Turning Cannibalisation in Retail from a Risk into a Competitive Advantage

Turning cannibalisation in retail from a risk into a competitive advantage

Writer

ADC

Date

February 23, 2026

Contact us

What is product substitution / cannibalisation?

With the proliferation of online shops and ever-increasing competition, retailers are continuously expanding their product ranges to offer more choice to customers. As a result, retailers face a critical challenge: providing sufficient variety to satisfy customers while balancing internal product competition. This is especially true in the fashion industry, where trends shift rapidly and product lifecycles are short.

Substitution (often referred to as cannibalisation) occurs when two or more similar products compete for the same customer, causing sales of one product to come at the expense of another. A well-known example is how the launch of Coca-Cola Zero impacted the sales of Coca-Cola Light. Although distinct products, their similarity meant they targeted overlapping customer segments. Substitution is thus not something inherently bad, it is part of any retail business, but it becomes problematic when left unmanaged or ignored.

How does substitution impact your business

Substitution effects significantly complicate pricing and promotional decisions. Discounts intended to boost sales for one SKU may simply shift demand away from another, eroding margins without increasing total revenue.

At scale, pricing complexity becomes overwhelming quickly. Let’s demonstrate this with a simple example.

Imagine we want to set prices for just two products: SKU A and SKU B. For each product, we need to decide between four possible prices. Because changing the price of SKU A affects how much of SKU B we sell (and vice versa), we can’t decide on each price in isolation. Instead, we have to consider every possible combination of prices.

With two products and four price options each, that already means 16 different price combinations we need to consider. Adding just one more product increase this to 64 combinations. Each additional product multiplies the number of possibilities again.

This growth is exponential. In a store with 200 products, each with four possible price points, the number of possible price combinations becomes bigger than the number of atoms in the universe, and far beyond what any human or computer could evaluate.

"Substitution is thus not something inherently bad, it is part of any retail business, but it becomes problematic when left unmanaged or ignored."

A scalable approach to modelling substitution

The naïve approach to modelling substitution effect, evaluating all possible combinations, is thus impossible. To solve this problem, we recommend to implement a structured simplification without losing impact by applying 3 steps to greatly reduce the number of price combinations our algorithm needs to consider.

The first step is to reduce the initial discounts we consider. We use business and legal criteria to exclude certain discount options. For example, discounts that would lead to negative margins are ruled out.

The second step is to classify SKUs as high or low impact. Low-impact SKUs contribute little to overall revenue or inventory risk. While they may still experience substitution, their limited financial impact allows us to safely ignore these effects in favour of computational efficiency.

Thirdly, we limit the number of SKUs that substitute our high impact SKUs. High impact SKUs are important, and we want to make sure we get their discount right. But instead of calculating the substitution effect of all SKUs on our high impact SKUs, we only consider SKUs that are highly similar. A pair of black shoes will cannibalise the sales of another pair of black shoes, but the same does not hold for a pair of black shoes and a pink shirt.

Thus, by tuning the number of discounts, the threshold for high impact SKUs, and the threshold for high substitution potential SKUs, we can adjust the number of price combinations our algorithm needs to evaluate. These 3 steps give us the ability to effectively incorporate substitution effects while still having the problem be computation feasible. For a full deep dive into our method from a theoretical as well as practical angle, please read our report.

Proven impact on profitability

The economic upside of properly accounting for substitution is substantial. Research shows that optimising assortments while considering substitution can increase retailer profits by 37.3%, rising to 43.7% when pricing is optimised simultaneously (Rooderkerk et al., 2013).

By intelligently reducing the number of price combination, these gains become achievable across real-world product ranges and marketing campaigns.

Interested in improving the performance of your next pricing campaign while avoiding hidden cannibalisation effects? Contact us to learn how our approach can help.

Explore the full method behind our substitution modelling approach

See exactly how the method works in practice and how you can apply it to improve pricing outcomes.

Read the report

Talk to our experts

Let's create real impact together with data and AI

Petr Pushkar

Senior Manager, Retail

Petr Pushkar