The Classification Analysis in DemandCaster follows the Pareto Principle with one significant enhancement for companies that manufacture, assemble, or kit finished goods - the analysis is extended through the entire bill of material. In addition, when the DRP (Distribution Requirements Planning) module is enabled, the analysis is location specific to allow each item within a location to be ranked relative to its location.
We recommend running a full classification on a monthly basis just after the close of the previous month.
Please note that if an item is deactivated or disabled, the ranking is not updated to allow other items to take its place. If the item was ranked 1, for example, and was deactivated, the ranking will now start at 2 until the ranking changes as new items sales start exceeding the previously deactivated items sales.
In DemandCaster, customers may be set up as multi-currency in order to view sales of the customer in their local currency. Location based currencies are based on the companies standard currency in order to ensure the classification analysis is performed with a single currency so all items and locations are weighted the same.
There are four classification options that may be applied to an item that may be viewed by an items independent demand or independent and dependent demand. By definition, independent demand is sales that the item independently generates where as dependent demand are sales that the item is helping to indirectly generate because the item is a component. The basis of the classification is set in the System Settings Classification Basis option.
- Revenue/Sales: Based on the total sales dollars generated by a given item by location (multi-location / DRP is enabled) and overall. This is calculated as the total sales quantity by line item for the given item times the sales price by line item. Though a dependent item does not have any independent sales it inherits the total sales it is responsible for generating.
- Margin: Based on the total margin dollars generated by a given item. This is calculated as the total sales quantity by line item for the given item minus the cost of the item by line item.
- Cost of Sales: Based on the total sales cost dollars generated by a given item.
- Units: Based on the total units sold for a given item.
By taking this approach DemandCaster helps answer the question, regardless if the item is directly or indirectly responsible for generating sales, how much revenue, cost, or margin is put at risk if a specific item is not available for sale or production. This way, stocking decisions can be made relative to an items overall contribution to the financial performance of the company.
Running a Classification Analysis
First, the user must define the basis of the analysis by selecting revenue, margin, or cost in the System Settings.
To run a Classification Analysis click on the “Run Analysis” button as shown below. There is no need to select items since the classification analysis is automatically calculated for all items.
There are two iterations of Classification analysis being performed in sequence based on an items independent versus dependent demand.
- Independent (I): Revenue, margin, or cost generated as a result of direct customer sales.
- Dependent (D): Revenue, margin, or cost inherited from direct customer sales through the bill of material.
- An item may have both independent and dependent sales. The total of this is designated as I&D in the classification analysis.
The first iteration calculates the ABCD using either revenue, margin, or cost for items with independent customer demand. The second iteration extends the independent demand down through the components. The two iterations are summed and then paretoed as follows:
- A Items: Top 80% of sales, margin, cost, or units.
- B Items: The next 15% of sales, margin, cost, or units.
- C Items: The remaining 5% of sales, margin, cost, or units.
- D Items: Items with no demand or consumption over the last 52 weeks.
It is feasible for a component item with 0 independent demand to be ranked as an A if it carries a significant cost component or is widely used across many independent demand items.
For multi-location situations, the following logic applies:
- Independent demand for a Manufacturing and Distribution Center (a warehouse that feeds another warehouse where the sales are generated) locations is the Manufacturing / Distribution Centers locations direct sales only.
- Independent and Independent (I&D) demand for a Manufacturing / Distribution Center location is the sum of the locations independent sales, the sales of the warehouses which belong to this location and the pass down from the finished goods.
The classification analysis also includes trend indices. The A,B,C code is followed by two symbols.
The first symbol designates an overall trend for the item comparing the items rolling 13 week average to its rolling 52 week average. The second symbol designates the direction of the trend since the last analysis or upload was performed (minimum of a weekly period). The symbols are either two plus signs (++), two minus signs (--), two equal signs (==), or a combination of the three signs. Coupling this information with the significance of the change as defined in the Alert, Concern, and OK Demand Change Status alerts, the user can use this information to apply safety stock policies or adjust forecasting. The trend indices are as follows:
- The first +, -, or = sign means the 13 week average is either increasing, decreasing, or holding flat relative to the 52 week average.
- The second +, -, or = sign means the trend as compared to the last analysis (minimum of 1 week span) is either increasing, decreasing, or holding flat.
Thus, an A++ classification means a top 80% revenue, margin, or cost generating item is on a continual upward demand trend. Such an item should be monitored to assure the forecasts are accommodating such growth and there is sufficient safety stock to assure there are no stock-outs during the replenishment lead time.
As part of the classification analysis, the forecastability of an item within a location (when applicable) is also calculated. Forecastability is a measure that helps a user understand if an item has a profile that is relatively easy or hard to forecast. The formula is F = standard deviation / 52 avg independent demand. An item is relatively easy to forecast if the value is 100% or less. This is represented with the letter H. It is difficult to forecast if the value is greater than 100%. This is represented by the letter L. The result is N/A if the item has 100% dependent demand.
The result of the calculation is viewable in the Segmentation Analytic, Forecast Analytic, and Item Selection.
The following described each analysis output column in the Classification Interface. Columns are sortable by clicking on the header. Default sort order is top down business importance ranking. You may filter the data to view independent demand or dependent demand by clicking on the buttons in the filter and select menu bar (see arrows in image below). You may also download the data as a report by clicking on the excel button in the filter and select menu bar.
- Details: Click to view the items requirement planning details
- Item: The item number
- Name: The item name
- Location: If the DRP module is enabled, the location of the item. This is a physical geographic location and not an inventory location within a warehouse.
- Product Category: This is the items product class.
- BI: Business Importance of the item / location. Top 80% revenue, margin, or cost of items are A, next 15% are B, remaining 5% C, and D's are items with no demand over 52 weeks.
- Calc. Class Revenue: The revenue based result for business importance.
- Calc. Class Margin: The margin based result for business importance.
- Calc. Class Cost: The cost based result for business importance.
- Calc. Class Units: The unit based result for business importance.
- User Defined Class: The classification assigned to the item / location by the user. This is often the case when the user wants to force an item to take on a higher or lower service level compared to what was calculated for the item.
- 52W Demand: The items total independent or dependent demand over the last 52 weeks. I&D is the sum of both Independent and Dependent demand. A component with no independent demand will have 0 I demand.
- 52W Avg. Demand: 52 average of the above
- 13W Avg. Demand: 13 average of the above
- 52W Sales: The items total independent or dependent sales revenue over the last 52 weeks. I&D is the sum of both Independent and Dependent sales. A component with no independent demand will have $0 I sales.
- 52W Cost: The items total independent or dependent cost over the last 52 weeks. A component with no independent demand will have $0 I cost.
- Margin $'s: The items total independent or dependent margin over the last 52 weeks. I&D is the sum of both Independent and Dependent margin (Revenue - Cost). A component with no independent demand will have $0 I margin.
- Avg. Qty on hand: Average on hand quantity since the DemandCaster process started. DemandCaster stores the on hand value with each upload and calculates the average quantity over time.
- Avg. Turns: Last 52W Average Demand I&D / Average Qty on Hand based on the items independent or independent and dependent demand. It is most suitable to calculate turns based on combined independent and dependent demand.
- Current On Hand $'s: On hand $’s as of the latest upload (current on hand times current cost).
- Current On Hand: On hand units as of the latest upload.
- Current Turns: Last 52W Demand I&D / Current Qty on Hand.
- Avg Contr: Average Contribution. Margin $ - (Average Quantity On Hand * Cost of Inventory Percentage defined in system settings).
- Tag: User defined tags assigned to the item.