• Household appliances
  • via Proxify

Electrolux

Forecasting next month's warranty service calls for every product group.

The challenge

During the warranty period, typically one year, Electrolux sends technicians to customers' homes. The task was to forecast the volume of these service calls from historical data.

What I did

I built a regression model based on Markov Random Fields, trained on one year of data, that predicts next month's service calls. The target is the Service Call Ratio (SCR): service calls normalised by the number of items sold.

An uncertainty band reflects the uncertainty of the sales-volume estimate. I showed the results across all product hierarchies (product line, group and subgroup) in an interactive sunburst chart.

Results

For the food preservation category (mean sales volume 176,735 items per month), the Mean Absolute Percentage Error on the SCR is 4.2%. Across all product hierarchies the MAPE stays below 14%.

Contact

Bring me your analytics question


Send a short email. I reply within two working days with whether I can help, a rough plan and a fixed quote.