Many companies often start forecast projects driven by technology. What is missing for really good forecasts? The technical relevance.
Forecasts should not be based on "tech enthusiasm", but on a specific need from the specialist area.
The following questions arise in this context, for example:
- Which KPIs cause uncertainties in planning?
- Where are there risks in terms of costs, personnel or sales?
- What would have a clear control benefit if it could be anticipated?
Professional relevance before technology: what you should really be forecasting
Not every KPI is equally suitable, neither from a data nor a benefit perspective. The selection of the "right" KPIs requires:
- Good data quality and sufficient history. Without these basics, no meaningful models can be trained.
- Expert assessment of whether a forecast is relevant to management at all.
Tip for decision-makers: Ask yourself or your team: "Would predicting this metric improve our decisions?"
More than just playing with numbers: what forecasts really do for you
Well-made forecasts have clear advantages across all areas:
Early recognition of trends and deviations
Developments such as liquidity bottlenecks, declining supply or staff shortages can be identified at an early stage - before they become a problem. This creates real room for maneuver.
Better planning and more realistic targets
Reliable forecasts provide realistic values for budgets, resource planning and sales. This improves target agreements and planning quality in the specialist departments.
Risk minimization through a sound basis for decision-making
Forecasts provide expectations and probabilities based on past trends. This reduces uncertainties and supports well-founded decisions - both strategically and operationally.
Greater involvement of the department
A good forecast brings data teams and specialist departments closer together: Models are discussed and results are evaluated together. This anchors data thinking in everyday working life.
Proactive management instead of retrospective control
If you know early on where things are heading, you can act better. Forecasts enable real control - not just the observation of past deviations from targets.
From KPI to action: integrating forecasts into processes
A forecast is not an end in itself. It only develops its added value, when he is involved in decision-making processes. Important:
- Forecasts must be linked to responsibilities.
- Data owners and specialist departments must work together from the outset.
This is how it works: Your forecast in 6 steps
- Define goal and clarify benefits
- Coordinate data situation and use case
- Explorative analysis (trend, seasonality, etc.)
- Model selection (e.g. Holt-Winters, ARIMA)
- Create and evaluate forecast
- Regular review and adjustment
Conclusion: Forecasts start in the specialist area - not in the tool
Technology is important. But Forecasts only become valuable when they solve real business problems. To do this, you need relevant KPIs, good data and cooperation between the specialist and data worlds. Then forecasts will become your management tool.
Sources:
- J. B. Kühnapfel, Sales Forecasts, 2nd ed. Wiesbaden: Springer Fachmedien Wiesbaden, 2023, isbn: 978-3-658-42875-4
- R. Hyndman and G. Athanasopoulos. "Forecasting: principles and practice, 3rd edition." (2021), address: https://www.otexts.com/fpp3 (visited on 16.06.2025)
- M. Hebing and M. Manhembué, Data Science Management : From Initial Concept to Governance of Data-Driven Organizations. Heidelberg, GERMANY: o'Reilly, 2024, isbn: 9783960108085
- M. Lang ed., Data Literacy. Munich: Carl Hanser Verlag GmbH & Co. KG, 2023, isbn: 978-3-446-47585-4.
