Will a bundle cannibalise your single SKU? A decision framework for D2C brands
Bundles lift average order value and quietly eat single-unit sales. Here is how to frame the bundle decision, which segments to explore, what cannibalisation actually costs, and the one test that tells you whether to launch.
Published 15 September 2026 · 3 min read · By the Precheck team
Key takeaways
Cannibalisation is only a problem if the bundle margin per unit is lower than what it replaces. Do that arithmetic first.
Repeat buyers are the natural bundle audience. First-time visitors usually need a single before they will buy three.
Offer the bundle where the segment lives: post-purchase for new customers, upfront for repeat buyers.
Test one bundle price on one segment for two purchase cycles before rolling it out.
A three-pack at ₹1,499 against singles at ₹549 is a decision every D2C brand faces eventually. The bundle looks like a clean win: higher order value, fewer shipments, a customer stocked for months. The risk is invisible on the launch dashboard. Repeat buyers who would have paid ₹549 three times over the quarter now pay ₹1,499 once. Revenue per order goes up. Revenue per customer can go down.
Frame it as arithmetic, then as a decision
Start with two numbers.
- Margin per unit on singles. Price minus landed cost minus per-order fulfilment, divided by one.
- Margin per unit on the bundle. Bundle price minus three units' cost minus one order's fulfilment, divided by three.
If the bundle's margin per unit is higher, cannibalisation is not a threat; it is the point. Every single-unit buyer who switches makes you more money per unit and ships less. If it is lower, every switch costs you, and the bundle has to earn its place by bringing in buyers who would not have bought three singles.
Then frame the options you would actually choose between:
- Launch the bundle at ₹1,499 to everyone.
- Launch at ₹1,299 to repeat buyers only.
- Keep singles, offer the bundle only after a first purchase.
- No bundle.
Explore by segment
In our illustrative simulations, three segments behave differently enough that the "right" bundle decision changes with the audience.
- Repeat buyers already stock up. A bundle gives them a reason to buy ahead, and price matters less than the sense of a fair deal. The risk here is pure cannibalisation, so this is where the arithmetic decides.
- First-time visitors will not commit to three units of something they have never tried. Showing them a bundle first can lower conversion. The bundle belongs after the first order.
- Lapsed customers rarely respond to a bundle at all. They left for a reason that was not price, and a bigger box does not address it.
A simulation can rank these for your own audience and surface the barrier in each. Precheck does this with data-informed simulations; we are open that predictive accuracy has not been established yet, so use the output to pick the test, not to replace it.
Run the one test
The test almost always looks like this: bundle at one price, one segment, two purchase cycles, against a control group on singles.
- Segment: repeat buyers, because that is where cannibalisation lives and where the bundle has the best chance.
- Price: the one your exploration flagged as most uncertain, often the lower one.
- Window: long enough to see whether bundle buyers come back. A bundle that lifts order value and then goes quiet for four months is not a win.
- Read: revenue per customer over the window, not revenue per order.
What to do with the answer
If the bundle wins on revenue per customer among repeat buyers, roll it out to that segment and offer it post-purchase to new customers. If it loses, you have learned something cheaper than a launch: your repeat buyers were already buying three, and the bundle only discounted them.
Either way, keep the record. A decision, the segment, the price, and the outcome is exactly the material that lets you calibrate any simulation against your own history before you trust it with the next one.
Questions people ask
- What is bundle cannibalisation?
- It is when customers who would have bought single units at full price buy the bundle instead, at a lower price per unit. It is a problem only when the bundle's margin per unit is lower than the singles it replaces, and the bundle does not bring in enough new buyers to make up the gap.
- Should the bundle be cheaper per unit than singles?
- Usually, a little. The discount buys commitment: three units instead of one, and a longer gap before the customer considers a competitor. Size the discount so the margin on the bundle beats the expected margin on the singles it replaces.
- When should a D2C brand not launch a bundle?
- When most sales come from first-time buyers who have never tried the product, or when singles are already margin-thin. In both cases a bundle mostly moves existing revenue around.