Demand Forecasting Basics

Use past sales to estimate future demand so you buy ahead of need, not after a stockout.

4 lessons 35 min 5-question assessment 75% to pass

What you’ll learn

Explain why forecasting beats reacting

Forecast from sales history and averages

Account for seasonality and trends

Check and improve forecast accuracy

Lesson 1 of 4 Reading 8 min

Why forecast

Forecasting is estimating future demand so you can buy and stock ahead of need. The alternative — reacting only after stock runs low — always lags real demand by at least the supplier lead time.

It matters because lead time means today’s order serves next week’s demand; if you only react when shelves are bare, you are already too late. A forecast lets you order before the gap appears.

If your supplier takes 7 days and demand jumps next week for a holiday, reacting on the day of the stockout means a week of empty shelves. A forecast that sees the holiday spike coming lets you place the larger order this week, so stock lands before the rush instead of a week after it.

Key takeaways

Forecasting estimates future demand to buy ahead.
Reacting after a stockout lags by the lead time.
Today’s order serves next week’s demand.
Example: a 7-day lead time means reacting late loses a week of sales.

Lesson 2 of 4 Practice 9 min

Forecasting from history

The simplest forecast uses sales history: average recent demand to project forward. A moving average (say the last 8 weeks) smooths out random spikes and dips into a usable baseline.

History-based forecasting matters because past sales are the best cheap evidence of future demand for steady items — far better than a gut feeling, and it improves as you gather more data.

If a SKU sold 90, 110, 100, 120, 95, 105, 100, 110 over eight weeks, the 8-week average is about 104/week — a sensible baseline to plan next week’s stock. One odd week of 200 (a one-off bulk order) would be visible as an outlier to exclude, rather than tricking you into permanently over-ordering.

Key takeaways

Average recent sales to project demand forward.
A moving average smooths random spikes and dips.
Past sales beat gut feeling for steady items.
Example: eight weeks averaging ~104/week sets the baseline.

Lesson 3 of 4 Reading 9 min

Seasonality and trends

A flat average misses two real patterns: seasonality (regular peaks like December or back-to-school) and trend (steady growth or decline). A good forecast adjusts the baseline for both.

Accounting for these matters because a plain average under-buys before a known peak and over-buys after it — and ignores a SKU that is quietly growing 5% a month until it stocks out.

If a SKU sells 100/week most of the year but 300/week each December, planning December on the annual average of ~120 stocks you out badly. Likewise, a line trending up 5% monthly will out-run a flat forecast within a quarter. Layering last year’s December factor and the upward trend onto the baseline keeps the plan honest.

Key takeaways

Adjust the baseline for seasonality and trend.
Seasonality is regular peaks; trend is steady growth or decline.
A flat average mis-buys around peaks and ignores growth.
Example: a SKU jumping 100→300/week in December breaks a flat forecast.

Lesson 4 of 4 Reading 9 min

Checking accuracy

A forecast is a guess you must grade. Compare what you forecast to what actually sold, track the error, and feed the lesson back so next month’s forecast is better. Forecasting is a loop, not a one-off.

Checking matters because an unchecked forecast can drift wrong for months unnoticed — consistently over-forecasting builds overstock, consistently under-forecasting builds stockouts, and only the error tells you which.

If you forecast 104/week but actual sales average 130 for two months, your forecast is running 20% low and you are quietly stocking out; the tracked error is the signal to raise the baseline. Reviewing forecast-versus-actual monthly turns each miss into a correction rather than a repeated mistake.

Key takeaways

Grade forecasts against actual sales and track the error.
Feed the lesson back so next forecast improves.
Unchecked forecasts drift into overstock or stockouts.
Example: forecasting 104 against actual 130 signals a 20%-low baseline.

| AWRA OpsHub

Demand Forecasting Basics

What you’ll learn

Course content

Why forecast

Forecasting from history

Seasonality and trends

Checking accuracy

Finished the material?

More in this path

Inventory Foundations

Warehouse & Locations

Units of Measure & Conversions

Need a quick answer while you read?

AwraIQ