What is demand planning?

Demand planning is the process of deciding how much of each product you expect to sell, and agreeing that number across sales, operations and purchasing. A forecast is one input. The demand plan is that forecast after your team has corrected it for what the numbers cannot know, such as a new customer or a planned promotion.

Demand planning vs demand forecasting

	Demand forecasting	Demand planning
What it is	A calculation from past data	A process run by people
Who does it	A model or an analyst	Sales, operations, purchasing and finance together
Output	Expected sales by item and period	One agreed number that drives buying and production
Main risk	Bad history or the wrong method	Everyone keeps their own number

The demand planning process

1. Clean the history. Correct months affected by stock-outs and one-off spikes.
2. Forecast each item with the method that fits its pattern.
3. Add what the data cannot know: promotions, new items, phase-outs and customer news.
4. Agree one number across sales, operations and purchasing.
5. Feed the supply plan, so purchasing and production work from that number.
6. Measure the error and learn from it. See the forecast accuracy formula.

What demand planning software should do

• Choose a forecast method for each item by testing methods on months it held back.
• Correct history for stock-outs and unusual orders, and show what it changed.
• Handle seasonality, new items and phase-outs.
• Show the error for each item, not just one average.
• Let people override a forecast, and record whether the override helped.
• Pass the result to reorder points, purchase orders and the production plan.

Does a small business need demand planning software?

Not at first. With a few dozen steady items, a spreadsheet and a monthly conversation are enough. Software starts to pay when you have too many items to forecast by hand, when seasonality or promotions make simple averages wrong, or when sales and operations keep arguing about whose number is right.

How to judge the forecasts

1. Ask for a backtest on your own history. Hold back the last few months and compare the forecast with what really sold.
2. Compare it with something simple, such as last year’s sales or a moving average. A tool that cannot beat that on your data is not worth paying for.
3. Look at both error and bias. A forecast can be accurate on average and still run high every month. Our forecast accuracy guide shows how to calculate each.

AI and demand planning

AI methods can help when you have long history, many items and useful outside signals. They still depend on clean data and need the same backtest as any other method. Read about AI demand forecasting software and whether your data is ready.

Connecting demand to the rest of the business

A demand plan only pays off when it changes what you buy and build. That happens in a monthly planning meeting. See S&OP for small manufacturers and how S&OP, SIOP and S&OE differ. If you would rather have help setting it up, see our demand planning consulting.

Where the Statadventure Optimization Lab fits

The Optimization Lab forecasts each item from your monthly sales history, including promotions, stock-outs, new items and phase-outs. It chooses the method on held-out months, shows the error, and explains why each method was picked. The forecast then feeds reorder points and suggested orders.

Demand forecasting is on the Planner and Growth plans. See the full comparison.

Frequently asked questions

What is the difference between demand planning and supply planning?

Demand planning decides how much you expect to sell. Supply planning decides how to meet that number with purchasing, production and inventory.

How much history does demand planning software need?

Twelve months is a workable minimum. Two years or more lets a tool see seasonality and test itself on months it did not use.

Can I do demand planning in Excel?

Yes, for a modest number of items. It gets hard when one person owns the file, when seasonality needs a different method by item, or when nobody has time to update it.

How accurate should a demand forecast be?

It depends on the item. Steady, high-volume items can be forecast closely, and erratic, low-volume ones cannot. Track your own error by item class and measure improvement against a simple method.