Bundling
Date: 2026-08-16
Selling several things as one. It raises basket size, obscures per-item comparison, and moves slow stock — and it destroys margin whenever the bundle discount exceeds the incremental contribution it creates.
What it is
Bundling offers multiple products together, usually at a price below the sum of the parts.
Three forms:
| Form | Offer | Use |
|---|---|---|
| Pure | Only available as a bundle | Rare in retail; suits kits and sets |
| Mixed | Available separately or bundled | The standard case |
| Cross-sell | Suggested at the point of purchase | Not strictly a bundle, same intent |
The margin arithmetic
The calculation that decides whether a bundle works:
sold separately
item A £30 contribution £9.00
item B £20 contribution £6.00
─────────────────
total £50 contribution £15.00
bundled at £45 (10% off)
revenue £45 contribution £10.00
If the customer would have bought both anyway, the bundle costs £5.00 for nothing.
It only pays when it changes behaviour:
customer would have bought A alone
A alone contribution £9.00
bundle at £45 contribution £10.00
─────────────────
gain +£1.00
A bundle is worth £1.00 when it adds an item and costs £5.00 when it doesn’t — so it needs roughly five incremental buyers for every one who’d have bought both anyway. Which means the whole question is what proportion of bundle buyers were going to buy both — the same cannibalisation question as any discount — Discount Impact on Margin, Incrementality Testing.
Where bundles genuinely help
- Attaching a high-margin item to a low-margin driver. The classic and best use — the driver brings the customer, the accessory carries the margin — Basket Composition
- Moving slow stock without discounting the fast-moving item on its own
- Obscuring comparison. A bundle isn’t directly price-comparable with a competitor’s single item, which reduces price salience — Price Anchoring
- Reducing decision effort. A “complete set” removes a choice, which converts better than presenting the components
- Raising AOV towards a shipping threshold with items that carry margin rather than padding
Where they don’t
- Discounting things people were buying together anyway. Pure margin loss
- Bundling two slow items. Two things nobody wants aren’t more wanted together
- Deep bundle discounts that reset the reference price for the components — customers stop buying A alone at £30 once they’ve seen it at £27 inside a bundle
- Complex bundles. If the saving needs arithmetic to see, it isn’t influencing anyone
Designing one
- Pair a driver with an attachment, not two equals. The driver creates the visit; the attachment carries the margin
- Show the component prices alongside the bundle price. That’s what makes the value visible — and it’s an anchoring mechanism that makes no historic-price claim, so it avoids the reference pricing regulatory question entirely
- Keep the discount shallow. The convenience and the framing do much of the work; the discount often needn’t be large
- Protect the standalone price of the driver, or you’ve cut its price by stealth
- Check stock on both components. A bundle with one item unavailable is worse than no bundle — Stockouts and Availability
Measuring it
- Contribution per order, not AOV. A bundle that raises basket size and lowers contribution is a common outcome — Average Order Value
- Attach rate before and after, to see whether it’s creating attachment or capturing it
- Standalone sales of the components. A fall means cannibalisation
- Return rate. Bundles are returned more often when one component disappoints, and a partial return of a bundle is an operational and margin problem — Return Rate and Reverse Logistics