The Forecast Analysis is an item/location specific analysis that analyzes the forecastability of a given item. As such, it is applied against items with independent demand only. If S&OP is enabled in your account, a similar analysis is done in the Forecast section of S&OP based on the forecasting structure deployed.
Even with S&OP enabled, this analysis provides valuable information to help users determine how best to plan a given item in a given location. It provides information relative to the demand history of the item and in turn based on an items classification, recommends how best to plan.
Forecast Analysis Column Definitions
- View: Click to view the items requirement planning details
- Number or Name: The item name or number depending on the value to use in system settings. Placing your cursor over the value will show the item name or number depending on the primary view option chosen.
- 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.
- BI: Business Importance of the item / location. Calculated in Classification Analysis. 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. This also reflects if an item is assigned a User Defined Class.
- F: Forecastability. This metric considers the value of CoV to determine the type of forecast approach to apply to the item. This is based on item/location even if S&OP is enabled. Threshold is based on the CoV threshold system setting. One letter H or L based on the CoV threshold.
- Age (Wks) How old the current item is. The count of weeks to the first week where demand is created not including linked history. Please note that if you see 6208 it means the item has not data to calculate.
- Wks No Demand: How many consecutive weeks from the current period backwards an item has had no independent or dependent demand.
- Linked New. Does the item have historical linking. Yes or No. From item data maintenance.
- Forecast Acc. % New. This will be a new measure based on independent demand forecast only. The calculation is: (1-(ABS(Actual Independent Demand - Forecast Independent))/Forecast Independent Demand))*100
- Actual Demand: Actual independent demand over the measurement period. The measurement period is the items lead time or the last 5 weeks, whichever is greater.
- Forecast Demand: Independent forecast demand over the measurement period. The measurement period is the items lead time or the last 5 weeks, whichever is greater.
- Demand Type: Type of demand. Types are independent demand from direct customer sales, dependent demand from either a BOM explosion or inventory transfer, or both
- Bucket Type: Specifies if the forecast is based on weekly (w) or monthly (m) buckets.
- MAPE: Mean absolute percent error
- RMSE: Root mean squared error
- Std. Dev.: The measure of the volatility of independent demand over the safety stock time frame setting.
- CoV: Coefficient of Variation. This is Standard Deviation / 52 W average independent demand.
- ADI: Average Distribution Interval. Measure of spread of demand. If there is no 0 demand during the measurement period, the result equals 1 since 0 divided by 0 equals 1. The threshold between smooth and intermittent demand is 1.32.
History Type: This uses the measures of ADI and CoV to state the type of demand history an item has. The results are Intermittent, Lumpy, Smooth, and Erratic.
- Smooth demand (ADI <= 1.32 and CoV <= 0.7): The demand is very regular in time and in quantity.
- Intermittent demand (ADI > 1.32 and CoV <= 0.7): The demand history shows very little variation in demand quantity but a high variation in the interval between two demands.
- Erratic demand (ADI <= 1.32 and CoV > 0.7): The demand has regular occurrences in time with high quantity variations.
- Lumpy demand (ADI > 1.32 and CoV > 0.7): The demand is characterized by a large variation in the quantity of demand and in the interval between two demands. This is quite impossible demand to reliably forecast, no matter what forecasting tools and methods are being used.
- Forecast Model: The type of statistical forecast model applied when S&OP is not enabled. If S&OP is enabled, the specific forecast model is not provided since the source is a sum of context based forecasts for the given item. In this case, the model is S&OP.
Approach: Recommends how best to forecast a given item item using BI and Forecastability. A items are considered H value and B and C low. See further information below.
- Collaboration: If the item has an A classification and CoV > 0.7 the planning method should be collaboration
- Statistical: If the item has an A classification and CoV <= 0.7 the planning method should be statistical
- Min/Max: If the item has a B or C classification and CoV > 0.7 the planning method min/max
- Statistical: If the item has a B or C classification and CoV <= 0.7 the planning method statistical
- Trend Alert: Describes the rate of change of demand by comparing the percentage difference between the last 13 week average demand and the last 52 week average demand. Hyperlinked to Classification Analysis where it is calculated. If the change is less than or equal to 15% the item receives an "OK" designation. If the change is between 15% and 50% the item receives a "Concern" designation. If the change is greater than than 50% the item receives an "Alert" designation. The + or - after the designation communicates if the trend is positive or negative i.e. negative means the 13 week average is less than the 52 average.
- Stability Index: Compares the last 13 week independent demand against the next 13 week forecast demand.