T1A helps a shopping channel network to improve the demand forecast and call centers’ capacity planning

T1A team boosted the quality of phone calls forecast at a shopping channel. Within a month T1A team increased the accuracy of predictions to client’s call centers to optimize schedules and get an improved level of customer service. In addition, T1A held several workshops on data science basics and analyst enablement to improve data literacy and set up guidelines for all future models.

The COVID-19 pandemic has made drastic changes in everyday life. Customer behavior has become more hectic and unpredictable. Therefore, foreseeing the amount and the issues of calls has become more challenging. As a reaction to these changes, some call centers had to change the working hours, which led to increased cost of operations.

T1A team started out with an audit of existing models - a crucial step that enabled quick onboarding and a solid grasp of potential improvements. In the next phase, T1A significantly improved the analytics data mart, used for modeling, by expanding the number of features from dozens to hundreds within the same data sources and provided modeling techniques and tested dozens of models to find the best one. T1A used a holistic approach to the project and improved all pillars of proper forecasting – analysis, data and model. The team predicted the volume of calls in call centers for the most important geo regions and instructed the employees from others on how to reproduce the results.

As a result, in a month the predictions have become 2% more accurate, saving 7% of the working hours of call centers’ employees as well as clients’ waiting time. 

In conclusion, we can say that the demand forecast has been an overwhelming issue for all the retailers lately. Many organizations have even abandoned attempts to make algorithmic forecasts for certain product groups.  Projects like this one, give us very good faith that these dark days are over, and with proper methodologies, we can go back to precise planning even in turbulent environments.
T1A T1A Europe