History Horizon in forecasting

Hi All,
I have given the 6 months duration for  historical data in the forecast profile. We are also using fiscal variant here.
When I run the univariate forecast profile I see some forecasted data.my confusion is when I click the tab "Horizon" and there if I change the "from" field in history and after that I run the univariate forecast - it is not considering my changes.
The fiscal variant we are using is for 4-4-5 calendar.
Thanks a lot in advance,
Prabhat

There could be sevaral resons
1. Make sure you have selected the 'Period Intervels' radio button and not the 'Number of periods'.
2. Check if there is any data for the horizon you selcted
3. It may also depend on the forecast model that you chosen. For example constant model may show you the same results if the history data is same for the horizon you selected as that of preivious seclection. To test this select a different model and execute the forecast and check if the results have changed

Similar Messages

  • Copy history as the forecast result

    Hi Experts,
    Has anyone ever use the forecast model that copying the past 12 months history as the future period forecast result in PP forecast model?
    I know this function is availbale in APO DP, But I want to use this function in R3 not in APO.
    thank you in advance!

    Hi, here is the steps by step procedure to create SOP flexible plan along
    with trasaction codes. Let me know if you have further problems
    S&OP Flexible Planning Checklist
          Step #
         Description
          1
         Create Field Catalogue MC18
          2
         Create Info Structure MC21
          3
         Create Update Rules for Characteristics and Key Figures MC24
          4
         Activate Update Rules OMOZ
          5
         Testing (Use Sales Order, Purchase Order, Material etc.) to test data
    MC30
          6
         Set Parameters for Info Structures MC7f
          7
         Create Planning Type for Info Structure Planning Table MC8A
          8
         Generate Master Data and generate proportional factors MC9A
          9
         Create Planning Hierarchy MC63
          10
         Create Forecast Profile MC96
          11
         Create Planning Activity MC8T
          12
         Setup Mass Processing Job to generate forecast MC8D then MC8G
          13
         Maintain all Material Master records with unit of measure conversions
    so that the planning table can aggregate materials into a common unit of
    measure. MM02
          14
         Apply OSS patch program that enables visibility of all nine hierarchy
    planning levels in planning table. (Check version 4.5 and prior)
          15
         Set up program to collect Historical transaction information (Need
    Custom ABAP program to populate the SOP table with historical data)
          16
         Run Monthly Sales Forecast MC8D
          17
         Transfer to COPA (if applicable)
          18
         Run COPA Analysis
          19
         Transfer forecast to Demand Management MC94 (Extra Menu -> Transfer to
    DM)
          20
         Demand Management receive the updated Sales Forecast quantities MD61
    -Farrukh

  • Moving average forecast does not work correctly

    Hello,
    I have a forecast model with the strategy moving average 6 months. When I checked the forecast values in the planning book, I recognized that there is a mismatch with the calculation of the forecast:
    There is a drift of 2 months when showing the values in the appendant buckets. The first bucket which is filled with forecast values (bucket: current month + 1) contains the moving average value from month - 8 until month - 2.
    Example: forecast value for Feb 12 is calculated as follows:
    (Nov 11 + Oct 11 + Sep 11 + Aug 11 + Jul 11 + Jun 11) / 6
    It is the same procedure for all other following months.
    The right calculation should be as follows:
    (Jan 12 + Dez 11 + Nov 11 + Oct 11 + Sep 11 + Aug 11) / 6
    I checked several settings in the activity and job creation and in the setting of the forecast profiles. I expected that an offset of 2 months is set, but there is nothing.
    So, I do not know how to explain this system behaviour....
    Hope anybody can help!
    Thanks and kind regards
    Heinz

    Hello Heinz,
    The only place to check is the forecast profile. Let me assume that you are not talking with reference to SPP. Also, that the period indicator in your master profile is Months. So, this would mean that you have maintained forecast profile as follows:
    You need to cross-check that the offset in the section history horizon of the master profile should have value +1, since you are running the forecasting during January, and you want January history to be considered in average calculation. Periods in history horizon should be 6.
    Similarly, in the forecast horizon in the master profile, you should have an offset of +1, since you are running the forecast in Jan, but you want the forecast to happen from the month of Feb onwards.
    Also, check in the univariate profile tab that you are checking for the right KF for history values, when you look in the planning view. 
    There is nothing more to standard forecasting apart from few things like 'like profile' or 'phase-in, phase-out' profile. Other option for changes is some custom logic, but it doesn't seem that there is a need for it if calculation is for simple moving average.
    So, if the above points don't explain your situation, then better to raise OSS for SAP, to see if there could be some bug.
    Thanks - Pawan

  • Basics about the forecast key figure

    Hello
    There are lot of Key figures in use -
    Statistical Forecast
    Ex-Post forecast
    Corrected forecast
    Adjusted Sales history
    Any idea what each stand for ?
    Thank you in advance
    KK

    Hi
    Statistical Forecast : This is system generated FCST based on FCST Technique used
    Ex-Post forecast:   Ex-Post Forecast is just calculating the forecast in the history horizon .This is calculated through the forecasting technique adopted by you .
    Corrected forecast : After system has created Stat FCST, we can manipulate as per our requirement and this is called as correct FCST. 
    Adjusted Sales history :- It regarding the clearing the sales history, In case of some business extreme condition we have some peaks and down in sales history. We require to correct them , to get appropriate Statistical Forecast. This corrected sales history which is also called as adjusted sales history get used for Forecasting.
    Thanks
    Amol

  • What is ExPost Forecast ? how to calculate it?

    Hi Frnds
    I would like to know more about Ex-Post Forecast Method kindly let me know the clear details about this ?

    Ex Post Forecast is generated by the system during the statistical forecast calculation and only appears in historical periods.  It shows you how the system is interpreting the demand history as it processes it.  In this sense, it shows you how the system is "learning" the history.  All forecast error metrics are calculated by comparing the ex post fcst to the demand history, so lower forecast error numbers mean the forecast profile assigned to the selection fits the history better. 
    In our implementation, we have included a key figure in the planning area for the ex post forecast so that the planners could see the comparison of ex post to history in the dataview graph.  This lets them see at a glance how well the model traces the history.
    You may also want to include a key figure for Corrected History in the planning area and in a dataview.  The system actually copies history to 9AVCORHIST at runtime to use as the basis of the statistical forecast, and the results of background processes like Lifecycle Planning (Like, phase-in and phase-out profiles) and promotion planning show up here -- this makes it easier to trace what the system is doing and how it arrived at the statistical forecast.

  • Error FCEXEC in creating forecast

    I am trying to create forecast in AWM with foodmart database. There are 5 dimensions, dim_timeall, dim_customer, dim_product, dim_store and dim_promotion, with 1 cube cub_salesfor. I followed example in OLAP Application's Developer Guide 10.2. Dim_timeall includes all of history period and forecast period with day as base level.
    First trial :
    I made 2 measure with short decimal type and then I used them in fcexec :
    FCEXEC handle TIME dimtimeall INTO forecast_sales SEASONAL forecast_sales_seasonal cub_salesfor
    After I compile, there were error message :
    ERROR: (ORA-33686) The time-series parameter to FCEXEC is not dimensioned by AW_20MAR!DIM_TIMEALL.
    Second trial :
    I made 2 measure with short decimal type and I selected DIM_TIMEALL, DIM_TIMEALL_HIERLIST, DIM_TIMEALL_LEVELLIST, DIM_TIMEALL_TIME_START_DIM and DIM_TIMEALL_TIME_START_DIM1 in tab Dimension, and of course with the same fcexec.
    But, I got the same error message.
    Somebody help me, please...

    Check the following:
    * forecast_sales is dimensioned by dim_timeall
    * forecast_sales_seasonal is dimensioned the same as forecast_sales. It may have extra dimensions, but it must have all the dimensions of forecast_sales. If it has extra dimensions, the status of the extra dimensions must be limited to a single value.
    * cube_salesfor is dimensioned the same as forecast sales. It may have extra dimensions, but it must have all the dimensions of forecast_sales. If it has extra dimensions, the status of the extra dimensions must be limited to a single value.
    There is usually an SMSEASONAL parameter, but I don't know if it is required.
    An example forecast program follows. It might be useful as a template.
    " The forecast engine needs a scaler integer variable to be used for
    " the forecast handle. Define the a variable if it does not already
    " exist.
    IF EXISTS('FORECAST_HANDLE_UNITS') EQ no
    THEN DEFINE forecast_handle_units VARIABLE INTEGER
    " Set status of dimensions to the values the will be forecasted. This
    " forecast will be run as the lowest levels of the dimensions.
    PUSH time channel product customer
    LIMIT time TO time_levelrel 'MONTH'
    LIMIT customer TO customer_levelrel 'SHIP_TO'
    LIMIT product TO product_levelrel 'ITEM'
    " Limit channel to some specific values.
    LIMIT channel TO '2' '3'
    " Run the forecast. Note that the variables for the SEASONAL,
    " SMSEASONAL, and SOURCE measures are dimensioned the same as the
    " target.
    FORECAST_HANDLE_UNITS = FCOPEN('units')
    FCSET forecast_handle_units METHOD 'automatic' HISTPERIODS 79 PERIODICITY 12
    fcexec FORECAST_HANDLE_UNITS -
    TIME time -
    INTO forecast_units_cube_forecast_units_stored -
    SEASONAL forecast_units_cube_sys_forecast_units_seas_stored -
    SMSEASONAL FORECAST_units_cube_sys_forecast_units_smseas_stored -
    units_cube_units
    FCCLOSE forecast_handle_units
    SHOW 'Completed running forecast for method.'
    " Return status (constraints) of the dimensions to where they were before
    " running this program.
    PUSH time channel product customer

  • History cannot be entered in book

    Hi,
    i am getting following error while running a process chain.
    History cannot be entered in book, you cannot save data
    Message no. /SAPAPO/SDP129
    Diagnosis
    The history horizon of your planning book is not ready for input (you cannot make entries here).
    System Response
    Data that was calculated in this horizon cannot be saved.
    Procedure
    Change the history horizon so that it is ready for input. Make sure you always use different planning books for background and interactive planning.
    i am not able to understand the message.
    please help me what is to be done when i get this message. ( with T-codes if possible)

    Dear Chinmay Kulkarni,
    You get this message due to note 531274. It is an info-message stating that you can not save data in the past, because the past in the planning book is not ready for input. If you do not intend to change data in the past, then everything is okay.
    You can avoid the message if you simply define your dataviews in a way that only horizons which are really updated are used. SAP always recommend to created separate planning books / data views for batch processing. If you need the historical horizon in order to read data for the macro execution you are able to define the historical horizon as ready for input. In this case the message will not occur. Certainly the behaviour you are observing does not represent an error. The message is just informational und is raised correctly.                                                              
    I hope this helps.
    Regards,
    Tibor

  • Urgent!! How to update Planning book and change when planning will start

    Hi Gurus,
                 I had configured the system and we can initilize planning area every month and my question is 
    1. How to update planning book and change planning start date every month?
    Thanks a Lot
    Regards,
    Raj

    Hi,
    In the planning book-dataview, you will assign the time bucket profile, and will be on rollling basis. if you want the forecast profile on rolling basis, maintain the forecast and history horizon as periods then that will also be on rolling basis.
    Regards,
    Raj

  • Issues with statistical modeling in calendar months vs 13 fiscal periods

    In our demand planning, we plan at the fiscal period level (13 fiscal periods across a year - not matching with calendar months or year).  We are having issues (strange results) trying to run the statistical modeling at the fiscal period level versus the calendar month level.  Is there a way to run the stat modeling over fiscal periods rather than calendar months?  Can we run stat modeling over a specific date range rathn than including the current month?

    Hi Stephen, I guess I can assume the following:
    1) In the master forecast profile, you have maintained the period indicator and the fiscal year variant
    2) You are choosing a forecast strategy in univariate profile (?) that makes use of seasonality, and you are trying to maintain the number of periods in a season as 13 (in our case, our periods are custom, but there are still 12 periods, so different from your case)
    3) Your history horizon in master forecast profile is more than the number of periods in the season i.e. history has more than 13 periods of data
    If above things are already done and you still face issue when you try to use 13 periods in a season, looks to me like a SAP bug, and it would be best to raise an OSS message to SAP.

  • Lead time information in Demantra

    Hi
    For a POC, We need to get lead time data from ERP(R12) to Demantra. Can I use Lead time as one of the attributes in data model? is there any other better way to represent the same in Demantra?
    What are the inputs for forecast generation in Demantra (Other than Historical series, Causal factors and engine/non engine parametes)
    Thanks
    DNP

    Mismatch of forecast values having similar history on 7.2.0.2 VS 6.2.6
    Hi,
    check the following
    Hope it will help you
    Demantra - history period to be considered for forecast ???
    Forecast generation horizon
    Demantra Forecast clarification
    System Parameter
    Demantra - sytem parameter - max_fore_sales_date
    Thanks
    MJ

  • Seasonal Index in SAP APO

    Good Morning,
    I am having difficulty replicating the Seasonal Index for the Seasonal Model in APO DP.  Looking at a 36 month history horizon, I have been able to replicate the initialized seasonal indices for months -24 to -13 (with current month, 4/2015, as month zero).  This is documented to be the order history for the same month divided by the average demand for months -36 to -25.
    However, the next 12 months' seasonal indices (months -12 to -1) are not working according to the documentation (shown below):
    When working with a very low gamma (such as 0.01), my values using the formula are coming very close to the values that SAP is returning in the Exceptional Statistical Forecast Planning Book.  However, when I change the gamma to something higher (such as 0.30) in both my formula and the Exceptional Statistical Forecast Planning Book, the values are increasingly different.
    Anybody have any ideas on what I may be missing?
    I appreciate the help!
    Best Regards,
    Steven DeLisle

    Thanks Gurucharan for your response. That was quite helpful and I understood the conceltp.
    Varma

  • Time based KPI Calculation

    Hi All,
    I need to define a report where I need to calculate Supply - Demand for the complete horizon.
    Following is the example: -
    Week #                                     1               2               3                4             5                 6             7
    Sales Orders                                                            100                           150                             120
    Supply                                    500                                                                                400
    Projected Supply                     500             500        400               400        250                650          130
    How to calculate the value of ''Projected Supply'?
    Projected Supply = Projected Supply from Previous Week + Supply From Current Week - Sales Orders from Current Week
    Appreciate help on how to calculate this time dependent KPI.
    Regards,

    Hi Rajesh,
    I am working on SNC Customer collaboration scenario similar to your scenario.
    I am facing a small problem. The history sales data is imported and stored in VMIW1 time series data area. But while releasing the forecast to short term forecast screen though the forecast is dis-aggregated to daily form but the sales history is not visible here. I assume this is a standard SNC function to display the weekly sales history from sales forecast planing area to short term forecast screen, but unfortunately it's not showing. So I think I am missing something.
    Can you tell me whether I need to define another planning area for short term forecast again or is there any other standard configurations needed for it.
    Thanks & Regards
    Praty

  • Difference between expost and outliers

    Hi
    in the univerate fcst profile, under control parameter, Outlier correction is there, if we drill down this option then it will display Ex-post method and Median method options. My doubt is
    what is the difference between Ex-post Fcst and Outliers  corrections. How the system calculate the Expost and outliers? How the value comes in the ex-post fcst field in interactive plng once we generate the planning?Why ex-post fcst is there in front of OUtliers. is that both are same?

    Hi
    The Ex-post and Outlier is having totally different meaning.
    Ex-Post Forecast is just calculating the forecast in the history horizon .
    This is calculated through the forecasting technique adopted by you .
    Please find below the more information regarding.
    http://help.sap.com/saphelp_apo/helpdata/en/ac/216b77337b11d398290000e8a49608/content.htm
    Regarding the Outlier correction, it just cleans your history for plusminus
    tolerance calculated by the sigma .
    Please find below the more information regarding.
    http://help.sap.com/saphelp_nw70/helpdata/en/44/03961e0bbe025ce10000000a1553f7/content.htm
    Please let us know, if it helps you and also let us know if you have any more doubts regarding it.
    Thanks
    Amol

  • Error while intercative planning in SNP

    What we did is as below
    1.     created MOPS
    2.     created Planning area in DP
    3.     created time series and assigned version 000 in DP
    4.     Created Planning Book and Data view in DP
    5.     BW settings to extract data from OLTP and we have the data with planning version 000 in BW info cube
    6.     Load the history data from Info cube
    7.     Calculated proportional factors for next year
    8.     created Forecast profile.
    9.     Run forecasting after reading the history
    10.     Now forecast is ready in DP.
    11.     Released the same to SNP
    12.     When I goto SNP interactive planning and trying to load the data using 9SNP94 planning book/Data view i am getting the error - key figure not defined
    Can anyone tell me how to resolve the same?

    Dears,
    I tried all the steps given by you..but all in vain, I will explain you in detail what i ahve done..
    In /SAPAPO/MSDP_ADMIN - Administration of Demand Planning and Supply Network
    I have defined MOPS with following character(no SNP tick)
    9ALOCNO
    9AMATNR
    Defined planning area with following key figures
    SALE_HIST     Sales history
    9AREVFC     Sales Forecast
    9AMANUP     Manual Correction
    9ATOTFC     Total Forecast
    9APRICEFC     Planned Price
    T_PRICE     Price
    Storage bucket profile used is MONTH
    Created time series and it is valid up to 31.12.2009
    In BW, I have created a infosource/infocube with  9ALOCNO and 9AMATNR as character and SALE_HIST     Sales history as key figure and OCALMONTH as time character. Uploaded data from excel and now I can see the data in infocube.
    Now uploaded the data from BW to SAP-APO using - /SAPAPO/TSCUBE - Load Planning Area Version
    Calculated proportional factor using /SAPAPO/MC8V - Calculate Proportional Factors.
    In /SAPAPO/MSDP_ADMIN - Administration of Demand Planning and Supply Network, created chrematistics combination.
    Created forecast profile with UA profile
    Defined Planning book and data view
    Defined macros in /SAPAPO/ADVM u2013 MacroBuilder
    For calculating 9ATOTFC     Total Forecast = Sales Forecast 9AREVFC + Manual Correction  9AMANUP
    For calculating T_PRICE = 9ATOTFC * 9APRICEFC
    Now I goto /SAPAPO/SDP94 - Interactive Demand Planning and search for my PB
    I could see sale history filed is copied form BW.
    Run Forecast and got the forecast in 9AREVFC.
    Released the DP figures above to SNP using /SAPAPO/MC90 - Release to Supply Network Planning.
    In /SAPAPO/SNP94 - Interactive Supply Network Planning, used std planning book/DV u2013 SNP94(1).
    When I load I am getting the error as u201Cno key figure selection was foundu201D
    Any clue where i am wrong?

  • Subtraction of two key figures normalized to result not working as expected

    Hello SAP Community!
    I am having problems with getting the right result from a subtraction of two KFs which are "normalized to results" which means the KFs really have values expressed as percentages. The substraction that should be performed is of two percentages (e.g.: 87.298% - 85.527% = 1.77%) but my report prints out the result as "number of units" instead (e.g.: 87.298% - 85.527% = 71,514.00 EA). The two normalized KFs actually "point" to two stock KFs, hence the "number of units".
    In order to explain the problem I am facing please analyze below text:
    1) Let's assume I have below data:
    LOAD MONTH  PLANT    MATERIAL HORIZON MONTH     FORECAST UNITS
    200805         PLANT-A  MAT-1            200805         510,235.00
    200805         PLANT-B  MAT-1           200805          74,240.00
    200805         PLANT-A  MAT-1           200806         438,721.00
    200805         PLANT-B  MAT-1           200806          74,240.00
    200805         PLANT-A  MAT-1           200807         356,981.00
    200805         PLANT-B  MAT-1           200807          74,240.00
    200806         PLANT-A  MAT-1           200805               0.00
    200806         PLANT-B  MAT-1           200805               0.00
    200806         PLANT-A  MAT-1           200806         510,235.00
    200806         PLANT-B  MAT-1           200806          74,240.00
    200806         PLANT-A  MAT-1           200807         438,721.00
    200806         PLANT-B  MAT-1           200807          74,240.00
    2) Then, assume I have a comparison report, restricted by load month for two months May and June 2008 (filter restricted by two month variables) with FORECAST units spread accross columns for whole horizon (two months also in this case).
    Material  Plant                                 2008/06     2008/07
    ===================================================================
    MAT1      PLANT-A  
                       Base Units (May 2008)        438,721.00  356,981.00
                       Comparison Units (June 2008) 510,235.00  438,721.00
              PLANT-B  
                       Base Units (May 2008)         74,240.00   74,240.00
                       Comparison Units (June 2008)  74,240.00   74,240.00
              TOTALS   Base Units                   512,961.00  431,221.00
                       Comparison Units             584,475.00  512,961.00
    3) Now, let's suppose we want to know the proportions (%) of Base vs Comparison units, so
    we normalize forecats to results an we get the below report:
    Material  Plant                                 2008/06     2008/07
    ===================================================================
    MAT1      PLANT-A  
                       Base Units (May 2008)        438,721.00  356,981.00
                       Base Units % (May 2008)      85.527%     85.527%
                       Comparison Units (June 2008) 510,235.00  438,721.00
                       Comparison Units %(Jun 2008) 87.298%     82.784%
              PLANT-B  
                       Base Units (May 2008)         74,240.00   74,240.00
                       Base Units % (May 2008)       14.473%     15.702%
                       Comparison Units (June 2008)  74,240.00   74,240.00
                       Comparison Units %(Jun 2008)  12.702%     17.216%
              TOTALS   Base Units                   512,961.00  431,221.00
                       Comparison Units             584,475.00  512,961.00
    4) Finally, let's suppose we want to know the deltas (differences) of Base vs Comparison
    units, for both number of units and %. The report now look as below:
    Material  Plant                                 2008/06     2008/07
    ===================================================================
    MAT1      PLANT-A  
                       Base Units (May 2008)        438,721.00  356,981.00
                       Base Units % (May 2008)      85.527%     85.527%
                       Comparison Units (June 2008) 510,235.00  438,721.00
                       Comparison Units %(Jun 2008) 87.298%     82.784%
                       Delta base vs. comp. units %  1.77%       2.74%
                       Delta base vs. comp. units    71,514.00  81,740.00
              PLANT-B  
                       Base Units (May 2008)         74,240.00   74,240.00
                       Base Units % (May 2008)       14.473%     15.702%
                       Comparison Units (June 2008)  74,240.00   74,240.00
                       Comparison Units %(Jun 2008)  12.702%     17.216%
                       Delta base vs. comp. units %  -1.77%      -2.74%
                       Delta base vs. comp. units         0.00        0.00
              TOTALS   Base Units                   512,961.00  431,221.00
                       Comparison Units             584,475.00  512,961.00
    5) PROBLEM:
    In my report, the "Delta base vs. comp. units %" is not working as expected and
    calculates number of units just as "Delta base vs. comp. units" does instead of calculating the % difference.
    So my report looks as follows:
    Material  Plant                                 2008/06     2008/07
    ===================================================================
    MAT1      PLANT-A  
                       Base Units (May 2008)        438,721.00  356,981.00
                       Base Units % (May 2008)      85.527%     85.527%
                       Comparison Units (June 2008) 510,235.00  438,721.00
                       Comparison Units %(Jun 2008) 87.298%     82.784%
                       Delta base vs. comp. units %  71,514.00  81,740.00 <<<WRONG!!
                       Delta base vs. comp. units    71,514.00  81,740.00
              PLANT-B  
                       Base Units (May 2008)         74,240.00   74,240.00
                       Base Units % (May 2008)       14.473%     15.702%
                       Comparison Units (June 2008)  74,240.00   74,240.00
                       Comparison Units %(Jun 2008)  12.702%     17.216%
                       Delta base vs. comp. units %       0.00        0.00
                       Delta base vs. comp. units         0.00        0.00
              TOTALS   Base Units                   512,961.00  431,221.00
                       Comparison Units             584,475.00  512,961.00
    The formulas are:
    a) Delta base vs. comp. units %
      Delta base vs. comp. units % = Comparison Units % - Base Units %
    b) Delta base vs. comp. units
      Delta base vs. comp. units = Comparison Units - Base Units
    The KFs
    - Comparison Units %
    - Base Units %
    Are RESTRICTED key figures (restricted to Base and comparison month variables) which
    are setup as:
    1) Calculate Result As:  Summation of Rounded Values
    2) Calculate Single Value as: Normalization of result
    3) Calculate Along the Rows
    The KFs
    - Delta base vs. comp. units %
    - Delta base vs. comp. units
    are FORMULAS setup to:
    1) Calculate Result As:  Nothing defined
    2) Calculate Single Value as: Nothing defined
    3) Calculate Along the Rows: user default direction (grayed out)
    Thanks for the time taken to read in detail all of this. Long text but necessary to understand what the problem is.
    Any help is highly appreciated.
    Thank you.
    Mario

    Hi,
    The subraction will be carried out before doing the normalization of your KF's. So, it is displaying "number of units". Create a calculated keyfigure and subtract the KF's and in the properties of this calculated keyfigure, change the enhancement as "After Aggregation".
    I hope this will solve your issue.
    Regards,
    S P.

Maybe you are looking for