Most corporate planning works with point estimates: “We expect revenue of €2.4 million in Q3.” That number sounds precise — but it isn’t. It hides how certain the estimate is, under which assumptions it holds, and what happens if those assumptions don’t hold.
I build forecasting models that make uncertainty explicit. The result is not a single number but a robust corridor — useful for operational steering, investor reports and strategic planning alike.
What are forecasting models — and when do you need them?
A forecasting model is a statistical method that estimates future values based on historical data and known influencing factors — with quantified uncertainty.
Forecasting models make sense when:
liquidity planning is needed on a monthly or weekly basis.
revenue or cost development has to be presented convincingly to investors or lenders.
seasonal fluctuations, growth trends or external factors complicate planning.
you want to weigh different scenarios (best case / base case / worst case) on a sound basis.
My approach
1. Check the data foundation
Forecasts are only as good as the underlying data. I begin with an analysis of your historical time series: completeness, consistency, structural breaks. Data gaps or errors are identified and documented — before the model is built.
2. Model selection
Depending on the data and the forecasting goal, different methods are used: classical time series models (ARIMA, ETS), regression models with external predictors, or ensemble approaches. The model choice is justified, not just implemented.
3. Uncertainty quantification
Every forecast includes confidence intervals or prediction intervals — depending on the use case. You see not only the expected value, but also the plausible corridor and the main drivers of uncertainty.
4. Scenario analysis
On request, I develop several scenarios with explicitly defined assumptions — not an arbitrary ±10%, but model-based ranges that follow from the data.
5. Handover
The finished model is documented and handed over in a form your team can follow and update — as an R or Python script, as a Power BI integration, or as a structured Excel model with a clear separation of logic and data layers.
Methods and tools I use
I work primarily with R (packages: fable, forecast, prophet) and Python (scikit-learn, statsmodels) for forecasting models. For integration into existing reporting environments I use Power BI and SQL. I add time series clustering and multivariate analysis methods where needed.
Who I work with
My clients are typically CFOs, managing directors and finance leads in mid-sized companies and startups in the EU who want to move their planning from Excel estimates to statistically sound models. I work project-based or hourly — remote or on-site in Berlin.
Reference projects and further reading
Practical insight into my methodological approach:
E-Commerce Analytics Dashboard (Power BI + R) — includes revenue forecasting components with R models, embedded in Power BI.
Time-Series Clustering with R’s dtwclust Package — a methodological article on time series analysis.
Let’s discuss your project
Do you want to put your revenue, cash flow or cost planning on a more robust foundation?
Pricing & Packages
- Fixed scope, defined in advance
- 1–2 weeks duration
- 1 revision round
- Summary deliverable document
- Post-delivery support
- Iterative collaboration
- Complete project with iterations
- 3–6 weeks duration
- Unlimited revisions within scope
- Full documentation & handover
- 2 weeks post-delivery support (async)
- Ongoing monthly support
- Monthly, min. 3 months
- Continuous iterations
- Living documentation
- Priority access
- €90/h for additional work
- Monthly review call
Hourly rate for ad hoc requests: €90/h. All prices plus VAT.