apply SPI CPI to measure Resource

[Edward.hd.wang]

Hi all,

Do you have any experience or think it is workable to use SPI & CPI to
measure the performance of each resource. I mean, in MSP, we usually
use EV to meaure the tasks, but I found it also support use BCWP, BCWS,
ACWP for resource, and we also could customerize the some field to
calculat the SPI & CPI for resource.

By delopying the EV, PM and high levels could get the benefits from it,
but I still belive this could bring some value to the resource managers.
By producing the report of SPI & CPI of each resource, the resource
manager could get more direct data about his members.

 

[davegb]

Hi all,

Do you have any experience or think it is workable to use SPI & CPI to
measure the performance of each resource. I mean, in MSP, we usually
use EV to meaure the tasks, but I found it also support use BCWP, BCWS,
ACWP for resource, and we also could customerize the some field to
calculat the SPI & CPI for resource.

By delopying the EV, PM and high levels could get the benefits from it,
but I still belive this could bring some value to the resource managers.
By producing the report of SPI & CPI of each resource, the resource
manager could get more direct data about his members.


--
Edward.hd.wang
------------------------------------------------------------------------
Edward.hd.wang's Profile: http://forums.techarena.in/member.php?userid=19009
View this thread: http://forums.techarena.in/showthread.php?t=630908

http://www.techarena.in

It seems that this would only work if each task had only one resource
assigned to it, which is doable on small, less complex projects, but
not on others. This is just one limitation I can think of up front.

I've been doing this a long time (almost 30 years), and I've never
heard of using EV attributes on resources. Would be interesting to have
a discussion about the relative merits of doing so. I do know a bit
about Measures of Performance (MOP). Basically, a person's performance
metric should always be measured in a way that promotes getting their
job done effectively. For instance, paying a CEO a bonus based on stock
options merely rewards them for inflating the stock price, then cashing
out. Not that something like that would ever happen for real! Paying
a brick layer per brick placed encourages haste and sloppy work. Some
sort of quality factor needs to be included to make it work.

In class, I use the example of call centers. What would be a good
metric for the resources taking support calls from your customers? In
some call centers, the people are rated on how many calls they handle
in an hour. This encourages them to get each customer off the line as
fast as possible, and leads to poor service quality. Other call centers
rate personell on their rate of return calls on the same ticket. This
encourages the support person to take care of the customer's needs on
the first call, even though doing so will increase the time, and
therefor the cost, of the individual call. But it makes for good
customer service, if that's your goal.

So on this basis, what would be the ramifications of measuring resource
performance on CPI and or SPI? I'm going to think about that a bit. Any
ideas?

 

[Dave]

It seems that this would only work if each task had only one resource
assigned to it, which is doable on small, less complex projects, but
not on others. This is just one limitation I can think of up front.

I've been doing this a long time (almost 30 years), and I've never
heard of using EV attributes on resources. Would be interesting to have
a discussion about the relative merits of doing so. I do know a bit
about Measures of Performance (MOP). Basically, a person's performance
metric should always be measured in a way that promotes getting their
job done effectively. For instance, paying a CEO a bonus based on stock
options merely rewards them for inflating the stock price, then cashing
out. Not that something like that would ever happen for real! Paying
a brick layer per brick placed encourages haste and sloppy work. Some
sort of quality factor needs to be included to make it work.

In class, I use the example of call centers. What would be a good
metric for the resources taking support calls from your customers? In
some call centers, the people are rated on how many calls they handle
in an hour. This encourages them to get each customer off the line as
fast as possible, and leads to poor service quality. Other call centers
rate personell on their rate of return calls on the same ticket. This
encourages the support person to take care of the customer's needs on
the first call, even though doing so will increase the time, and
therefor the cost, of the individual call. But it makes for good
customer service, if that's your goal.

So on this basis, what would be the ramifications of measuring resource
performance on CPI and or SPI? I'm going to think about that a bit. Any
ideas?

I can't really see how it can work. The suggested approach seems to
involve measuring resource performance against a task baseline into
which the resource may have had little or no input. If the baseline for
a given task/resource association is unrealistic, then they will deviate
from it. In the same plan, there may be a more realistic baseline for
another task (because these things happen in real life even with
competant project managers, estimators and planners) and so the resource
will not deviate significantly from it. Then what do the metrics
actually tell you? You certainly can't compare them across tasks with
any meangful results.

Dave

 

[davegb]

I can't really see how it can work. The suggested approach seems to
involve measuring resource performance against a task baseline into
which the resource may have had little or no input. If the baseline for
a given task/resource association is unrealistic, then they will deviate
from it. In the same plan, there may be a more realistic baseline for
another task (because these things happen in real life even with
competant project managers, estimators and planners) and so the resource
will not deviate significantly from it. Then what do the metrics
actually tell you? You certainly can't compare them across tasks with
any meangful results.

Dave

The limitation that the resource may or may not have had input to the
baseline is very real, though usually not neccessary. In many cases,
the PM can give them the opportunity to help determine the estimated
duration of the work. Certainly, in the real world, this is rarely
done. But it really works well, in most cases, when you do it.

In addtion to Dave's comment, I thought of a few other issues:

1. Some resources' tasks are primarily internal, and not dependent on
other's performance. Others' tasks are primarily external, and
therefore, dependent on others' performance. This method of evaluating
their perfomance puts the second group at a disavantage.

2. Certain types of work are more prone to delays and overruns than
others. I.e., Civil work is more likely to exceed time expectations
(and therefore, costs) than welding or carpentry. Would be an apples to
oranges comparison.

3. One of my problems with EV is that it doesn't incorporate any
quality considerations unless they are built-in to the project, as they
should be. But emphasizing cost and schedule performance without
emphasizing quality performance first invariably leads to poor quality.
Using CPI/SPI as a resource performance measure without quality somehow
worked in would surely do the same.

Just some food for thought, in case you're hungry!

Hope this helps in your world.

 

[Andrew Lavinsky]

I've used EVMS on the resource level, and it works pretty well....I don't
apply it to the individual resource, but rather group the resources by department,
then apply EVMS to the department overall.

The goal of EVMS is then to measure performance against goals. If they're
not making their goal, that's not necessarily a bad thing, just an indicator
that something is amiss (performance, plan, whatever). That just signals
the PM to pay attention. You can use this to identify things such as need
for additional training, or poor planning....and then potentially track performance
improvement or use it in the lessons learned.

You can also identify the EAC for each department, which is certainly a metric
of interest to the organization...

I would contend that applying EVMS to resources on the departmental level
is where you get the biggest bang for the buck...

-A