Tracking Status for beginners

[tdemelle]

I think I may be missing some critical concept when it comes to tracking progress in Project.

I work in a very complex production environment with a high degree of fluctuation in tasks. I've managed to get to the point where I have thousands of tasks with durations, resource assignments and dependencies in place that are automatically being updated by a tool that imports and merges task data from a request and task tracking tool on a daily basis.

Durations and priorities are added to all the tasks that have been imported into Project where they are then leveled and massaged (dependencies removed, durations fine tuned, etc as necessary) to model a way to meet all known goals by their deadlines.

New multi-step, dependent tasks are imported into Project daily that are costed, prioritized and leveled. Work is being done throughout this process, sometimes ahead of requests being filed so sometimes tasks come in that are already completed.

Here's where I run into my current Project challenges. I often find myself with tasks in Project that are already fully or partially finished ahead of the leveled schedule. When I mark them 100% complete, although the gantt bar gets filled showing 100% completion, the start, finish and duration of the task don’t change and the resource assigned to the task remains booked into the following week on a task that is already marked as done.

I thought I might be able to get around this by filling in “actual duration” with the (smaller) number of actual days the task took, but that adjusts % complete instead of shortening the gantt bar (even when the initial duration is marked ‘estimated’).

I routinely get to the point where my Project is so far out of whack with reality because resources are assigned to work that's already complete, I manage it by adjusting duration of completed tasks to 0, changing the project start date to today’s date and leveling to get a good forecast for what work remains. Unfortunately, apart from being hacky, this method undermines the ability to meaningfully track overall % complete.

I'd really love to get my head around Project enough to be able to use it to better manage this complex production process.

What *should* a typical status update involve? "% complete" obviously, but then also adjusting "Duration" for each task to correct for errors in the initial scheduling estimate? How then do you compare reality against your initial estimates? Would that not more appropriately be handled by adjusting "work"?

Thanks for any help or tips you can offer!

The best way to handle updates to the schedule especially where actual start
and finish dates differ from planned, is to fill in the actual start date and
actual finsh date, not the actual duration.

Using the "update tasks" button complete the actual start and actual finish
dates. Project will calculate the actual duration and the whole schedule
will then be recalculated to reflect the status. This should also complete
the assignment of resources.

Welcome to this Microsoft Project newsgroup :)

This simple answer is to not use %Complete when updating. Use Actual Start,
Actual Work and Remaining Work, thus letting Project calculate the
percentages. If you enter %Complete without and actual data, Project will
assume the original planned dates have been met.

[tdemelle]

Thanks for the tips… I really appreciate the thoughtful replies! The level of support on this forum is remarkable!

I’m trying to enter status updates using your suggestions staying away from %complete but am still finding Project’s behavior a bit confusing. I’m going into the Tracking table and have grouped together columns for Actual Start, Actual Work and Actual Work remaining. Here’s what I’m seeing as I try to update status:

When I enter the actual start date of a task, Project behaves as expected by moving the work’s Gantt bar so that it begins at the actual start date. So far so good!

When I adjust the ‘actual work’, the start date jumps five months into the past! I am able to correct this to an extent by going back and re-entering the actual start date. When I do this, however, the work that is complete moves where it should but the overall task bar gets split with the work remaining pushed to the very end of task list. Re-leveling even with an appropriately high priority does not seem to be able to rejoin the work to be done with the already completed portion of its Gantt bar and the work to be completed remains stranded in the distant future far beyond when the actual work is going to take place.

Hrm.

Here’s another one I’m curious about. It seems when a task has a predecessor its start date gets locked at the predecessor’s end date even if that puts the work to be done on the task entirely in the past. I’d like to use predecessors to ensure tasks follow one another in a certain chronological order, but it’s not important that they follow one immediately after the next (and in this case that presumption is creating scheduling problems!) Is there a way to have successor tasks lag when leveled in those types of situations without necessarily having to add a specific lag figure or manually delete the dependency?

I can't reproduce your problem. Do you have any tasks constrained?
View/Table/More Tables.../Constraint Dates. It is best to have only As
Soon As Possible, as that is, in effect, no constraint at all, thus leaving
Project to schedule with the maximum flexibility. Constraints can create
the sort of problems you are experiencing.

Once again Michael, you are referring people to this same newsgroup. In
which community are you reading these postings? I would recommend a
newsgroup reader or a least use the microsoft site here:
http://www.microsoft.com/office/comm...ang=en&c r=US .
The google site is way behind.

sure I referre to this newsgroup since I found the mentioned thread
here ;-)
Whats the problem?

It's not a problem, but why waste your time and the time of anyone reading
this newsgroup (microsoft.public.project) being referred to this newsgroup
(microsoft.public.project) that they are already reading? It just does not
make any sense!

I think the "problem" is that Michael is trying to point to a specific
thread (hence saving everyone time in hunting), but the fully URL for
that thread isn't actually getting published. For me the URL's are
clipped off. Why? dunno.

Here is the link -

http://groups.google.com/group/micro...a7aa44c479ac0b