Modelling Worker Efficiency (Productivity)

[Mark E]

Hi.

I have a problem which I fear may be out of Project's capability to solve
and wondered if anyone out there may know different?

If I define a task as being of a specific duration (effort) and want to
model the effects on project duration by adding worker resources. If we say
a skilled worker has an efficiency of 1 and a trainee worker has an
efficiency of 0.75, this would mean that the trainee has the capacity to
produce 75% of the work of a skilled worker in the same time.

I would like to model the effect of varying the ratio of skilled/trainee
resources employed on the task but so far have been unable to find a way of
modelling their relative productivity. I would have thought that this would
be a powerful feature to have within the application and would be mildly
surprised if indeed it is not possible.

Many thanks in anticipation of your kind comments.

 

[Jim Aksel]

Project will not do this without help. However, there may be some companion
products that are available that might be of assistance.

Try here: http://project.mvps.org/comprods.htm

I am thinking that a product called Risk Plus would help as it models
changes in duration & work using curves that can be specified. I think
something like a Monte Carlo simulation would be in order and Risk Plus does
that.

As a thought, you can try this in Project. Create a custom column where you
enter worker efficiency, a number near 1.0. In another custom column create
work times efficiency. Copy your [work] column to a spare column such as
Text1 which will memorialize your orginal work data (or just save a baseline).

Now, You have a column of work x efficiency. Paste the contents of that
column into work, overwriting what is there. The calculated numbers will
change but so will your work column. Now you can see the impact on Cost.
However this will not show you the impact on schedule. So, change the worker
availability to some other number like 75% which will push out your schedule.
Clunky, but it will give you some insight.

If I think of anything else, I'll post again.

 

[Mark E]

Thanks for the reply Jim.

I've already tried to solve using worker availability, but this introduces
two new problems.

One - if the task duration is less than the multiple of availability x
day-length, then the worker in question is fully committed to the task and no
account is taken of their relative inefficiency.

Two - if the multiple of availabilty x day-length is less than the task
duration, then the worker contributes to the task only a proportion equal to
his availability (which in this case models his efficiency), which is good.
However, when the cost of his work is calculated, the availability quotient
is applied to the calculation, making it appear that the cost is a fraction,
equal to the availabilty percentage, of that which would actually occur!

Your idea of adding custom columns is an interesting one, but as you point
out, it does not help in modelling impact on scheduling which is really what
I'm trying to do. So far as I can see, the companion products you point to
do not seem to offer the functionality I'm after (unless I'm missing
something!).

In your experience, do you think it would be possible to over-ride the
effort calculation in VBA, so as to take account of efficiency and if so
would you have any idea which functions/routines would be the ones to look at?

Again, thanks in anticipation.
Mark

Project will not do this without help. However, there may be some companion
products that are available that might be of assistance.

Try here: http://project.mvps.org/comprods.htm

I am thinking that a product called Risk Plus would help as it models
changes in duration & work using curves that can be specified. I think
something like a Monte Carlo simulation would be in order and Risk Plus does
that.

As a thought, you can try this in Project. Create a custom column where you
enter worker efficiency, a number near 1.0. In another custom column create
work times efficiency. Copy your [work] column to a spare column such as
Text1 which will memorialize your orginal work data (or just save a baseline).

Now, You have a column of work x efficiency. Paste the contents of that
column into work, overwriting what is there. The calculated numbers will
change but so will your work column. Now you can see the impact on Cost.
However this will not show you the impact on schedule. So, change the worker
availability to some other number like 75% which will push out your schedule.
Clunky, but it will give you some insight.

If I think of anything else, I'll post again.

Hi.

I have a problem which I fear may be out of Project's capability to solve
and wondered if anyone out there may know different?

If I define a task as being of a specific duration (effort) and want to
model the effects on project duration by adding worker resources. If we say
a skilled worker has an efficiency of 1 and a trainee worker has an
efficiency of 0.75, this would mean that the trainee has the capacity to
produce 75% of the work of a skilled worker in the same time.

I would like to model the effect of varying the ratio of skilled/trainee
resources employed on the task but so far have been unable to find a way of
modelling their relative productivity. I would have thought that this would
be a powerful feature to have within the application and would be mildly
surprised if indeed it is not possible.

Many thanks in anticipation of your kind comments.

 

[Trevor Rabey]

That MSP doesn't do this is not a lack of capability but just the way it is
designed, and this goes right back to the Critical Path Method.
In its raw form, CPM does not even involve resources.
The only real reason that resources are in MSP is to allow a calculation of
the costs of the tasks, oh an do leveling of course.
The tasks type, and effort driven and W=ExD calculations are really just a
bit of optional auto calculations that kicks in after your initial resource
assignments.
I think you are missing something, which is that project management is
primarily about the project and the tasks within it, not primarily about the
resources.
It's not an HR and Training application, nor is it an accounting system.
It also doesn't have a database of "production rates" for different types of
work, but there are plenty of books and software apps that do.
How you arrive at estimates for durations of tasks is outside MSP's
knowledge.
When you estimate and plan that it takes one guy to lay 10000 bricks in 10
days, it is implicit that it's possible and you can get someone who is
"efficient" enough to do that.

Projects don't usually succeed or fail because of relative "efficiencies" of
the human resources, but rather for a myriad of other reasons, including
insufficient scope definition and, in the case of the example here, by the
kind of coordination failure which ensures that the bricks and/or bricklayer
don't arrive when needed.

Carefully rating your human resources and using that information to estimate
the durations that you put into MSP might be a useful thing to do.
Anyway, Frederick Taylor already did it, but without the software.

Thanks for the reply Jim.

I've already tried to solve using worker availability, but this introduces
two new problems.

One - if the task duration is less than the multiple of availability x
day-length, then the worker in question is fully committed to the task and
no
account is taken of their relative inefficiency.

Two - if the multiple of availabilty x day-length is less than the task
duration, then the worker contributes to the task only a proportion equal
to
his availability (which in this case models his efficiency), which is
good.
However, when the cost of his work is calculated, the availability
quotient
is applied to the calculation, making it appear that the cost is a
fraction,
equal to the availabilty percentage, of that which would actually occur!

Your idea of adding custom columns is an interesting one, but as you point
out, it does not help in modelling impact on scheduling which is really
what
I'm trying to do. So far as I can see, the companion products you point
to
do not seem to offer the functionality I'm after (unless I'm missing
something!).

In your experience, do you think it would be possible to over-ride the
effort calculation in VBA, so as to take account of efficiency and if so
would you have any idea which functions/routines would be the ones to look
at?

Again, thanks in anticipation.
Mark

Project will not do this without help. However, there may be some
companion
products that are available that might be of assistance.

Try here: http://project.mvps.org/comprods.htm

I am thinking that a product called Risk Plus would help as it models
changes in duration & work using curves that can be specified. I think
something like a Monte Carlo simulation would be in order and Risk Plus
does
that.

As a thought, you can try this in Project. Create a custom column where
you
enter worker efficiency, a number near 1.0. In another custom column
create
work times efficiency. Copy your [work] column to a spare column such as
Text1 which will memorialize your orginal work data (or just save a
baseline).

Now, You have a column of work x efficiency. Paste the contents of that
column into work, overwriting what is there. The calculated numbers will
change but so will your work column. Now you can see the impact on Cost.
However this will not show you the impact on schedule. So, change the
worker
availability to some other number like 75% which will push out your
schedule.
Clunky, but it will give you some insight.

If I think of anything else, I'll post again.

Hi.

I have a problem which I fear may be out of Project's capability to
solve
and wondered if anyone out there may know different?

If I define a task as being of a specific duration (effort) and want to
model the effects on project duration by adding worker resources. If
we say
a skilled worker has an efficiency of 1 and a trainee worker has an
efficiency of 0.75, this would mean that the trainee has the capacity
to
produce 75% of the work of a skilled worker in the same time.

I would like to model the effect of varying the ratio of
skilled/trainee
resources employed on the task but so far have been unable to find a
way of
modelling their relative productivity. I would have thought that this
would
be a powerful feature to have within the application and would be
mildly
surprised if indeed it is not possible.

Many thanks in anticipation of your kind comments.

 

[Jim Aksel]

If you use Risk+ you can setup maximum and minimum durations (somewhat like
the triple time estimate in Project). Example: Project Task Durstion 5
days. Risk+ will let you set the minimum duration and maximum duration to
any value you wish. Most people wil use +/-10% or +/-25% or similar.

In this case (10%) Risk plus will randomly select a duration for that task
anywhere in the range of 4.5 days to 5.5 days and then calculate your costs
from there. To similate an experienced worker you could say 5 days +/-10%
and for an inexperienced worker you could specificy 5 days -0% +25% which
would give you a range from 5 to 6.25 days. Pick any number you want, even
400% if you like.

Risk+ will perform a Monte Carlo Simulation for you and give you the time
and cost results. It will run the simulation for you as many timesas you
want, for example 3,000 and plot the results on a graph. I can't imagine you
can get better than that. It doesn't take too long, maybe an hour (depending

Thanks for the reply Jim.

I've already tried to solve using worker availability, but this introduces
two new problems.

One - if the task duration is less than the multiple of availability x
day-length, then the worker in question is fully committed to the task and no
account is taken of their relative inefficiency.

Two - if the multiple of availabilty x day-length is less than the task
duration, then the worker contributes to the task only a proportion equal to
his availability (which in this case models his efficiency), which is good.
However, when the cost of his work is calculated, the availability quotient
is applied to the calculation, making it appear that the cost is a fraction,
equal to the availabilty percentage, of that which would actually occur!

Your idea of adding custom columns is an interesting one, but as you point
out, it does not help in modelling impact on scheduling which is really what
I'm trying to do. So far as I can see, the companion products you point to
do not seem to offer the functionality I'm after (unless I'm missing
something!).

In your experience, do you think it would be possible to over-ride the
effort calculation in VBA, so as to take account of efficiency and if so
would you have any idea which functions/routines would be the ones to look at?

Again, thanks in anticipation.
Mark

Project will not do this without help. However, there may be some companion
products that are available that might be of assistance.

Try here: http://project.mvps.org/comprods.htm

I am thinking that a product called Risk Plus would help as it models
changes in duration & work using curves that can be specified. I think
something like a Monte Carlo simulation would be in order and Risk Plus does
that.

As a thought, you can try this in Project. Create a custom column where you
enter worker efficiency, a number near 1.0. In another custom column create
work times efficiency. Copy your [work] column to a spare column such as
Text1 which will memorialize your orginal work data (or just save a baseline).

Now, You have a column of work x efficiency. Paste the contents of that
column into work, overwriting what is there. The calculated numbers will
change but so will your work column. Now you can see the impact on Cost.
However this will not show you the impact on schedule. So, change the worker
availability to some other number like 75% which will push out your schedule.
Clunky, but it will give you some insight.

If I think of anything else, I'll post again.

Hi.

I have a problem which I fear may be out of Project's capability to solve
and wondered if anyone out there may know different?

If I define a task as being of a specific duration (effort) and want to
model the effects on project duration by adding worker resources. If we say
a skilled worker has an efficiency of 1 and a trainee worker has an
efficiency of 0.75, this would mean that the trainee has the capacity to
produce 75% of the work of a skilled worker in the same time.

I would like to model the effect of varying the ratio of skilled/trainee
resources employed on the task but so far have been unable to find a way of
modelling their relative productivity. I would have thought that this would
be a powerful feature to have within the application and would be mildly
surprised if indeed it is not possible.

Many thanks in anticipation of your kind comments.