locked
Leaf and Objective KPIs Questions RRS feed

  • Question

  • Hi -

     

    We are in the beginning stages of what will be a PPS roll-out in a couple months and I'm creating some prototype KPIs for a 'discovery phase' effort. Having done some reading on the subject of normalized scores, I thought that I would start my prototyping by choosing a volume metric, and a couple ratio metrics to evaluate the integrity of the roll-up value in the objective KPI.

     

    For the purposes of this thread and my current issue which necessitates this post, I'm running into an apparent roadblock on the precision available for setting the KPI thresholds. Maybe I just don't know the application well enough yet, hence throwing these out on the table.

     

    Let's say that I have an objective KPI called 'Cash and Sales', along with two ratio leaf KPIs and another volume leaf KPI (will not detail this one here)... The two ratio leaf KPIs are called, 'Sales Chg' and 'Cash Shortage' respectively.

     

    Sales Chg has specific thresholds to drive the status indicator as follows: any sales growth value which is below the prior year is Red, growth values between zero and four percent are yellow, and values above four percent are green.

     

    Cash Shortage has the following thresholds which are very small numbers (causes my apparent precision issue). Cash shortages larger than 0.07% (errr, smaller numerically speaking that is...) need to receive a Red status, shortages between 0.05% and 0.07% are yellow, and shortages smaller than 0.05% (here we go again.... err, the number is actually larger on the number scale) will be green.

     

    Since I have specific targets, I set up the Sales Chg KPI to use a banding method based on the actual value. It seems like the same methodology will apply to the Cash Shortage KPI as well, but the threshold precision in the UI can only be set to two decimal places and I need four (please set me straight on this if I'm wrong ...).

     

    Here's a couple of questions: (1) Will I be able to roll-up these leaf KPIs into an Objective KPI if I use this banding method choice for the leafs? I understand the non-additive nature of ratio values such as these. (2) .... this is kinda related to #1 by worded a bit differently .... While I've read in post after post regarding the use of the Normalized banding method as the only banding method that should be used for Objective KPIs, does this mean that leafs must use this banding method also if they are going to be rolled up? (3) Does the Stated Score banding method have any application here to resolve my precision issue? And if so, can anyone point me to HOW to do that..... I've read the book, Monitoring and Analyzing with MSFT Office PPS 2007, but it did not go into great enough detail to explain the mechanics of setting up a Stated Score data source (can this be a table of values in an Excel worksheet?, or does these values have to exist in an AS source?).

     

    I've tested my Sales Chg KPI with both banding methods, Actual and Normalized. As expected, I get the correct status alerts when I compare my Actuals against the thresholds, but do not get the right status when I use the Normalized metholdology.

     

    Any help would be greatly appreciated as I'm spinning my wheels here. I think this has the potential to be a really great BPM tool, but obviously I need to understand how it works better to be really effective building our a DSS for our company.

     

    Thanks in advance for any and all help !

     

    Mel

    Wednesday, May 21, 2008 10:50 PM

Answers

All replies

  • Thank you for your prompt reply Alyson....

     

    I've had an oppty to read through your links and those documents cleared up some of my questions, but I feel that my Q related to threshold precision remains unaswered.

     

    Let me repost the relevant part of my initial entry here, with one small change by adding the minus sign in front of the numbers to reflect that they are negative amounts (I forgot to do that initially)....

     

    Cash Shortage has the following thresholds which are very small numbers (causes my apparent precision issue). Cash shortages larger than -0.07% (errr, smaller numerically speaking that is...) need to receive a Red status, shortages between -0.05% and -0.07% are yellow, and shortages smaller than -0.05% (here we go again.... err, the number is actually larger on the number scale) will be green.

     

    In the UI, I can enter the target value to the level of precision I need, e.g. -0.0005, but when I attempt to enter the same level of precision for the thresholds as follows, < -0.0007 = Red, between -0.0005 and -0.0007 = Yellow, and > -0.0005 = Green, that level of precision is not available. Is there way to accomplish what I need for threshold precision w/o changing the way the measure is calc'd in the cube?

     

    Mel

    Thursday, May 22, 2008 5:45 PM
  • The current precision in the thresholds only goes to 2 decimal places.  I will note your request to increase the precision as a feature suggestion for a future release.  In the meantime, I think your only options are to either adjust the value in the cube itself or write an MDX expression to multiply the value by 100 and then band by that.  I am assuming you are trying to band by actual is that correct?

     

    Thanks,

    Alyson

     

    Thursday, May 22, 2008 6:32 PM
  • Yes that is correct .... I am attempting to band by the actual value. Can walk me through a procedure to do as you suggest by using an MDX expression? If PPS was ProClarity BLS, I would know how and where to do this, but in PPS I could you some guidance...

     

    Mel

     

    Thursday, May 22, 2008 8:12 PM
  • You would need to enter an MDX expression for the value of your KPI.  Something like:

     

    [Measures].[measurename])*100

     

    This would take your measure and multiply it by 100 so that then your number would be larger in order to band by it.  You could band by numeric value of actual in this instance.

    If you don't want your KPI to show the number multiplied by 100 you would need to enter a target and band by stated score, in your stated score calc you would do a comparison of the metric in your stated score calc something like:

     

    iif(  [Measures].[measurename])*100 > .5, 0, 1)

     

    Does that help?

    Thanks,

    Alyson

     

    Tuesday, May 27, 2008 4:24 PM
  • My question is more specific to where in PPS's UI would I insert/type the MDX expression? Reviewing the options that are available in the KPI element screen (data mappings, calculation etc) and through the Thresholds settings dialog process ('Set scoring pattern and Indicator'), its not apparent to me where I can refer to/supply an MDX override expression.

     

    Can you clarify this for me?

     

    Thanks, Mel

     

    Wednesday, May 28, 2008 11:08 PM
  • When creating a KPI you would enter the MDX in the Data Mapping in the "Use MDX tuple formula" box.  This will allow you to create either the actual or the target via an MDX expression.

     

    If you are going to band by numeric value of actual you would enter the MDX expression in the actual value in the data mappings area.  Then no further changes are needed just enter your threshholds.

     

    If you want to band by stated score, you have to create your actual and your target using the data mappings column and entering hte MDX in the "use MDX tuple formula" box. You also then need to enter the same expression for the target in the threshhold settings in the "specify data mappings" area.  This allows you to enter an expression that will be the band by expression - then enter your thresholds normally after that.

     

     

    Hope this helps,

    Alyson

    Thursday, May 29, 2008 3:11 PM