The Segmentation Analysis process is an important element of DemandCaster. It is built on top of the Classification Analysis to help direct an organization's attention and efforts toward the company's most significant items. Each item a company manages from finished goods through components are ranked on the basis of Business Importance, Order Frequency, and Accuracy.
We recommend running a full segmentation on a monthly basis just after the close of the previous month. This process saves historical data that can later be used to track actual to plan in the item detail view. Partial segmentation may be run with each data upload via a setting in the system settings. A partial segmentation only updates the OF ranking of the item and not the accuracy.
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.
- Click "Run Analysis" to commence the Segmentation Analysis. If a requirement plan has not previously been run, please run the requirement plan first as documented in the article Running a New Requirement Plan (when S&OP IS NOT enabled) or Running a New Requirement Plan (when S&OP IS enabled)
- BI Overall: Overall business importance across all locations when DRP module is enabled. Calculated in Classification Analysis.
- BI Rank: The item / location (If DRP is enabled) rank of the item calculated during Classification Analysis. The lower the ranking the higher the business importance. This also serves as the default sort order for all analytics and selection interface.
- BI Rank Overall: The overall item (If DRP is enabled) rank of the item across all locations calculated during Classification Analysis. The lower the ranking the higher the business importance.
- OF: Order Frequency. This is a measure of how often an item is ordered within a location (if DRP is enabled) over the past year (customer order line items). Components adopt the order frequency of their finished goods. For example, if an item is a component of three finished goods that are each ordered one time, the component will receive 3 orders whereas the finished goods will receive one each. This places the ranking of the component higher than each of the three finished goods.
- OF Overall: Overall OF across all locations for an item. Only when DRP is enabled.
- OF Rank: Order Frequency Rank. This is the numerical ranking of the OF calculation by item / location (if DRP is enabled).
- OF Rank Overall: Overall OF Rank across all locations for an item. Only when DRP is enabled.
- A: Accuracy. Items with greater than 80% accuracy are considered High (H), between 50% and 80% are considered Medium (M), and below 50% are considered Low (L). The basis is the Actual to Forecast Accuracy calculation below.
- Demand Type: Presents the demand stream of the item. I = Independent Demand. D = Dependent Demand. I&D = Independent and Dependent Demand. As a reminder, I demand is direct customer order demand via a sales order. Demand type is a calculated value that looks at item, and if applicable, locations with sales orders first. These are then designated as independent demand locations. The dependent demand locations are calculated by looking at if the independent demand is passed down to a different item / location. This is done by looking at the bill of material and also the source / destination relationship in the item location text file.
- Actual Demand: Used during the Accuracy calculation. Displays the actual demand of the item over the items lead time or the last 5 weeks, whichever is greater. This applies only to sales orders.
- Forecast Demand: Used during the Accuracy calculation. Displays the actual demand of the item over the items lead time or the last 5 weeks, whichever is greater.
- Act. to Forecast Accuracy: Accuracy measure of forecasted demand versus actual demand over an items lead time or 5 weeks whichever is greater. This analysis is focused on lead time accuracy over period to period accuracy since the lead time is the basis for the quantity that needs to be ordered. The basis for this calculation is the forecast. The calculation is: (1-(ABS(Actual-Forecast))/Forecast))*100
- Calc Order Period (EOQ): Calculated as EOQ/Average Weekly Demand. As a reminder, EOQ is calculated as SQRT((2*Average Weekly Demand*Ordering Cost)/(item cost/weekly cost of inventory percentage)).
- Order Period: User defined order period in weeks. How many weeks of average demand to order at a time.
- Periodic Order Quantity: Quantity to order at one time based on the Periodic Order Quantity basis selected.