A chain of high-street bakery outlets presenting multiple cities in Pakistan. The brand has opened 20 branches across Pakistan and most recently, has gotten into the airline catering business.

The Problem

A national bakery chain in Pakistan has a manual process for estimating SKU demand. The process of ordering stock with respect to SKU for the next day is done by the branch manager of each site and is completely intuition based. This leads to loss of sales and an additional wastage because the demand is not based on any data points. Moreover, the demand sheet was sent on paper to the factory to communicate numbers for the next day.

Data Pilot has built predictive models for the top 100 SKUs for top 5 branches to predict demand of each SKU. Moreover, we have also built a web-based application which has dashboards and sheets which show demand for each SKU coming from the model output. The web-based application is accessed by the manager of each branch and the production personnel at each factory. The demand is sent automatically to the factory through the portal after the branch manager approves it.

This would lead to increased revenue, less wastage and reduced stock-outs. The error rate of model prediction for demand for each SKU is 6. For e.g. If model predicts 25 cupcakes will be sold tomorrow, the real sale will fall within the range of 19-31.

Case Studies

This is some text inside of a div block.