Skip to content
Omnikalc

Demand forecast calculator: what will you sell next period?

Enter your last six periods of demand to get the next period forecast by exponential smoothing and moving average, with the past error of the smoothing.

Updated on
Your inputs

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Results

Next period forecast (exponential smoothing)

139.5

Smoothing has followed past demand closely (error under 10 %)
Next period forecast (3-period moving average)
146.3
Naive forecast (same as the last period)
147.0
Average past error of the smoothing (MAD)
13.0
Average past error in percent (MAPE)
9.1%

Ordering stock, staffing a shop or planning a production run all start with a guess about next period’s demand. Two of the simplest and most used methods are the moving average and exponential smoothing. This calculator applies both to your last six periods of demand and compares them with a naive forecast, then reports how accurate the smoothing would have been on your own past data.

How the calculation works

  • Moving average (3 periods) = average of the three latest periods
  • Naive forecast = demand of the latest period
  • Exponential smoothing: the first forecast equals the first demand; then forecast = α × latest demand + (1 − α) × previous forecast, repeated up to the next period
  • MAD = average of |demand − forecast| over periods 2 to 6
  • MAPE = average of |demand − forecast| ÷ demand over periods 2 to 6

Example

Weekly demand for six weeks is 120, 135, 128, 142, 150 and 147 units. With α = 0.3, the smoothed forecasts are 120, 124.5, 125.6, 130.5 and 136.3 for weeks 2 to 6, and 139.5 for week 7. The 3-period moving average gives 146.3, and the naive forecast is 147. The smoothing’s average past error is about 13 units, or 9 %, and it lags a rising demand: the moving average is closer here.

Why it matters

A forecast need not be perfect, but it should be better than a guess and come with a sense of its error. Smoothing methods are cheap, quick and transparent, which makes them a good baseline: if a more complex method cannot beat them on your data, it is not worth it. The past error also feeds the safety stock of the reorder point.

Good to know

At least a few periods of history are needed, and unusual periods (a stock-out, a promotion, a holiday) should be adjusted before using them. Since both methods lag behind a trend, they will under-forecast in a growing market and over-forecast in a declining one. Seasonal demand is better handled by comparing with the same period of the previous year. This tool is for education.

The reorder point calculator turns a forecast and its error into a safety stock, the exponential moving average calculator smooths a series of prices, and the EOQ calculator finds the best order size.

Frequently asked questions

What is exponential smoothing?

A forecasting method that gives the most weight to the latest observation and less and less to older ones. The forecast for the next period is α × the latest demand + (1 − α) × the previous forecast.

How do I choose α?

A low α (0.1 to 0.2) gives a smooth, stable forecast that reacts slowly, good for steady demand. A high α (0.5 or more) reacts quickly to changes but also to noise. Try several values and keep the one with the lowest past error.

What do MAD and MAPE mean?

MAD, the mean absolute deviation, is the average size of past forecast errors in units. MAPE is the same error as a percentage of the actual demand. Both tell you how much to trust the next forecast.

When do these methods fail?

Simple smoothing and moving averages cannot handle a trend or a seasonal pattern: they lag behind a steady rise and ignore the seasons. For such demand, use methods with trend and seasonal terms, or compare the same month from earlier years.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.