Case Study: Freska predicts customer churn with Aito.ai

Freska uses Aito.ai to forecast customer churn with 2 monthly data sources

Freska, a Nordic home cleaning company, faced the challenge of accurately predicting customer churn. Their existing RFM model was insufficient, and while neural networks offered a potential solution, they were not scalable and lacked explainability. To improve prediction accuracy and understand the reasons behind churn, Freska began a proof of concept with Aito.ai's predictive database service.

Aito’s solution involved aggregating and preprocessing Freska's time-series customer data, including bookings and monthly metrics, into a format suitable for analysis. Using its Predict API endpoint, Aito.ai's database could then forecast the likelihood of a customer churning the next month. Crucially, the solution provided explainable results, highlighting the factors contributing to a customer's probable churn. The proof of concept was ongoing at the time of writing, with the next steps focused on maximizing prediction accuracy and incorporating additional data sources.

Freska