top of page
Beyond Putting Data to Work Strapline


revenue potential


This clothing retailer was looking to optimise sales through better predicting customer demand. However, getting the balance of stock right was proving challenging.  At times ordering too much of a SKU and it not selling, at other times ordering too little and stock was quickly selling out, leading to lost sales (loss demand) and customer disappointment.

What We Did

We developed a view of the customer journey to understand how customers encounter an out of stock product and how we could estimate when out of stock products would have sold. We built and tested a Loss Demand model that estimated demand and provided a proportional allocation method with alternative sense checks.


This model answered a number of important questions such as: - Which products could have sold more and by how much? - What does the optimal stock position look like? - What are the common attributes of products with high loss demand? - To what extent does substitution exist in purchase behaviour? Using the model identified the potential to increase net revenue by £149m through reducing this loss demand.

bottom of page