A precise monthly forecast of future demand is a central factor in a reliable supply of medicines. Deviations between the forecast and actual demand have a direct impact on the ability to deliver and must therefore be kept as low as possible.Volume planning at HES was previously based on a wide range of different input sources. Customers provided monthly forecasts of varying quality, and not all manufacturers supplied forecasts consistently. Missing data was supplemented by sales estimates. Budget planning required customer-specific coordination between sales and controlling. This led to increased manual effort, limited forecast accuracy, and a planning logic that was difficult to scale overall.

At the same time, significant decisions rely on these forecasts: staff deployment and transportation planning at the depots, strategic network and depot planning over an 18-month period, and budgeting. HES was therefore looking for a partner for a superior, granular, and at the same time leaner forecast.

paretos built the planning together with HES in an agile setup, iteratively and close to day-to-day operations. Today, two forecasts are live in production. The daily forecast (Daily Parcel Prediction) delivers a rolling forecast per customer group, distribution center, shipping route, and depot over 12 weeks, with custom logic such as postal code mapping, regionalization, and a dynamic time offset between distribution center and depot.

The monthly forecast (Tactical Outbound) was switched from a weekly to a monthly cycle in early 2026 and extends 18 months into the future. Both forecasts interlock: the first three months of the monthly forecast are consistent with the daily forecast. The result is one forecast serving many decisions and stakeholders at once, from the distribution centers to the depots.

The daily forecast has significantly improved its accuracy over the years: forecast deviation fell from 19 percent (2024) to 12 percent (2025) and down to 9 percent (2026). The monthly forecast is already live for 10 of 14 customers and outperforms the previous customer forecasts in accuracy for 12 of 14 customers, even after just the second month of use.

Instead of two manually maintained Excel worlds, customer forecasts and sales forecasts, HES now receives a granular daily forecast, a level of detail HES could not achieve on its own. This reduces manual adjustments by around 22 hours per month and replaces estimated customer forecasts. The resulting monetarybusiness impact is substantial

Together, the plan is to roll out the monthly forecast to additional customers. Looking ahead, new paretos features such as the AI data analyst Socrates are set to make planning even more accessible. For both sides, the partnership is strategically central.