Product cannibalization: when a new item takes your own sales

A brand adds a mango flavor to its sparkling water line. In its first two months, mango sells 200 units a week across the stores that carry it. The team celebrates. Then someone checks the rest of the line and finds the original flavors are down by 100 units a week in those same stores. That's product cannibalization: sales a new item takes from your own existing products.

Some cannibalization is normal and expected. The question is how much, and whether the new item earns its place anyway.

How do you measure cannibalization?

Compare the whole line before and after, in stores that added the new item. In the hypothetical above, the line sold 900 units a week before mango arrived. After, it sells 1,000 including mango's 200. The line grew by 100 units. Mango sold 200. So half of mango's sales were new to the brand and half came from the existing flavors.

That ratio, cannibalized sales divided by the new item's sales, is the cannibalization rate. Here it's one in two.

The comparison is stronger with control stores. If stores that didn't get mango also saw the original flavors decline, some of the drop wasn't mango's doing. Subtracting the control group's change gives a cleaner read.

Timing matters too. New items often get a trial bump in their first weeks. Measure over a longer window, a few months if you can, so novelty doesn't distort the result.

When is cannibalization acceptable?

When the new item brings something the old ones don't. Three situations come up often.

The new item carries a better margin. If mango earns more per unit than the flavors it replaces, the brand can come out ahead in profit even with heavy cannibalization.

The new item defends shelf space. If a rival was about to fill that gap with its own mango, taking some of your own sales is better than losing them to a competitor.

The line total grows. Any net gain in brand sales means the extension added something, even if it also shifted some volume around internally.

Cannibalization becomes a problem when the line total is flat, the margin is no better and the new item takes a facing from a stronger seller.

Where should you look in the data?

At store level first. Cannibalization often varies by store. In a store with plenty of shelf space, a new flavor may sit alongside the existing ones. In a small store where it replaced a facing of the best seller, the original item may lose sales simply because there's less of it on the shelf.

Then at size and pack level. A new multipack might pull from singles of the same product, which changes both units and dollars in ways worth separating.

It also helps to check price. If the new item launched at a lower unit price than the existing flavors, some shoppers may have switched for price rather than flavor. That's a different kind of cannibalization, and it affects margin more than volume.

What would you do with this on Monday?

Before the next launch, write down what you expect: how many units the new item should sell, and how many of those you're willing to see come from the existing line. After launch, check against that expectation at four, eight and twelve weeks.

For items already on shelf, run the line-level comparison now. Background on how line extensions are typically planned is a useful refresher. For category context, NRS Insights reports show how the broader category moved over the same period.