National Tire Distributor Optimize Demand

The client was formed as a joint venture between two global tire manufacturers and, following this merger between the two companies, it was evident that a more accurate demand forecast needed to be implemented. 

Previous forecasts proved to be very rudimentary, and the data history was obviously reflective of two completely separate companies, rather than just one. These inadequate demand forecasts were having a disastrous knock-on effect on the business as large numbers of sales were being lost due to a combination of stock shortages and increased storage costs due to overstocking. A difficult balancing act.

An optimization of demand planning was critical for the future success of the business and the client engaged us to build a Machine Learning model to facilitate this.

We developed multi-layer models with a live connection to an SQL server to predict how many of exactly which products would sell where. This also resulted in a recommendation of how much inventory to store in each warehouse across the country.

Results were integrated into the business with cloud-based interactive tools to support inventory planning and resource management. This tool allowed users to adjust predicted inventory levels based on their knowledge of promotions and sales. Model results were evolved and empowered by intuitive dashboards. 

The forecast methodology deployed succeeded in achieving an average of 95% accuracy.  

As a direct result of this, lost sales from stock shortages were kept to an absolute minimum and storage costs decreased. 

What’s more, this efficient forecasting further increased sales as the data enabled client to be more progressive and strategic in terms of effective marketing and supplier negotiations.