Filson’s leadership wanted to have a high degree of confidence in their e-commerce growth investments through 2026. They wanted a way to understand various scenarios of customer acquisition efficiency, customer lifetime values, and advertising investment levels.

The model has adjustable inputs for customer lifetime value, customer acquisition cost, and advertising investments. The model outputs forecasts in various forms based on those inputs.

For the model’s first big test, Filson used the model to predict their e-commerce sales for the 2023 Holiday Season.

The teams compared forecasts to actuals a few months into the Holiday Season and were delighted to learn that they were very precise, with one month forecasted at 99.6% of the actual. Other months were forecasted within a few percentage points. The teams also learned even more ways to improve the models accuracy in future time periods.