shutterstock_600560726 (1)


Determining the right shelf assortment is important to retail success.


One of our leading European FMCG retail clients in its region, with a clientele of more than 1.2 million customers and more than 200 stores, was interested in improving their shelves assortment.

Within each store, they had multiple planograms and different strategies. Their business goal was to optimize their product categories, reduce the number of products on the shelves, reduce the number of planograms and identify the ideal and most profitable planogram.


The current way of deciding which products will be on the shelves was based on shelves’ space and favorite products derived from customer surveys. Our Data Science team proposed an integrated method. To take into consideration the consumer behavior data for each store and incorporate product substitution probabilities.

We also had to ensure that the optimal planogram was placing the most productive SKUs in the right locations, in the most relevant stores to maximize profit. Our client also wanted to decrease the number of products at the chain level, decrease the number of planograms and also provide price variation on the shelves.


The algorithm was initially built to oversee local preferences and trends for each store and match each store’s customer preferences. You can learn many things about your customers from their purchase behaviours and product consumption. Especially if you focus on your top customers, based on highest Recency, Frequency, Monetary value (RFM). And this is what we did for our client. We tracked consumers behavior over a pre-defined period and identified all switching patterns.

Additionally, the algorithm selects products based on:

  • Estimated sales of a combination of products
  • Substitution probabilities (The proportion of the products’ sales, that can be allocated to all other products in the assortment, in case the former is delisted)
  • Share of each product sales
  • Ratio of each price category of each planogram
  • At least one product from each price category


RetailZoom delivered a tool with more than 40 statistics, helping the client to act faster with much more condifence and achieve the highest and best use of shelves. The tool provides suggested assortment proposals at store level, to maximize profit.

In parallel it provides the user with all the information required to simulate different combinations and see the impact on each sales, quantity and products.

Assortment Optimization allows you to:

  • Identify the optimal number of SKUs within the category
  • Define the optimal SKU mix
  • Better awareness of product positioning
  • Optimize product variety
  • Add, delete and keep products optimally to maximize profitability
  • Unified planograms

In parallel it provides the user with all the information required to simulate different combinations and see the impact on each sales, quantity and products.

An evaluation of the new assortment was performed and a test-cotnrol method was applied:

  • Overall decreased the number of planograms by 77%
  • Decreased the total number of unique products by 12.5%
  • 8% sales increase in tested categories

Key Findings:

65 % ⇩
Overall decrease in the number of planograms
10 % ⇩
Decrease in the total number of unique products
1 % ⇧
Sales increase in tested categories