Unsupervised clustering techniques to rationalise products across the globe
Global retail company with over 80,000 different products (spectacles). Key question is how to rationalize this product landscape to provide clusters of spectacles and profiles of retail chains with similar product ranges.
We collected sales data and product attributes. Using this data we designed a dynamic dashboard to slice & dice the product sales along various dimensions and applied machine learning techniques (unsupervised learning) to provide clusters of spectacles and profiles of retail chains with similar product ranges.
We provided insight in the products sold, such as in which quantities in what country, colours, material etc. For example, some spectacles are popular for particular target groups that are mainly covered by specific retail chain(s).
Finally, we provided a tailored open-to buy list per profile of retail chains. This enables the customer to rationalise its product range and write guidelines for their different retail chains across the globe.