Good folks,

It's been a while (unfortunately) and I think I know the answer I just want some confirmation.

Within the 5.2.x and below versions of PMF, there is no way to weight the score/value/count of one specific dimension VLAUE over another.
Say Region=East is to be given more weight Region=West when rolling up a measure to Division.

I don't think this can be done.

Anyone have experience with this?

Thanks

That feature's been in PMF since 4.5 and - we think - unused since then (certainly nobody has mentioned it). In 4.5 you had to turn on a setting to show it; now we just always leave that switch on, hoping someone decides to use it.

What are you trying to do with it?

=b

Thanks, Bob, for your reply.

Initially I was trying to answer a question posed to me.

Now I'm curious how this is implemented?

Can you direct me on the setting up this feature?

Thanks.

Around page 294 of the manual (depending on what version you're running)

"Setting Dimensional Measure-to-Objective Weighting"

Bob

Thanks again, Bob, for your response.

I reviewed the passages on Measure-to-Oblective Weighting.

I think my question still needs improvement.

What I'm trying to understand is:
Can PMF weight a specific value in a measure? eg. Can Region=Boston carry more weight (75%) than Region=Chicago (25%) when rolling up a measure to Division?

Thanks.
Greg

OK let me ask a question then...

Suppose you have a dimension with country and region levels and two regions within the country USA are Boston and Chicago, and we're looking at a Metric called Sales.

If Boston region for the current period says Sales=$100 and Chicago likewise says Sales=$100, are you saying you want the total dollars for Sales at the USA level to add up to $200? Or should they add up to something else? Of course this question is asked because people do want things to add up.

Now, with Dimensional Weighting, you do have the ability to change the priority of the metric contribution at the next level up - the goal/risk level. You can have different Dimension level values separately weighting to the higher-level stuff in the PMF model.

Also, with Dimensional Tolerances you can set the H-M-L tolerances differently, so that although the values don't change, the expected tolerance to lower performing values (which depend on the directionality of the metric) can vary.

Hope this logic path makes sense... now please give us examples of what you expect.

thanks

The idea was that if Region Boston had $100 and Region Chicago had $100 in Sales, then when these are rolled up with a weight on specific regions where Boston's weight is 75% and Chicago is 25% that the Boston value would be multiplied by 125% and Chicago by 75% to get to a total of $200.

As I write this I don't see how this can be done. Am I wrong?

Thanks.

You are right. Plus I'm not sure how you'd possibly show what you're trying to show in any sense in any system... Once you aggregate the data on a report, the data is already summed, yea? You still want it to be $200, at that level you're looking at country, not region any more. And when you drill down you still want to see both regions with $100, yea?

If that's not the case, and you want to scale the data, you can do it before the load. Apply a scaling factor to the data inbound. Or if you want to weight it, use the weighting features at the objective/risk level.

If you need something else, we got to figure out what that is...

thanks

This is as I thought. Thanks, Bob.
You've answered my curiosity. If/when I get more interest I may have some follow-up. For now this is what I wanted to understand.

PMF continues to rock and so does the support!

We're blushing. All of us. :-)

Ok, so here's an update to this challenge.

I have an Objective that has weighted measures that roll-up to it. The measures are not eually weighted.

When I run a Analysis Designer report using Objective Across Dimension I can only choose a PCTRCH field, no values. These measures are Ratios.

My client wants to see the Actual, Target and Stretch values across month at the Obj level using the weighting at the measure level when rolling them up.

Makes sense?
Any one have a suggestion on how to get there?

I'm thinking of pulling the Act, Trgt and Strch and weighting and doing the calc myself but I'm hoping there is better 'PMF' way that is available out-of-the-box.

Let me know what you think.

IS.

At the Objective level, many measures can be combined.

Does $1,000,000.00 + 323 production line defects + 17 jobsite accidents = 1,000,340.00 dollar.defect.accidents?

No, right?

So we abstract them by their performance. Hence the Percent Reached is the only thing that can be aggregated, and that is all that is aggregated.

However, you can sort a Measure report BY objective in analysis designer.

You can get dimensionally aggregated metrics broken out in Scorecard order. You can sort across objectives, or across metrics, or across dimensions. One of these is probably what they want.

When you show it to them they'll like it.