Upsell, cross-sell and bundles: raise the basket without slowing the sale

Upsell, cross-sell, bundling: three different levers

On Google France, “upsell cross sell” is searched 480 times a month, “vente groupée” (bundling) 70 times. The three terms cover distinct moves, which do not belong in the same place:

• Upsell (trading up): offer the better version of the product being viewed. Example: the complete guide with its templates rather than the guide alone.
• Cross-sell: offer a product that complements the one chosen. Example: the case with the device, the settings logbook with the plan.
• Bundling: sell several products together at one combined price. Example: a bundle of three guides on the same topic.

Marketing research also distinguishes product bundling (you create a new offer) from price bundling (you sell unchanged products together): Stremersch and Tellis (2002) put forward 12 proposals to choose the right strategy for the context.

Why the average order value decides profitability

Every order carries fixed costs: the fixed part of the payment fee (1.5% + €0.25 at Stripe for a standard European card), fulfilment, customer service and, above all, the cost of acquiring the visitor. On a small basket these costs eat the margin; on a larger basket they are spread out.

That is the reason for our decision of 12 October 2026 on our own sites: bundles from €49 to €149 that group the small products, rather than sales at €5 to €12. The bundle becomes the recommended offer, and the single product stays available for anyone who only wants one.

Build a bundle that sells

A good bundle answers a use, not leftover stock. Our method:

1. 1
   Start from a complete need: “everything to get started”, “the exam kit”, “the first-year equipment”. The buyer must understand in one line what they get.
2. 2
   Three or more products that work together: a bundle of unrelated products reads like a clearance sale.
3. 3
   A clear combined price, compared with the real total of the products bought separately from your own catalogue.
4. 4
   Few bundles: two or three at most. The experiment by Iyengar and Lepper (2000) shows that people buy more when offered a limited choice of 6 options rather than 24 or 30.
5. 5
   A page of its own: title, detailed contents, a photo of the whole set, structured data, just as for a product.

Where to place each offer in the journey

Timing matters as much as the offer. What works, and what drives buyers away:

• On the product page: the upsell (better version) and the bundle, shown as options of the same choice, next to the main button.
• After payment, on the thank-you page: the cross-sell. The buyer has paid, trusts you, and a complementary product in one button no longer puts the first sale at risk.
• In the delivery email: one useful add-on, once, with no insistent follow-up.
• Never between the “Buy” button and the payment page: a pop-up of offers at that moment adds a step and a moment of doubt.

On our sites, the add-on is offered after the purchase; since 12 October 2026, a repeat purchase takes one button, the card being stored by the payment provider with the buyer’s consent. And since 17 October 2026, the cart only appears from three products: below that, buying directly stays the shortest path.

The mistakes that lower sales

Each of these mistakes shows in the numbers: an average order value that does not move, a conversion rate that drops, refunds that rise. Hence the value of measuring before and after every change.

• Offering too much: five suggestions on a product page scatter attention and push the decision back.
• Offering unrelated items: a randomly generated “you may also like” damages trust.
• Pre-ticking an extra product: the buyer must choose; adding an item by default is a practice to ban.
• Inflating the comparison price: the “bought separately” total must match your real prices; a fake reference price counts as a misleading discount.
• Forgetting the margin: a bundle discounted too heavily can earn less than selling the product alone.

Measure the effect

Three indicators are enough: average order value, conversion rate and the share of orders that include a bundle or an add-on. Change one thing at a time and compare periods of the same length. If the basket rises but conversion falls by as much, you have gained nothing.