I was working at Rossmann in the Audit department, mainly dealing with stock correction data. One day, my manager got promoted to the Purchasing Department, and I started working on sales and stock data.
My first task after the transition was to collect a dataset that included product information, sales figures, stock quantities, and more. Using Python, I quickly completed the task. The report, presented in an Excel sheet, was very useful for my colleagues.
Reflexively, I felt the need to add a "stock coverage days" column to the report. To calculate this, I first determined the sales speed: dividing inventory quantities by sales speed gave me stock coverage days. It was a simple approach, and I applied the same calculation to every product without exception.
After adding the new column, I was curious to hear my manager's feedback. It turned out that the calculation was not accurate for new products. Some new products had only been on sale for 10 days, but I had calculated their sales speed over a 90-day period. This was just one of many exceptions. I then realized that I needed to classify products and develop unique calculations for each category.