Demand forecasting and inventory planning
Short answer
SwankyForge builds demand forecasting that predicts sales per product, store or warehouse and turns the forecast into order and stock recommendations. Forecasts add up consistently from product to category to company total and plug into the planning process your team already uses. A typical project takes 1–3 months and costs €10k–€28k.
Sounds familiar?
- Bestsellers run out while slow lines fill the warehouse.
- Planning runs on one growth rate in a spreadsheet.
- Promotions and seasons are adjusted by hand.
- Cash is tied up in the wrong stock.
- Nobody can say what a stockout actually costs.
What we build
Forecasts per product and location
Weekly or daily, with a range rather than a single number.
Promotions and seasons
Effects of price, campaigns and holidays learned from your history.
Order recommendations
Quantities from the forecast, lead times, minimum orders and your service level.
What-if scenarios
Planners test a promotion or a supplier delay before committing.
Accuracy tracking
Forecast error measured every week against what actually sold.
Exception lists
The few products that need a decision, instead of a full spreadsheet.
How a project runs
- 1
Measure today's error
We score the current planning method on past data, so every improvement has a baseline.
- 2
Clean the history
Stockouts, promotions and one-off events are marked, so the model learns demand rather than what happened to be on the shelf.
- 3
Forecast and reconcile
Product forecasts are made consistent with category and company plans.
- 4
From forecast to orders
Forecasts become order quantities using lead times, minimum orders and the service level you choose.
- 5
Planners review exceptions
The system flags products where it is unsure or where demand has shifted, and planners decide.
What you get
- Forecasts and order recommendations in your planning tool or ERP
- Backtest report against your current method
- Weekly accuracy dashboard
- Retraining pipeline and runbook
- Code and models in your repository
Typical price
€10k–€28k
Typical timeline
1–3 months
From the published rate card. The final price is fixed once the scope is agreed.
Estimate your projectQuestions
How much sales history does demand forecasting need?
Ideally two years, so each season repeats at least once. With less, the model leans on similar products and categories.
Will it replace our planners?
No. It removes the spreadsheet work and leaves planners with the decisions a model cannot make, such as a supplier problem or a new range.
Can it work with our ERP?
Yes. It reads sales and stock from your existing systems and writes recommendations back where planners already work.
How do we know the forecast is better?
It is scored against your current method on the same past periods before go-live, and accuracy is tracked every week after.
Do we need machine learning for this at all?
Not always. For a small, stable range a well-built statistical baseline may be enough. We check that first.
Tell us about the task.
The first call is free. If this system is not the right answer for your numbers, we say so.