Module: Analysis of stockouts

Executive summary

The module identifies and subsequently analyzes periods with no stock on hand, i.e. stockout. This data is then used to compute lost sales at stockout of a specific item in your warehouse.

Thanks to this, we are able to:

• Evaluate losses in revenue and profit that the company experiences due to the stockout – and to compute a desired service level.
• This module is an input to the module: Managing customer service levels.
• Straighten historical sales and provide thus an appropriate input to sales forecast. The sale forecast, clearly, must reckon with the fact that sales would have been kept on should there be no stockout.

Historical data obtained from the module Analysis of stockouts are the primary method of computing the customer service level. ModuleManaging customer service levels.

Functional description

Identifying stockouts

STOCK can estimate quantity of a product to be sold within a certain period of time (including seasonal products and promotions). Furthermore, it also estimates – using the average purchase and sale prices at the given time – lost revenue and profit.

The module can be set so that a certain percentage of sales lost in stockouts is computed in the data used for sales forecast. Stockout analysis thus prevents that sale history – deformed due to the stockout – is negatively reflected in future forecasts.

A stockout is not only zero quantity of a given product in a warehouse, but also a level of stock lower than the average daily sales.

At the same time, however, in some cases it can be set (mathematically or managerially) not to count the situation as a stockout despite the fact that there was zero stock of the product. There are 2 possible cases:

Seasonal stockout

If a defined minimum sale forecast occurs at least 2x in the same month in two different years, this sales forecast is considered seasonal – and thus ignored.

Too long a stockout

If sales forecast exceeds the maximum defined period, such a case is taken to be due to other causes than stockout – and as such, it is ignored when sales forecast is computed.

Computation of lost quantity (of sales)

The lost quantity of sales (caused by a stockout) is computed based on various components, if these are available. If none of the components can be traced, the sales forecast is erased.

Otherwise, the first half of the forecasted quantity of sales is computed from the neighboring months (the amount of months to be considered can be set). The period for which the sales forecast is being set is divided into three equal parts. If any of the thirds contains more than 60% of the sales in the given period, it is only this third that is used for the calculation of the sales forecast. This method ensures that the forecasted sales are based on the period as closely related to the stockout as possible.

The second half is computed from the sales in the neighboring years with a minimum share of 60% of sales days. Moreover, 25% of those sales is taken from the year before the sales forecast a 25% is taken from the year after the sales forecast. Thanks to these restrictions, the computation of the lost quantity of sales takes a seasonal component into consideration as well.

Promotion stockouts

If the module Analysis of promotions and forecast of promotion sales is active, then stockouts for promotions and outside promotions are automatically identified. This holds also for overlapping promotions.

A big stockout is divided into several sections, as indicated in the following figure.

Subsequently, each stockout is assigned to an appropriate promotion (always to only a single one of them; in case of overlapping promotions, the promotion that started later is selected). The lost quantity is automatically computed only from the given promotion and from the time when the product was not in stockout.

Extensions

Setting stockout levels for each month using the SIDI

The basic module is able to identify the average daily sales and – based on that – also the level of stock defined as stockout. If the (level of) stock at the beginning of a day is lower than the average daily sale, it is likely that during that day, there will be a stockout.

This extension can change the (automatically computed) stock level for a stockout, moreover for each month separately. The required stock level is imported from your ERP system via a data bridge.

There are two different scenarios for which the extension can be useful

1. We deal with a highly seasonal product: The threshold for a stockout during the season is completely different from the threshold for a stockout off-season.
2. We deal with a product with considerably lower sale if the quantity of the product is low: A typical example is a product bought largely by wholesalers (who always require a huge quantity of the product, say 500 or more pieces at a time). If the product is not available at your warehouse, the wholesalers won’t purchase anything. Yet, there is always a small percentage of customers who purchase smaller quantity of the product than wholesalers. If we take them as the base for the stockout computation, the stockout will occur at much a lower level than required, therefore we need the extension to enforce the higher levels for a stockout manually.