Like YOU, We think of TIME, DATA, and INVENTORY as MONEY... Managing them successfully: Pure Profit       









Three recent jobs for ICS had a reoccurring issue with how often they re-forecast their items.

The math concepts for re-forecasting all contain a variable for standard deviation in any supply chain system. How this math is applied in a system will dictate what results the system drives down to the store or warehouse. The misuse of the reforecast feature and conceptual misunderstanding about the math involved creates company expectations that are grossly inaccurate. In each of the companies examined several occurrences of both Out of Stock and Over Stock issues were creating horrible financial results.

Any supply chain system used, E3 or other, has a collection of components and parameters that need to be set to each individual companies business needs and goals. Standard settings often applied at implementation fail to achieve the long term goals for most companies. Conceptual and technical knowledge are necessary next steps toward successful achievement of company goals and expectations.

Weekly vs. 4Weekly re forecasting

Upon review of your company level forecasting accuracy report you identify a large part of your item location business with lumpy or slow demand, then a review of your MADp should be the next step.

Recommendation: With lumpy demand and a high MadP, a 4-week reforecast will provide better service, lower inventories, and less buyer maintence.

Each time an item is reforecast in E3, a new deviation is calculated. The deviation is the difference between the forecast and the actual sales, expressed as a whole number. The E3 system creates a moving average of the results called Mean Absolute Deviation expressed as a percent (MadP). The system is measuring how lumpy your sales demand is in relation to the forecast over a period of time. Lumpy sales between reforecast will increase the MadP, the effect increases safety stock forcing you to carry more inventory which results in buyers questioning the validity of the system. Another result of a high MadP is increased history exceptions on the low sales side and possible missed sales opportunity due to the lack of a history exception flag on the items with increasing sales demand.

The MadP is an effective measure of how accurately you are forecasting your business over a three to six month time period. While the forecasting accuracy report in E3 will tell you the accuracy of the previous period sales versus forecast, the MadP provides a measure of how dependable the forecast is over a long period. This is often a critical measure for items that are fashion driven, or maintaining long lead times, or have constrained supply. Often E3 assigns these items a system class of slow or lumpy.

Lumpy Demand creates high safety stock and additional history exceptions that may fail to address the root issues. The Company Control settings have the history exception switch set to 5 deviations on the high side and 3 deviations on the low side for the MadP. These are the E3 default settings, there is no data currently to suggest a change to these settings. The key is to manage your forecast at the location level and determine which items need weekly versus four weekly periodicity (reforecast).

Review the following examples to help you determine best practices for your company.

Here is an example from a recent company ICS worked with:
In the first example using weekly reforecast, the MadP is 62.85%.
The weekly forecast for the item is 6.
A history exception will ONLY be incurred if weekly sales are above 24.86 or below 1.
62.85% * Forecast 6 = 3.77
3.77 * 5(company controls) = 18.86.
Forecast 6 + MadP measure 18.86 = 24.86.
Weekly sales must be above 24.86 for a History Exception check on the high side.
This is 400% greater than the monthly demand.

62.85% * Forecast 6= 3.77
3.77 * 3(company controls)= 11.31
Forecast 6 - MadP measure 11.31 = 0
Weekly Sales must be 0 for a History Exception check on the low side.

Weekly re-forecasting
Week Forecast Ships Standard Deviation MadP “X” Variation
WK1 6 13 4.9497 0.8250
WK2 6 0 4.2426 0.7071
WK3 6 12 4.2426 0.7071
WK4 6 0 4.2426 0.7071
WK5 6 14 5.6569 0.9428
WK6 6 12 4.2426 0.7071
WK7 6 6 0.00000.0000
WK8 6 0 4.2426 0.7071
WK9 6 5 0.70710.1179
WK10 6 0 4.2426 0.7071
WK11 6 12 4.2426 0.7071
WK12 6 0 4.2426 0.7071
MadP 62.85%

Here is an example of a similar item with a periodicity set for four week reforecasting: Using 4-week re-forecasting, the same item, same time period, the MadP is 15.71%. The 4-week forecast for the item is 24. (weekly 6 * 4weeks = 24) A history exception will ONLY be incurred if 4-week sales are above 42 or below 16. 15.71% * Forecast 24 = 3.77 3.77* 5(company controls)=MadP measure. Forecast 24 + MadP measure 18.852 = 42.852. Weekly sales must be above 42 for a History Exception check on the high side. This is only 75% greater than the monthly demand. 15.71% * Forecast 24= 3.77 3.77 * 3(company controls)= 11.31 Forecast 24 - MadP measure 11.31 = 12.69 Weekly Sales must be 12 or less for a History Exception check on the low side. 4 week re-forecast

Week Forecast Ships Standard Deviation MadP “X” Variation
Month1 24 25 0.7071 0.0295
Month2 24 32 5.6569 0.2357
Month3 24 17 4.9497 0.2062
MadP 15.71%


The analysis clearly shows that weekly re-forecasting is
  • Increasing your safety stock requirements.
  • Increasing History Exceptions on the low side
  • Missing History Exceptions on the high side.
  • Creating a validity/ accuracy question in E3. The effect is both overstocks and out of stock on best sellers at the same time. The buyer is not warned soon enough of the possible out of stock due to ships being above the demand forecast significantly over time.

    The Weekly Reforecast function in E3 is often mis-used. While the intention is to “watch” item sales more closely, the effect can be disastrous to the overall performance in the supply chain. While a “Zappo” can be used to decrease the MadP, the effect of a stand alone “zappo” would only hide the reoccurring issues of how you forecast and how you manage your inventory process.

    The weekly re-forecasting was designed for items with very high sales. The re-forecasting switch can be set at the item or vendor or company level. The system is designed to meet the needs of the individual items within the constraints of how a company wants to manage the inventory.

    *****
    This information is for review only. Any actions taken by an individual or company based upon their review of the information on this website is the sole responsibility of the individual and/or company; their decision entirely. ICS and its affiliates take NO responsibility or warranty for the actions taken by companies or individuals based on this information.

    *****
    "The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential, proprietary, and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from all computers."

    ICS encourages companies to contact ICS directly to learn how this information can benefit their Supply Chain solutions.
    www.InventoryGuys.Com
    Office 770-995-5054
    • Copyright 2003 Inventory Consulting Specialists 770.995.5054 ICS@InventoryGuys.com