Project Performance Indicator

[mike edwards]

I am no genius at creating formulas so I need some help. I am trying to
create a custom enterprise field called Project Performance (graphical)
indicator that will show performance based on CPI and SPI. I thought that if
you added the CPI and SPI together a sum of 2.0 would indicate an on budget
and on schedule project (green). Then anything between a 1% and 10% variance
would equal yellow and anything beyond 10% would equal a red condition. The
problem is that tasks that haven't started yet are showing up red as well.
Any recommendations? Is there someplace that I can get some sample formulas
that would allow me to show project performance?

 

[John]

I am no genius at creating formulas so I need some help. I am trying to
create a custom enterprise field called Project Performance (graphical)
indicator that will show performance based on CPI and SPI. I thought that if
you added the CPI and SPI together a sum of 2.0 would indicate an on budget
and on schedule project (green). Then anything between a 1% and 10% variance
would equal yellow and anything beyond 10% would equal a red condition. The
problem is that tasks that haven't started yet are showing up red as well.
Any recommendations? Is there someplace that I can get some sample formulas
that would allow me to show project performance?

Mike,
You know this sounds like something a CEO might ask for. Reduce
everything down to a single metric. Unfortunately in every case I've
seen the single metric is so far removed from providing any valid
information that it is totally useless. For example, what does a
combined CPI/SPI value of 2.0 tell you? Nothing. You could have a plan
whose CPI is 2.0 and SPI is 0 (or vice versa) and everything looks hunky
dory. Don't fall into that trap.

As it is, the CPI and SPI metrics already attempt to summarize
performance. CPI is a good measure, it measures overall cost performance
in terms of cost. On the other hand, SPI is not a good measure of
schedule because it measures a time related metric in terms of money. If
you believe that time is always money, then SPI probably works. I
personally don't buy into that approach - I think schedule performance
should be based on a time related metric, such as Total Slack. However,
that's just my opinion.

Well, that's my two cents. I'm pretty sure others will have differing
opinions

John
Project MVP

 

[Trevor Rabey]

Are you not trying to go to a lot of trouble to invent something which is:
time consuming
possibly counter-productive
maybe pointless and meaningless and uninformative
already invented better and already built in to MSP provided you set it up
properly in the first place

You need to define what "performance" actually is.
You need to define what "success" is.

Since the "status" of a project at any given point in time during execution
is likely to be a mixture of things which are "over/under/on schedule", and
"over/under/on budget" it is pointless trying to boil down the whole picture
to one number or colour indicator. How serious the status any of the parts
of the project are, and whether they are cause for joy or concern or
correction, and what form the correction should take, are matters for
assessment of the specific circumstances which include which Tasks are
involved and what point of time is being considered. This is the essence of
analysis.
For example, non-critical Tasks which are starting and/or finishing "late"
(ie later than scheduled) but well within their float, are nothing much to
be comcerned about.
In this game, even a simple word such as "late" requires qualification.

CPI and SPI come close to being valid, meaningful indicators by themselves.
Your invention of adding them together, and your arbitrary 10% variance, are
creative but how do they help?

Trevor Rabey

 

[davegb]

Are you not trying to go to a lot of trouble to invent something which is:
time consuming
possibly counter-productive
maybe pointless and meaningless and uninformative
already invented better and already built in to MSP provided you set it up
properly in the first place

You need to define what "performance" actually is.
You need to define what "success" is.

Since the "status" of a project at any given point in time during execution
is likely to be a mixture of things which are "over/under/on schedule", and
"over/under/on budget" it is pointless trying to boil down the whole picture
to one number or colour indicator. How serious the status any of the parts
of the project are, and whether they are cause for joy or concern or
correction, and what form the correction should take, are matters for
assessment of the specific circumstances which include which Tasks are
involved and what point of time is being considered. This is the essence of
analysis.
For example, non-critical Tasks which are starting and/or finishing "late"
(ie later than scheduled) but well within their float, are nothing much to
be comcerned about.
In this game, even a simple word such as "late" requires qualification.

CPI and SPI come close to being valid, meaningful indicators by themselves.
Your invention of adding them together, and your arbitrary 10% variance, are
creative but how do they help?

Trevor Rabey

I agree with John and Trevor, that trying to reduce a project to a
single performance variable is taking quantitative analysis way too
far, for the same reasons they stated.
I do disagree with John that Total Slack is a metric. You certainly can
use it to help priortize tasks, and as a warning when tasks might be in
trouble, but I don't think it's very effectived beyond that.
As John mentioned, SV is normally in dollars. John believes this
renders it ineffective because it's not in time units. To me, it's not
intuitive, but still useful once you've used it enough. However, if you
want to take the trouble, you can move the appropriate info into Excel
and measure SV in time units by measuring it's horizontal, rather then
vertical, variation from the baseline. This method has been published
at least once that I know of in a book I have at home.
More importantly, none of these standard measures of performance
measures the most important thing on any project. Customer
satisfaction. I don't care how carefully you measure everything else in
the project, or if you bring it in on schedule and within budget, if
the customer (internal or external) isn't satisfied with the result,
you've failed. And your CPI + SPI could be 277!
There is a rush these days to try to reduce everything to numbers, and
then beyond that, to a single number. And then focus on achieving some
numeric goal. This invariably leads to carefully metered failure. It's
your choice whether you want to follow this path, or try to do projects
successfully. Metric success is not project success!
Hope this helps in your world.

 

[Ranjana Mehta]

Hi:
The only measure I know that includes CPI and SPI is CPIxSPI.
Thanks
Ranjana