Demand Forecasting: The Methods
Demand forecasting means estimating future sales of a product or service in order to size production, inventory, and procurement in advance. It isn't an exercise in divination: it's a method that combines sales history, market knowledge, and the judgment of the teams. A good forecast is never perfectly accurate — it is simply reliable enough to make better decisions than intuition alone.
The main families of demand forecasting methods
Qualitative methods
These rely on expert judgment: sales teams, market opinion, panels. They are used when history is lacking — a new product, a new market — or to bring in information the numbers don't yet capture.
Quantitative methods (time series)
These extend the sales history: moving average, exponential smoothing, accounting for trend and seasonality. They suit established products with regular behavior, where the past reasonably illuminates the future.
Causal methods
These link demand to explanatory factors (price, promotions, weather, economic indicators). Richer, they also require more data and rigor to avoid confusing correlation with cause.
How to choose a demand forecasting method in an SME
The right choice depends on the data available and the stakes involved. A regular product with several years of history forecasts very well with a simple quantitative method; a launch will be handled by judgment, for lack of history. The useful rule in an SME: start simple, measure the forecast error, and add complexity only if it genuinely improves reliability. The forecast then feeds the decision — it's the first building block of the "decision" layer of the Two-Layer Model, the one that guides execution instead of being driven by it.
The common mistake
The most frequent mistake is chasing the perfect forecast instead of the useful forecast. Models get refined while the real question — what decision does this number drive? — goes unanswered. The other trap is never measuring the gap between forecast and actual: without tracking the error, there's no way to know whether the method is improving or deteriorating. One last pitfall lurks for SMEs: confusing the statistical forecast with the sales target. The former estimates what will probably happen; the latter expresses what the company wants to achieve. Mixing the two systematically leads to overestimating demand and over-stocking as a result. Keeping the two separate, then reconciling them consciously, is one of the hallmarks of a healthy forecast.
Take stock of your forecast
Want to know whether your demand forecasting is reliable enough to drive your inventory? Our forty-five-minute free diagnostic takes stock, with no commitment.
Frequently asked questions
What is demand forecasting?
It's the estimation of future sales of a product or service, in order to size production, inventory, and procurement in advance.
What are the main demand forecasting methods?
Qualitative methods (expert judgment), quantitative methods (time series such as moving average or exponential smoothing), and causal methods (explanatory factors).
Which demand forecasting method should an SME choose?
The one that matches the data available: judgment for a launch with no history, a simple quantitative method for a regular product.