service · forecast-paket
Planning on data, not gut feeling.
Rolling forecasts for sales, liquidity and demand — self-correcting and traceable in the portal. Not a black-box model, but a number whose drivers you can see and push back on.
The challenge
The most important number in the business is often an estimate.
In many companies, planning is born in an Excel spreadsheet: last year's revenue, a markup, a gut feeling. That works as long as nothing changes — and goes wrong as soon as season, demand or incoming payments deviate from the pattern.
The Forecast Package replaces the estimate with a model built from your own data. It doesn't calculate once a year, but on a rolling basis, and corrects itself when reality diverges from the plan. What matters is not the number alone, but that you see how it comes about.
«A forecast that no one maintains is only a snapshot.»
three forecast types
Three questions, the same data discipline.
Sales forecast
Demand by product, channel and season — so that purchasing and production don't follow gut feeling.
Liquidity forecast
Cashflow from booking and payment history — bottlenecks visible before they become a problem.
Demand forecast
Materials, staff, capacity — planned on the data rather than on last year's average.
how we work
Discover → Build → Test → Rollout → Operate.
A forecasting model is only worth as much as its operation. That's why our work doesn't end at the finished model, but at the model that keeps correcting itself.
The honest comparison
First we set your current planning against the model — on your historical data. If the model doesn't beat the spreadsheet, we say so, and you keep your spreadsheet.
A model with drivers
We build the forecast from your data and make the drivers visible: you see what moves the number and can push back. No black-box result.
Against reality
The model is validated against held-out periods and quantified with a clear accuracy metric.
Visible in the portal
The forecast runs in the portal — with scenarios, drivers and history. Every run is in the audit log, traceable down to the data source. A control gate at every step that triggers something.
Operated self-correcting
The model is retrained on schedule and corrects itself against new data. A forecast that no one maintains is only a snapshot — this one stays current.
What you get
A number you can trust — because you can check it.
- Rolling forecasts
- Sales, liquidity and demand, updated on schedule rather than estimated once a year.
- Traceable in the portal
- Scenarios, drivers and history viewable — no black-box model you trust blindly.
- An accuracy number
- Forecast quality is quantified and continuously checked.
- Self-correcting models
- Retrained on schedule, corrected against new data — the number stays current without anyone maintaining it.
- A complete audit log
- Every run, every data source traceable. Whatever triggers something passes through an approval gate.
for whom
For everyone whose planning could be better than a gut feeling.
Seasonal and demand pressure
Those who buy and produce before demand is settled — and plan today with last year's average.
Liquidity under watch
Those who must see cashflow bottlenecks early, rather than discovering them at month-end.
Planning that lives in Excel
Those who run central planning today from a spreadsheet and gut feeling — and know both are hitting their limits.
intro call
Show us your planning — we'll show what a model makes of it.
In the intro call we clarify the planning question that costs you the most — and the honest comparison: does a model beat your current estimate on your data?