Forecasting Demand from Order History: A Simple Method for Distributors

Forecasting sounds like statistics, but for a distributor it is a practical question: how much of each product will we sell before the next shipment arrives? Order too little and you run out. Order too much and cash sits in the warehouse. The best source of evidence is already in your records: what customers ordered, when and in what quantity. With clean history and a few simple methods, you can improve purchasing decisions without advanced tools.

Start with clean history

Forecasts inherit the quality of the data. Before you calculate anything:

• Check that orders and returns are recorded consistently.
• Separate one-off events, such as a large institutional order, from regular demand.
• Note periods of stock-out, because sales in those periods understate demand.
• Keep product codes consistent so history does not split across duplicates.

Choose a simple method

Method	How it works	Best for
Last period	Assume next period equals the last	Very stable products
Moving average	Average of the last few periods	Products with steady demand
Seasonal comparison	Same period last year, adjusted for growth or decline	Seasonal products
Customer-level view	Add up expected orders from key customers	Products with a few big buyers

Begin with a moving average and layer adjustments on top.

A hypothetical illustration

Suppose a product sold 80, 90 and 100 units in the last three months. A three-month average gives 90 units per month. If the product has historically risen in the following month because of a festival, you might adjust upward based on what you saw last time. If a large one-off order was included in the 100, you would remove it before averaging. These figures are only for illustration.

Adjust for known events

• Festivals and seasons: use your own history.
• Schemes and promotions: expect a lift while active and a dip after.
• Price changes: demand can shift.
• New outlets or new dealers: add expected volume.
• Lost outlets: subtract.

Write down each adjustment and the reason, so you can learn from it.

Turn forecast into purchase quantity

1. Forecast demand over lead time plus review period.
2. Add buffer for variation.
3. Subtract stock on hand and stock on order.
4. Round to a practical order size.

Review the buffer regularly; see how to prevent stock-outs at retailers.

Check your forecast

Each month, compare forecast with actual. Note products where the gap is persistent and investigate. Maybe a customer behaves differently, or a competitor changed the market. Forecast errors are information, not failures.

Dealer-held stock

If dealers hold stock, your primary sales may not reflect true demand. Use secondary sales and stock counts to understand actual sell-out; see reconciling dealer stock with secondary sales.

How to run it in gamitrack

1. Keep orders complete. Book every order, including those taken in the field or at the counter, so history is complete.
2. Use customer and product history. Review order history per party and product to see pace and gaps.
3. Track secondary orders and dealer sell-out in via-dealers mode.
4. Record returns so net demand is visible.
5. Watch low-stock and batch views to avoid mistaking stock-outs for low demand.
6. Use visit notes and no-order reasons to understand why demand dipped.
7. Set targets from your forecast, monthly for officers and quarterly for dealers.
8. Share forecasts with the team through broadcasts about upcoming pushes.

The owner's morning dashboard routine is a good place to start a forecast review habit. Beverage and agro-input companies with strong seasons can start from beverages and agro inputs.

A hypothetical example

Imagine Example Beverages (an invented name) preparing for a warm season. The planner reviews the same months in previous years for each product, removes one large one-off order, and sets a base forecast from recent months. She raises the forecast for the products that rose in earlier warm seasons and keeps others flat. After the first month, she compares forecast with actual: two products ran higher, one lower. She adjusts the next month's plan and notes the lesson for next year.

Monthly forecasting routine

• Clean last month's history.
• Update the base forecast.
• Apply known adjustments.
• Convert to purchase quantities.
• Compare last month's forecast with actual.

Common mistakes

• Using sales during stock-outs as true demand.
• Not removing one-off orders.
• Over-trusting last year without checking this year's trend.
• Forgetting lead time.
• Never comparing forecast with actual.

Adding customer intelligence to the numbers

Numbers tell you what happened, but not always why. Combine order history with what your field team learns:

• A big customer plans to expand or cut back.
• A competitor launches a scheme in a territory.
• A dealer changes its sub-dealer arrangements.
• A school term, harvest or festival is coming.
• A road closure will delay deliveries.

Add a short column of notes next to the numbers during the monthly review, so the forecast reflects both. Keep the notes brief and dated, and check them against what happened.

Forecasting new products

New products have no history. Use similar existing products as a guide, start with small lots, and watch the first reorders closely. Resist large initial purchases based on hope. After two or three months of actual orders, switch to a moving-average approach and compare with your first guess.

Who should own the forecast

Assign the forecast to one person, ideally the one who controls purchasing, and ask sales leads to review it each month. A forecast with no owner becomes a spreadsheet that nobody trusts.

Summary in one line

A useful forecast is clean history, a simple average, honest adjustments and a monthly check against actuals. Start small with your top products, write down your assumptions, and improve the method only when the simple one stops being good enough for your decisions.

Next step

Pick ten products, calculate a moving-average forecast and compare it with next month's actual orders. Plans are on the pricing page.