The demand planning process: 6 steps for small and mid-size companies

Demand planning turns what you know about customers into a forecast the rest of the business can plan with. In a small or mid-size company it does not need a department or an expensive system. It needs a short monthly routine that everyone keeps. Here are the six steps.

The six steps

1. Gather and clean the history. Use shipments or invoiced sales, not orders, and use 24 months if you have them. Mark the months a stockout cut sales, and take out one-off orders that will not repeat.
2. Build a statistical baseline. Forecast each item or family from its own history, and test the method on months it did not see. Seasonal items need a seasonal method, and slow or lumpy items need a method made for them.
3. Add what the numbers cannot know. Sales and customer service add new customers, lost accounts, promotions, price changes and new products. Ask for reasons, not only numbers.
4. Agree one number. In a short monthly meeting, sales, operations and finance agree the forecast that everyone will use. Record the decisions and who made them.
5. Hand it to supply planning. The agreed forecast drives purchasing, production and inventory: MRP for what you build, and reorder points for what you buy and resell.
6. Measure accuracy and bias. Compare last month’s forecast with what happened, by family and for your most important items, and look for a forecast that is always too high or too low.

Measure step 6 with the forecast accuracy formula: WAPE for the size of the errors, and bias for their direction.

A monthly calendar

Week	Step	Who
1	Close the month, clean the history, run the baseline	Planner
2	Collect market input and review the exceptions	Sales and planner
3	Agree the forecast	Sales, operations, finance
4	Release it to supply planning, and measure last month	Planner

The agreement in week 3 is the demand step of S&OP. If you also plan inventory and capacity in the same cycle, see S&OP vs SIOP.

How much detail?

Agree the numbers by product family, where forecasts are more accurate and the meeting stays short. Let the system break families down to items. Review items one by one only when they matter: your top sellers, new products and items with long lead times. ABC/XYZ classification tells you which ones those are.

Common mistakes

• Forecasting from orders, which double count when customers re-order, or from sales cut short by stockouts.
• Letting sales targets become the forecast. A target is what you hope for. The forecast is what you expect.
• Changing the numbers outside the meeting, so nobody knows which forecast is current.
• Never measuring accuracy, so the same errors repeat every month.
• Too much detail. A meeting that reviews 2,000 items reviews none of them well.

Spreadsheets or software?

A spreadsheet works for a few dozen families. It gets hard when you need a forecast for every item and location, a test of which method predicts best, and an accuracy report every month. That is what demand planning software is for, and AI demand forecasting software adds methods that learn from more signals. If your history needs work first, start with a data readiness check, and if you want someone to set the process up with you, see what a demand planning consultant does.

Frequently asked questions

What is the demand planning process?

A monthly cycle that turns sales history and market knowledge into one agreed forecast: clean the history, build a statistical baseline, add sales input, agree the numbers, hand them to supply planning, and measure how accurate they were.

How often should we run it?

Monthly suits most small and mid-size companies. Review fast-moving items weekly if demand changes quickly, but keep the full cycle monthly.

Who owns demand planning in a small company?

Often the operations or supply chain manager, with sales giving the market input. What matters is that one person owns the number and runs the meeting.