Focal Point Banner
Community Center Education Summit Technical Support User Groups
Let's Get Social!

Facebook Twitter LinkedIn YouTube
Focal Point    Focal Point Forums  Hop To Forum Categories  Performance Management Framework (PMF)    Definition of "Basis Value"
Go
New
Search
Notify
Tools
Reply
  
Definition of "Basis Value"
 Login/Join
 
Platinum Member
posted
What is "Basis Value", and what page in the documentation is it covered? 3 of us have looked and couldn't find it.

Thank you.


WebFOCUS 7.6.11, Linux, HTML, PDF, AHTML, EXCEL
 
Posts: 100 | Registered: September 15, 2010Reply With QuoteReport This Post
Master
posted Hide Post
Basis is the standard on which 100% is based. It prevents divide-by-zero in situations where you might have a target of 0, because in descending measures your target is the denominator in the calculation.

If you set a basis of 10 on a descending measure, and the target is 0, and your actual is 10, then you are at 0% achieved. If the basis were 20, you'd be at 50% achieved.

Not sure why this isn't in the doc but we'll correct that.

Also please see http://forums.informationbuild...=363101451#363101451

and

http://forums.informationbuild...=369104021#369104021

There is a search function on the Forum. That's how I found these.

thanks


Bob Jude Ferrante
Director of Business and Development
WebFOCUS Performance Management
Bob_Ferrante@ibi.com
917-339-5105

I'll take any questions about PMF - business or technical - anytime!

 
Posts: 919 | Registered: March 26, 2003Reply With QuoteReport This Post
Platinum Member
posted Hide Post
Thanks Bob.

We still were a little unsure, so for clarification, are either of these interpretetations correct?

1. In the situation where a descending measure's target is zero, the basis value is used instead of the target value to avoid division by zero. In this case is the formula to calculate Percent of Target Reached (% Achieved): (Basis - Actual) / Basis? This would would work for the first 2 examples below.

  
Basis    Actual    Target    % Achieved
-----    ------    ------    ----------
   10        10         0            0%
   10        20         0           50%
   20        10         0          -50%



How would this work if the basis were 20 and the actual were 10? Using (Basis - Actual) / Basis, (10 - 20) / 20 = -0.5, which is -50%.

2. The other interpretation we came up with was:
(Actual - Basis) / (Basis - Target)
which would be used regardless of whether the target is 0 or not for a descending measure.

We also need to be able to explain all this to our end users who supply the info we use to create the measures.

Thank you.


WebFOCUS 7.6.11, Linux, HTML, PDF, AHTML, EXCEL
 
Posts: 100 | Registered: September 15, 2010Reply With QuoteReport This Post
Platinum Member
posted Hide Post
As Bob said, Basis is used for descending Measures when the Target is 0 to avoid divide by 0 situations.

The formula for % Reached is ((BASIS- ACTUAL) / (BASIS - TARGET)) * 100.0. So expanding on your data:

 
Basis    Actual    Target      % Reached
-----    ------    ------     ----------
   10        20         0          -100%
   10        15         0           -50%
   10        10         0             0%
   10         5         0            50%
   10         0         0           100%
 



You can also use Basis when your Target is very small to avoid excessively large values of % Reached.

EricH
 
Posts: 164 | Registered: March 26, 2003Reply With QuoteReport This Post
Platinum Member
posted Hide Post
Thank you, Eric.


WebFOCUS 7.6.11, Linux, HTML, PDF, AHTML, EXCEL
 
Posts: 100 | Registered: September 15, 2010Reply With QuoteReport This Post
<cae>
posted
This is the first time using Focal Point so I hope I am doing this properly.
I have been trying to figure out the best way to calculate % target reached for descending measures. I understand that using Basis value avoids the Dividing by 0 issue but what if the target equals the Basis value? Aren't we back to the dividing by 0 issue. Also, I have found that if your target and actual numbers are greater than the basis value the Basis value calculation gives incorrect results. If there were a way to ensure that there were never targets of 0, wouldn't it always be best to use the Target Value calculation for % target reached for descending measures?
Any thoughts?
 
Reply With QuoteReport This Post
Master
posted Hide Post
For some reason I cannot fathom, a complete description of Basis isn't in the doc...

PMF calculates Percent Reached for descending Meausres by dividing the Target into
the Actual value and multiplying the result by 100. If the descending Measure had a
Target of 0, this would have resulted in a Divide by Zero, which would have been
mathematically invalid. Special handling is therefore required for Measure Targets that
are set to zero.

In cases where you had a target of 1 or a fractional target, and an Actual that was many
multiples higher (e.g., a Target of .05 and an Actual of 25), you would see a seemingly
abnormally high Percent Reached (sometimes so high as to overflow and display *******
as the Percent Reached).

Basis controls where you would have zero Percent Reached. The basis is the point you
would exceed with your Descending Measure greater than which your Percent Reached
would start moving into the negative. Any Actuals that were greater would show a
negative Percent Reached; any that were less will show a positive Percent Reached.
Setting the Basis gives you control over how Percent Reached is calculated so you can
see more reasonable-looking Percents Reached.

How Descending Measure Basis Handling Works

The special handling lets you pick a defined numeric value as your zero crossing for
Percent Reached. The value can be sufficiently large that you would never exceed it, or
you can adjust it as needed until your numbers look reasonable.

If you set a Basis for a Descending Measure, the calculation for Percent Reached
subtracts the Actual from the Basis amount and the Target for the Basis amount, which
rebalances the Percent Reached to give you more control when Target numbers get very
small.

-----
What this means: If the target for your descending measure SUMS out to equal the basis exactly (and note that if your measure uses continuous rather than discrete values the odds of this are QUITE LOW), you would in concept hit a divide by zero but you never do, since we defend against divide by zero deterministically by setting set Pct Reached to 0 rather than having "undetermined value" for an answer (and/or having WebFOCUS throw an error halting processing).

Hope this helps.


Bob Jude Ferrante
Director of Business and Development
WebFOCUS Performance Management
Bob_Ferrante@ibi.com
917-339-5105

I'll take any questions about PMF - business or technical - anytime!

 
Posts: 919 | Registered: March 26, 2003Reply With QuoteReport This Post
<cae>
posted
Thank you Bob.

I have now moved on to our "Range" measures. I was trying to figure out the formula for %target reached. I thought I had it:
(Target-ABS(Actual-Target))/Target. This works most of the time but still not for everything. Do you know the formula?
Thanks,
 
Reply With QuoteReport This Post
Master
posted Hide Post
Here's a tip: Have a look in A_COMPUTE_RPT_VARS.FEX and its children. All the answers to your questions are there. :-)


Bob Jude Ferrante
Director of Business and Development
WebFOCUS Performance Management
Bob_Ferrante@ibi.com
917-339-5105

I'll take any questions about PMF - business or technical - anytime!

 
Posts: 919 | Registered: March 26, 2003Reply With QuoteReport This Post
<cae>
posted
Hi Bob,
I guess I need to clarify that I am a Business Analyst, not a developer. I do not have the capability to look at system files or code. If there is a manual I could look there. Otherwise I need to try to get help from other people.
Thanks
 
Reply With QuoteReport This Post
Master
posted Hide Post
I'd love to... but am in the middle of getting PMF 5.3.1 ready to ship. Anyone else want to help her right away?

Otherwise, I'll get back to you next week.


Bob Jude Ferrante
Director of Business and Development
WebFOCUS Performance Management
Bob_Ferrante@ibi.com
917-339-5105

I'll take any questions about PMF - business or technical - anytime!

 
Posts: 919 | Registered: March 26, 2003Reply With QuoteReport This Post
Member
posted Hide Post
I am working on Linux.It has some problem over documents configurations.It is terminated by entering a function command.
 
Posts: 1 | Registered: April 16, 2012Reply With QuoteReport This Post
Master
posted Hide Post
Basically gross percent reached is:
100.0 - ABS(SUM OF(Target_Value) - SUM OF(Actual_Value))/SUM OF(Target_Value) 


In the case of percents and rations you have to take the division into numerators and denominators into account

100.0 - ABS(SUM OF(Target_Numerator_Value)/SUM OF(Target_Denominator_Value)) - (SUM OF(Actual_Numerator_Value)/SUM OF(Actual_Denominator_Value))/(SUM OF(Target_Numerator_Value)/SUM OF(Target_Denominator_Value)) 


If Target ever SUMs to a zero amount we set it to 0.

This doesn't show all the breakpoints we calculate to draw the colours (we do those based on applying thresholds you set), but all use the same pattern.

Hope this helps. Note that it's pretty tedious to type this stuff into this primitive online editor so I hope I didn't make any typing errors. But this should give you an idea.


Bob Jude Ferrante
Director of Business and Development
WebFOCUS Performance Management
Bob_Ferrante@ibi.com
917-339-5105

I'll take any questions about PMF - business or technical - anytime!

 
Posts: 919 | Registered: March 26, 2003Reply With QuoteReport This Post
<cae>
posted
Bob,
Thank-you, I think.
This will be tough to test for percents and ratios. A little challenge to keep me busy.
 
Reply With QuoteReport This Post
  Powered by Social Strata  
 

Focal Point    Focal Point Forums  Hop To Forum Categories  Performance Management Framework (PMF)    Definition of "Basis Value"

Copyright © 1996-2018 Information Builders, leaders in enterprise business intelligence.