Statistical vs Sales Forecast Blend Calculator
Most demand planning teams hold two numbers, a statistical forecast and a view from sales. Averaging them 50/50 is easy but ignores which one has been right. This calculator weights each by its recent accuracy, or by a split you set, and shows the gap between them.
Your numbers
Used only in manual mode.
Your results
Inverse-error weighting assumes errors of similar bias. Review bias separately and keep a record of why a manual override was chosen.
Blended forecast as sales MAPE worsens
Sweeps Recent MAPE of sales forecast from 5 to 80, holding your other inputs fixed.
Track forecast sources automatically
Optional: we can pull statistical and sales forecast accuracy from your planning data and recommend weights.
Optional. The tool and every export work without this. No spam.
How to use it
Enter the two forecasts for the same item, period and level. Add the recent MAPE of each, measured over the last three to six cycles at the same lag. In automatic mode the weights are proportional to one over MAPE, so the more accurate source gets more weight. Choose manual mode when you want to apply a policy such as 70/30 for mature items.
- Compare at the same lag and level
- Use at least 3 cycles of accuracy history
- Document each manual override
- Check bias separately from error
The method
Combining forecasts usually beats the average single source because errors partly cancel. Inverse-error weighting sets the statistical weight to (1 over MAPE_stat) divided by the sum of (1 over MAPE_stat) and (1 over MAPE_sales). The blend is then stat times the weight plus sales times one minus the weight. This is a pragmatic stand-in for optimal weights, which also need the correlation between errors.
Example
Statistical 5,200, sales 6,100, MAPE 22% and 35%. Weights are 61.4% and 38.6%, so the blend is 5,547, and the gap is 900 units or 17% of the statistical figure. If sales accuracy improves to 20% the weights flip toward sales and the blend moves to about 5,690. The chart shows how much a deteriorating sales forecast loses weight.
Common mistakes
Blending at a level where one source has no information, for example a sales forecast that is really a target. Using MAPE from different horizons for the two sources. Letting a consistently biased source drive the blend rather than correcting its bias first. And not tracking forecast value add, which is the real test that the human input improves on the statistical baseline.
Frequently Asked Questions
Is a 50/50 blend a bad idea?
Not always, but it ignores track record. If one source is consistently twice as accurate, equal weights give away accuracy. Weighting by recent error is simple and usually better.
Why use MAPE and not bias?
MAPE measures size of error and is easy to compare. Bias shows direction and should be tracked too: a consistently optimistic source should be corrected before it is blended.
What is forecast value add?
The change in accuracy caused by a step in the process, such as a sales override on the statistical forecast. If the step makes accuracy worse, it adds no value.
Can I use more than two sources?
Yes. The inverse-error idea extends to any number: each weight is one over its error divided by the sum of those values.
Netray can measure statistical and sales forecast accuracy from your ERP and S&OP data and keep these weights up to date for each product family.
Related Tools
MAD to Safety Stock Forecast Error Calculator
Use your measured mean absolute deviation of forecast error to set safety stock and reorder point instead of a guessed buffer.
Supply Chain & InventoryLinear Regression Trend Forecast Calculator
Enter six periods of demand and get the least squares trend, how well it fits and the forecast for the period you choose.
Supply Chain & InventoryBullwhip Effect Calculator
See how much order variability a supplier sees compared with end customer demand, based on replenishment lead time and the number of periods used in the moving average forecast.
Supply Chain & InventoryPromotion Lift and Cannibalization Calculator
Convert baseline demand, promo lift, cannibalization and pull-forward into net incremental units and promotion profit, so the demand plan carries the real effect.
More Supply Chain & Inventory tools
See all- Oracle Fusion Inventory Turns and Carrying Cost Calculator
- Distribution ERP Fill Rate Calculator
- Distributor Rebate Tracking Worksheet
- Branch Transfer Planning Worksheet
- Distributor Slow-Moving Inventory Review Worksheet
- Distributor Vendor Performance Scorecard Worksheet
- Distributor GMROI Calculator
- Distributor Backorder Cost and Fill Rate Calculator
- EOQ with Vendor Minimum Order Calculator