Controlling Purchase Runs and Min Max Update Runs using Readysell statistical measures
Overview
The most important aspect of forecasting demand is knowing when the data is not sufficient to calculate a result. Our statistical measures round, average, and filter the demand history so that the system can skip forecasting some products that have lumpy demand, one large sale, etc.
You probably need to adjust a combination of those statistical measures to get the result you need. Increasing the sales variance threshold will reduce the effect of a few large sales. But you may need to adjust the minimum weekly sales count to filter out sales where there are only a few entries in the demand history. Weeks of floor stock and minimum weekly sales quantity are also important and are two of the easier to understand measures.
The statistical measures
- 'Minimum weekly sales count' = each week the product within this run has to have at least this number of sale lines for that week (.9 is less than 1 sale).
- Use case: If you select a period of time covering eight weeks and the product only sells twice in that period. The average weekly sales is (number of sales/number of weeks) = 2/8 = .25. If you set the minimum weekly sales count to 0.9, that product would be excluded from the calculation
The “Minimum Weekly Sales Count” parameter would get the result you need, supplementing the “Avg Weekly Sales Variance Threshold”.
- Use case: If you select a period of time covering eight weeks and the product only sells twice in that period. The average weekly sales is (number of sales/number of weeks) = 2/8 = .25. If you set the minimum weekly sales count to 0.9, that product would be excluded from the calculation
- 'Weeks of floor stock' = number of weeks worth of stock that you wish to keep
- 'Minimum weekly sales quantity' = minimum quantity that must be sold each week before the product can be ordered (0.5 at least sell a quantity of half)
'AVG weekly sales variance threshold' = if in any week the quantity sold varies from the average by this amount then that week's sales will be ignored. This acts to stop once-off large sales from distorting the overall average quantity sold for a product.
Worked Example
the system can not determine which sale is the correct qty and which isn’t. When you extract a min-max run what the best practice is that you enter 3 in the “Avg Weekly Sales Variance Threshold” so that the system will be able to determine if there is a sale that is way above the average sold.
I have entered an example below for you. If the grid below only had 2 sales in it and the total sales will 10 and the cut-off would be 2 the average would be 5 * 2 the new average would be 10 which means it will recalculate the max on the product again.
More then 2 Sale
2 SALES
