Winning a new customer is often five times more expensive than getting extra revenue from an existing one. Yet most stores pour all their energy into driving traffic, leaving on the table the additional revenue that could come from a customer who's already on the site and ready to buy. When done right, upsell and cross-sell are the most efficient growth channel there is — raising average order value without spending a single cent on ads.
The difference between upsell and cross-sell
The two concepts are often confused, but they speak to different customer behaviors. Upselling means recommending a higher tier of the exact product the customer is looking at, or a bigger package: showing a shopper eyeing a 128 GB phone the 256 GB version, for example. Cross-selling means recommending a different product that complements the main one: showing a case or screen protector to a customer looking at a phone.
For both to work, there's one shared rule: the recommendation has to serve the decision the customer is making right now. An irrelevant "you might also like this" suggestion is a distraction that drags down conversion; a well-timed, contextually relevant recommendation, on the other hand, reinforces the purchase decision.
To draw a clear line between the two tactics, here's a side-by-side comparison:
| Dimension | Upsell | Cross-sell |
|---|---|---|
| Definition | Recommending a higher-tier or larger version of the same product | Recommending a different product that complements the main one |
| Typical placement | Product page, variant selection | Cart, post-checkout page |
| Example | Suggesting a 256 GB phone instead of 128 GB | Suggesting a case and screen protector for a phone |
| Primary effect | Raises unit price and margin | Increases the number of items in the cart |
Smart placement on the product detail page
The product page is the most natural home for upsells, because the customer is still in decision mode. The most effective placements are:
- A small card right below the price and "Add to Cart" button, clearly showing the price difference for a higher tier in the same category,
- An "others who bought this also bought" strip further down the page, driven by past order data,
- A comparison at the variant-selection step that highlights the per-unit price advantage of a bigger package (a three-pack instead of a single item, for instance).
The critical detail here is the number of recommendations: staying between three and five products makes it easy to decide without triggering choice fatigue. A strip crammed with dozens of suggestions is less effective than a single strong one.
Complementary products on the cart page
Once a customer adds a product to the cart, purchase intent is clear — this is the ideal moment for cross-selling. Showing complementary products that are frequently bought alongside that item, either on the cart page or in a small pop-up right after adding to cart, lets the customer add them with a single extra click.
Threshold logic works well here too: a message like "Add 150 TL more and get free shipping" both drives additional sales and gives the customer a concrete benefit. When paired with the right products, this kind of shipping-threshold nudge also lowers cart abandonment, because the customer is already inclined to spend more to avoid the shipping fee. We covered ways to win back visitors who abandon their cart entirely in a separate article.
"The best upsell recommendation doesn't feel like a sale — it feels like advice that completes the decision the customer was already about to make."
Checkout and the post-purchase opportunity
Recommending additional products at checkout is a delicate balance: get it wrong and you distract the customer, lowering your checkout completion rate. So keep the checkout screen down to a single, small suggestion — usually a low-priced add-on that requires no real decision (gift wrapping, extended warranty, a small accessory) works best.
The real opportunity is the thank-you page right after checkout is complete. The customer has already paid and made their decision; an offer here along the lines of "add this to your order and we'll ship it together" creates an extra micro-conversion without any purchase anxiety. The same logic can be repeated in a post-order email.
Thinking through recommendation placement along the customer journey, in order, makes it clear which tactic belongs on which page:
- Capture the upsell opportunity on the product page by showing a higher tier or bigger package.
- Offer complementary products and a shipping-threshold incentive in the cart.
- Create a micro-conversion at checkout with a single, low-risk add-on that doesn't distract.
- Repeat the post-purchase offer on the thank-you page and in the follow-up email.
Building ready-made bundles is a separate technique that reinforces this same sequence; we go into detail on bundle strategy in our article on product bundling.
Personalization and the recommendation engine
Static "you might also like" lists lose their effectiveness past a certain point. The real difference comes from computing recommendations dynamically, based on the customer's past behavior and the current cart contents. Even a simple rule engine — along the lines of "X% of customers who bought from this category also bought this product" — clearly outperforms static lists. We went deeper into building a recommendation engine for cross-selling in the cart in an article dedicated to that topic.
When building your recommendation engine, combine three data sources: co-purchase history (which products frequently appear together in the same order), category relationships (complementary categories), and margin data (priority can be given to higher-margin products). An engine that ignores margin may boost revenue while contributing little to profitability.
Measuring and testing
A single metric isn't enough to gauge the impact of your upsell and cross-sell efforts. There are at least three indicators worth tracking together: average order value (AOV), the click-through rate on recommendation strips, and the conversion rate of a recommended product from cart addition through to checkout. If you don't track all three together, a recommendation that gets clicks but doesn't convert can look like a success when it isn't.
Run regular A/B tests on placement, copy, and the number of recommendations shown. For instance, "Others also bought this" versus "Just for you" may perform differently depending on the product category — there's no single right answer; you need to test it with your own customer base.
Quick checklist
- Are upsell and cross-sell zones separately defined on the product page?
- Is there a recommendation tied to a shipping threshold on the cart page?
- Is the checkout recommendation limited to a single, low-risk product?
- Is a post-purchase offer shown on the thank-you page?
- Are recommendations dynamic, based on past order data, or are they fixed?
- Is margin data factored into how recommendations are ranked?
- Are AOV, click-through rate, and conversion rate tracked together?
The hard part of this setup isn't the idea — it's the infrastructure: calculating which products sell frequently together, surfacing that without delay while the page loads, and ranking it by margin rules is not something you can manage by hand. Şimşek Software's e-commerce platform delivers recommendation modules, fed by order history, as ready-to-use components for the product, cart, and post-checkout pages — you just set the rules: margin priority, shipping threshold, number of recommendations.