Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

MS Project increasing duration

Status
Not open for further replies.

wanstronian

Aerospace
Jun 23, 2007
17
0
0
GB
Hi,

I know this is an age-old problem/bug/feature of MS Project, but can someone please tell me why allocating resource to a task line in project INCREASES its duration unless the resource is added at 100% utilisation?

I have a task with effort of 54 days. I have two resources assigned, one at 68% and another at 26%. I believe, and my calculator agrees with me, that this is the equivalent of 94% utilisation. So why does MSProj tell me that the task will take 123 days?

I've fiddled with all the 'task type', 'effort-driven' options, and nothing makes any difference.

How do I make Project behave? Or, what am I misunderstanding and what do I have to do to get the result I'm expecting?

Thanks
 
Replies continue below

Recommended for you

I think I may have sussed this now - realised that the concept of effort is not a property of a task. Which means that Project can never support effort-driven scheduling [surprise].

Can anybody recommend an alternative tool that can provide this basic functionality?

Thanks
 
Depends on how the data is entered. Enter the duration only, followed by the allocated resources a[68%],b[26%], then MSP fills in the total work at 406.08 hrs.


TTFN

FAQ731-376
 
Exactly. Not effort-driven, duration-driven. Project can't handle effort-driven scheduling. I want to know if there's a tool out there that can.

I want to know how long a given task will take, if I know how much effort's involved and I know what resources I have.

As an example, I know, from various estimating techniques, that developing Widget X takes 200 person-days of effort. I know that I have lots of other tasks to do and I know that I have a team of, say, 6. What I need to do (and this seems like bread and butter to me, so I'm stunned that Project can't handle it), is say that George can work on this for 50% of his time, and Fred can work on it for 20% of his time. HOW LONG WILL IT TAKE?

Project seems unable to answer this; it seems to think that if two people work on a project, they'll each do 50% of the work, regardless of their %utilisation. Therefore, in the example above, the task will take as long as it takes Fred to do his half of the job at 1 day a week. That's not real-world, and it's not useful.

I'm sure Project is great for saying, "Fred's going to go and support Widget X on-site for 3 months, how much will it cost and how does it affect the other, duration-based tasks I've got?" - but it's crap for estimating how long a job will take for a given effort and resource (unless the resource is 100% utilised on the task).

As far as I can see, the only solution is to export the information into Excel, do some processing there to arrive at a task duration (basic maths, no reason why Project shouldn't be able to do it), then import it back into Project. Then, if anything changes, go and do it again... and again... and again. And this means that for effort-driven scheduling, the only thing that Project is actually any good at is displaying these calculations as a Gantt.

So I need something that can support effort-driven planning!
 
That's true, but how do you know how many people, and at what utilisation, to put on a task? You're saying you have to decide that before you enter the work, but the reality is that you'll want to play around with resourcing levels to see how it affects the schedule.

Here's an example of why Project doesn't work:

1. Create two tasks
2. On the first task, set the resources to "Me[50%], him[25%]"
3. Set the work to 5 days
4. On the second task, set the work to 5 days
5. Set the resources to "Me[50%], him[50%]"

The task information is identical for both, yet the durations are different. The one set up using your method is correct (effort / total utilisation (50 + 25 = 75%)). But don't get smug ;-)

6. Now, change the utilisation for both resources, on both tasks, to 50%.

Now the one set up using your method has an incorrect duration, whereas the one set up "properly" (to my way of thinking) is correct. But it's only correct by coincidence, because the utilisation for each resource happens to be the same. This is because Project divides the work by the number of resources, not the proportional utilisation of all the resources.

How can this be workable?
 
This is an old struggle for me. Here's my solution and conclusions:

The story goes: Working in a Manufacturing Engineering function doing project work, The Boss beats me up for missing all my schedules. As an experiment, I record my time for 3 weeks and realize I am only 40% efficient (too many meetings, etc.)

Solution: I make my project task list and schedule as if I was 100% efficient. I place lots of milestones and deadlines to anchor my tasks and provide "pivot points". I put in sufficient detail to drive it and account for the majority of activities. Then I return to my tasks and put my 40% resource allocation to it. Poof, MSP changes all the durations and the project begins to act like reality.

I start hitting all my schedules. I explain this to The Boss, he starts doing his job and reduces the amount of silly meetings and lets his team actually do work. Everybody wins. Eventually my efficiency increases and I start putting that into my project plans. Continuous immprovement.

Conclusion: the working business world is too dynamic to capture all the details and ineffiencies into MSP. You would spend weeks developing a project plan to do this. I've NEVER had the luxury of absolutely 100% perfect predictive task durations. I suppose they do exist somewhere, but I've never seen them. Everything is only your best guess, and MSP can be used effectively as a Prediction Tool. Some people create big fat task lists with lack of detail and then manage it day-to-day to fit in the their published deadlines. I choose a different approach. Put in a lot of detail (task lists provide daily to-do lists without thinking). But I realize I can never capture everything, and that conditions actually do change as time progresses (an amazing concept).

Then the truth is that most people only use MSP to create pretty pictures that get shown to Management at the Kickoff Meeting, then never again.

I set the expectations to Management / Customers that the prediction accuracy of the project is 100%-90% for a few weeks from the current date. Then the prediction accuracy rapidly falls off after that. Constantly updating and managing the project keeps managers informed and happy that there actually IS a plan, it is being worked, and provides a reasonable level of predictive accuracy. Besides, they have never met a schedule, either.

For projects that absolutely can't slip behind schedule, then there is a lot of fat built into it for recovery. This allows 100% success rate of hitting the milestones. But it's a very inefficient use of resources.

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering
 
We use MSP for initial scheduling and update it regularly as reality starts to bite. The fact that MSP can't accurately handle these changes without inaccurate forecasting of remaining work is exasperating and means that scheduling of work always take at least 5 times longer than it really should.

Which is why I'm desperate to find a tool that will accurately calculate and portray effort-driven, utilisation-flexible Gantts.
 
I've never used it, and don't know if it would suit your purposes. But I've heard of several "big" companies (aerospace, etc) using Primavera software . I don't know, but based upon what I've heard it's massive, expensive, requires dedicated resources, etc etc etc.

Then, again, I've been wrong about many things [shadeshappy] .

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering
 
Yep, unfortunately Primavera would be cost-prohibitive!

I've decided the only real solution is to copy-and-paste into Excel, parse the resources column and calculate the duration correctly, then type it back into Project.
 
I really love this whole discussion. My own conclusion is if MSP were to do it all then we would not be that necessary. I guess thats why MSP was added to MS Office because it can not stand alone really.
Correct me if I am wrong

xx
 
I don't expect MSP to do it all. I would expect it to be able to perform basic mathematics in support of a fundamental planning approach though!

I found a tool called OpenProj which does this adequately - unfortunately it has other shortcomings which means I can't use that either.
 
MS Project schedules tasks based on the project calendar until a resource is assigned then the resource calendar takes over. What I think you are seeing is based on the availability of the resource. Playing around with the settings of task type and effort-driven will not change the result. Effort-driven means that the work is distributed over the resources - it usually shortens duration. Fixed units means it is fixing the about of the resource allocation (how much you get of a resource).

Don't consider this an "age old problem" consider this as it is how the software works.
 
I think that you've misread the OP. The problem has to do with how MSP changes the duration and distribution of work, given multiple resources.

To recap, if you assign two resources, MSP splits the hours 50/50 regardless of the utilization factors of resources, rather than dividing the resources according to the utilizations. So, start with 100 hrs of work, if A is at 1% availability and B is at 99% availability, A still gets 50% of hours, jacking the duration out to 625 days.

The more correct distribution is to assign A 1 hr, and b 99 hrs, resulting in 12.5 day duration



TTFN

FAQ731-376
 
Obviously you are using effort-driven fixed work. If the calendars for the resources are the same other factors are either the assignment level or the max. units field on the resource sheet. Set resource units to decimal (tools, options, schedule, second option - decimal) and make sure the max. units are 1 for each resource.
 
Status
Not open for further replies.
Back
Top