What is cannibalization on a restaurant menu?

Cannibalization hides in plain sight because launch reporting celebrates the new item's sales without asking where they came from. Category growth is not proof of success: a new item's category can grow while the rest of the menu shrinks by the same amount, leaving the business exactly where it was, minus the launch cost.

The honest test is whether the total business grew beyond where it was already heading. That means comparing against the trend that existed before the launch and against the same period a year earlier, so momentum and seasonality don't get credited to the item. It also means looking at who bought it: genuinely new or returning-lapsed guests above the brand's normal rate, or existing guests visiting more often or spending more in total, are signals of new demand. Existing guests at their usual frequency, spending the same dollars differently, are substitution.

Some cannibalization is fine and even intentional — trading customers up to a higher-margin version of what they already buy. It becomes a problem when a lower-margin item displaces a higher-margin one, which is precisely the pattern value-priced LTOs fall into if nobody measures the net.

The same dynamic operates at the unit level. A new store that opens near an existing one can transfer sales rather than add them, which is why new-unit contribution should be judged net of the impact on neighboring stores, not on the new store's headline volume alone.