, the platform or a recommendation engine picks the pair from behaviour data.

Rules work very well for stores with a small catalogue and high margins. You control exactly what appears, so out-of-stock or unprofitable items never get pushed. In contrast, a catalogue of thousands of products makes manual pairing impossible. There, an algorithm becomes necessary.

My advice is simple. Set pairs manually for your top 20 to 50 bestsellers and most profitable products, then let the algorithm handle the rest. Also give the algorithm some limits: remove out-of-stock items, very low margin items and high return items from the recommendation pool. That way you combine the speed of automation with the control of rules.

How should cross selling look on the product page?

The product page is where cross selling starts most gently. The shopper is still evaluating, so the suggestion should read like information, not pressure. Descriptive headings such as "Often used with this product" or "What you need for setup" beat vague ones such as "Picked for you".

I follow two rules for placement. First, the cross-sell block must never compete with the add to cart button; the main decision always comes first. Second, keep it short: two to four products are enough. A ten item carousel creates choice fatigue and scatters attention.

Next, add a one click add option to each suggestion. Sending the shopper to another product page pulls them away from the main item and lengthens the path back. For the wider structure of the page, see my ecommerce product page guide, where I also discuss where the recommendation block sits in the page hierarchy.

Should you show recommendations at the add to cart moment?

Yes, because this is the most productive window for cross selling. The customer has just decided, and their mind is still on that product. A small cart drawer or confirmation panel that opens at that moment delivers "add this too" in the most natural way.

However, do not smother that moment with a pop-up. A full screen window that is hard to close pulls the shopper off the checkout path. Instead, use a slide-in mini cart or a simple box under the product. Show three items at most, each with a price and a one click add button.

Also check one more thing: can the customer continue without adding anything? The "View cart" and "Continue shopping" buttons should stand out more than the suggestions. Otherwise, cross selling can turn into cart abandonment. I explain how I balance this in my article on reducing cart abandonment.

How does cross selling work on the cart page?

The cart page is where customers review their order. Here cross selling has two jobs: reminding them of a missing piece and, where relevant, helping them reach the free shipping threshold. For instance, if the basket sits just below the threshold, highlighting complementary items in that small price gap genuinely helps the customer.

On the other hand, the cart page is where the main sale is most at risk. So place the recommendation block below the order summary and the checkout button. The shopper should see the total and "Proceed to checkout" first; the suggestions appear only when they scroll.

The threshold itself needs its own calculation. If you set it in the wrong place, cross selling ends up subsidising shipping rather than raising profit. I walk through that maths in my guide on how to calculate a free shipping threshold; plan your cart suggestions around that number.

Is it risky to show offers during checkout?

Yes, in most cases it is. A customer at checkout is about to enter card details, and anything extra on the screen breaks focus. That is why I do not put a classic product carousel on the checkout page.

Still, one exception works well: a small option directly tied to the order that the customer can add with one click. Gift wrapping, an extended warranty or a small part the product always needs fall into this group. It should be as plain as a checkbox and should not change the layout.

Also, never use pre-ticked boxes. Customers who buy extras without noticing may lift revenue in the short term, but that comes back as returns, complaints and lost trust. It is also a consumer law risk in many markets. In short, if checkout includes a cross-sell option at all, it must be clear, optional and fully transparent.

How do you cross sell after the purchase?

Post purchase is the least used and least risky area for cross selling. The main sale is already safe, so you cannot lose it. Moreover, the customer still carries the excitement of the purchase. You can use three main channels:

1. Thank you page: offers a one click addition to the order or a small incentive for the next order.
2. Order and shipping emails: pair usage tips with a reminder about the matching accessory.
3. Refill reminders: for filters or cartridges, a message that arrives before the typical usage period ends.

The key here is timing. Suggesting a care spray before the product even arrives feels early. In contrast, a "How are you getting on with your new product?" email a few days after delivery creates a natural cross-sell moment. So build your post purchase flow around the product's usage cycle.

Which recommendation spot does which job?

Comparing the placements in one table makes it clear what to show where. The table below is the starting framework my team and I use on ecommerce projects:

Notice the pattern: the closer the customer gets to paying, the fewer suggestions you show and the tighter they link to the main item. Once the purchase is complete, the space opens up again.

How should you price and discount cross selling offers?

The first rule: you do not have to discount a complementary item. A good pair sells at full price, because the customer already needs that product; you are simply reminding them. Keep discounts for cases where the pairing alone is not enough.

If you do discount, calculate the margin first. Complementary items are usually cheap, so even a small percentage can wipe out most of the profit. For example, an extra discount on an accessory with a thin margin grows the basket but not the profit. You can check percentages and net prices quickly with the discount calculator.

Also, do not confuse "second item discount" with cross selling. A second item deal usually sells one more of the same product or category, while cross selling answers a different need. I analyse the margin impact of that mechanic in my article on second item discounts and margin. Sets sold at a single price are a separate topic again.

What are some cross selling examples by category?

The pairing logic is the same in every sector, but the products change. Treat the examples below as general patterns to get you thinking about your own catalogue:

• Electronics: a case and charger for a phone, a bag and mouse for a laptop, a memory card and spare battery for a camera.
• Fashion: a belt or bag with a dress, care spray and socks with shoes, a scarf with a coat.
• Beauty: a sponge or brush with foundation, the matching conditioner with a shampoo.
• Home: pillowcases with bedding, filters and coffee with a coffee machine, a pot and soil with a plant.
• Pets: a bowl with food, an ID tag and waste bags with a lead.

Every example shares one trait: the suggested item makes the main product easier to use or completes it. For instance, suggesting products from the same range works well in beauty, because the customer already trusts the brand. In fashion, outfit logic leads; therefore showing the suggested piece with the main item in photos makes the pair concrete.

How do you write copy for cross-sell blocks?

The heading and short description of a recommendation block matter as much as the products themselves. Shoppers read the heading first and then look at the items. So the heading should explain in one line why the suggestion is there.

Use need-based headings instead of generic ones. For example, write "Cartridges that fit this printer" rather than "You might like". The first sounds like a sales pitch; the second actually helps. Also, a one line reason under each suggestion works well: a factual note such as "Charger not included in the box" outsells the most persuasive copy.

That said, avoid pressure language. Phrases such as "Only 3 left" or "Don't leave without this" burn trust fast when they are not true. In short, cross-sell copy should sound like advice from a good shop assistant: plain and honest. You measure copy with A/B tests too; a heading change is one of the cheapest tests you can run.

How do you measure cross selling performance?

Measurement starts by separating your recommendation blocks. The ecommerce setup in Google Analytics 4 offers a ready structure for this: you send view_item_list when a list appears and select_item when a user picks an item from it, and both events carry item_list_id and item_list_name. The details sit in the Google Analytics ecommerce measurement guide.

In practice, you give each block its own list name, such as "pdp_bought_together", "cart_addon" or "thank_you_page". That way you see how much each block sells on its own. These are the core metrics to track:

• Recommendation view rate: the share of sessions that see the block.
• Add rate from recommendations: how many viewers add at least one suggestion.
• Orders with a cross-sell item: the share of orders containing at least one suggested product.
• Revenue and profit per recommendation: contribution on margin, not just sales.
• Main product conversion rate: does the block hurt the main sale?

The last metric is the one teams forget most. If a block sells extras but lowers the main conversion rate, the net effect can be negative. You can compare rates with the conversion rate calculator.

How do you test the true impact of cross selling?

Counting sales that came through recommendations does not show the true effect. Some of those items would have sold anyway; the block may have simply claimed them. Therefore you need a control group to measure the incremental lift.

The cleanest method is a simple A/B test. Show the recommendation block to one group of visitors and hide it from another. Then compare revenue per order, profit and main product conversion between the two groups. If the difference is statistically significant, cross selling genuinely adds value. You can check significance with the A/B test calculator.

However, with low traffic a test takes a long time. In that case, limit tests to big changes: whether the block exists, where it sits or what type of item it shows. Small details such as button colour rarely reach significance on low traffic. Also run each test for at least two full weeks so that weekday and weekend behaviour both show up.

What are the most common cross selling mistakes?

I see the same mistakes again and again in store audits. Most of them are about logic, not technology:

• Suggesting out-of-stock items: hitting "add" and seeing "sold out" damages trust.
• Suggesting what is already in the cart: it signals carelessness.
• Suggesting incompatible items: a case for the wrong model or a filter in the wrong size turns into a return.
• Showing the same list everywhere: if the product page, cart and email repeat the same four items, customers stop noticing.
• Looking only at revenue: a block that pushes low margin items can grow the basket and shrink profit.

Compatibility errors in particular get expensive. So for technical products, tie compatibility data to your recommendation rules; if the model, size or version does not match, the block should not show that item at all.

When does cross selling not work?

Cross selling does not deliver the same result in every store. If you sell a single product, such as one software licence or one device, there may be no meaningful companion to suggest. In that case, rather than filling the space, consider a genuinely related option such as setup, a service plan or a warranty.

Also, stores selling very cheap items may see limited gains. If the profit from suggested items does not cover the cost of the recommendation tooling, the numbers do not add up. Likewise, in categories where shoppers buy with urgency and a single goal, such as a specific spare part, extra suggestions just slow them down.

Still, these cases are exceptions. In most stores the problem is not cross selling itself but weak pairs and poor placement. So before you remove the block, change the pairs and the position and test again.

What should you do in the first month?

If you are starting from scratch, do not try to build everything at once. This is the order my team and I follow with the stores we advise:

1. First week: pull the product pairs from your order data and build a pairing list for your 20 most profitable products.
2. Second week: set up blocks on the product page and at add to cart, each with its own list name.
3. Third week: rework the cart page around your free shipping threshold and add one offer to the thank you page.
4. Fourth week: read the first data, check main product conversion and start an A/B test on the biggest block.

This order gives you the fastest impact with the least risk. After that, build the post purchase email flows in month two using the data from month one. That way you write them knowing which products people really buy together.

If you want to rebuild your store's recommendation logic on solid data, my team and I can plan the whole process with you, from the pairing list to the tracking setup, as part of our ecommerce consulting. And if the recommendation blocks need design changes, our web design team can rebuild them without breaking your site's flow.