Learn how to use time series analysis and the four‑period centred moving average to predict future sales. Think of it as smoothing a bumpy road so you can see the overall direction.
A time series is a list of data points collected at regular time intervals. For sales, it could be monthly, quarterly, or yearly sales figures.
This method averages the sales of four consecutive periods, but the average is assigned to the middle of those four periods. It helps to reduce short‑term fluctuations.
Mathematically:
\$MA{t} = \frac{S{t-1} + S{t} + S{t+1} + S_{t+2}}{4}\$
Here, S is the sales figure for a period, and t is the centre period of the four‑period window.
Imagine driving on a road with potholes (sales spikes). A moving average is like a car with a suspension system that smooths out the bumps, so you can see the general slope of the road (overall trend).
Example Data
| Period | Sales (units) | 4‑Period MA |
|---|---|---|
| 1 | 120 | — |
| 2 | 135 | — |
| 3 | 150 | 127.5 |
| 4 | 160 | 142.5 |
| 5 | 170 | 157.5 |
| 6 | 165 | 172.5 |
| 7 | 180 | 177.5 |
| 8 | 190 | — |
The most recent 4‑period MA (for period 7) is 177.5. We use this as the forecast for period 8.
If the actual sales for period 8 were 190, the forecast was slightly low. You could refine the model by adding more periods or using a weighted moving average.