Monday, December 27, 2010

Schedule heresy

Here's a little heresy on schedules, just before the holiday break:

I don't like, and don't recommend, MSProject and similar tools for managing a project! 


Plan v Manage

There's too much administration and faux assumptions day-to-day managing dependencies for it to  be an effective management tool, especially for smaller projects.  There's always a mad scramble and a lot of time and effort taken up on evaluating ad-hoc task level interactions, most of which can be worked out by other means.

On the other hand, it's a great planning tool to get a project started, including a consideration for dependencies, resource conflicts, and other artifacts.  As a planning tool, so long as dependencies are restricted to finish-to-start, and there are no fixed dates, it's a good tool to host Monte Carlo simulations.  It's just that once a project is under way, milestone charts, gated criteria, and earned value spreadsheets are better tools on account of their efficiency, even on large projects.

Walk the talk

I finished an engagement a couple of years ago that went several years, consumed multiple hundreds of millions of dollars, involved hundreds of project staff, and had four blocks of deliveries over two years. 

And, we never had a task-level project schedule. 

What we had were swim lane milestones and major dependencies identified between swim lanes.  Within the lanes, there were one or more teams and planned interteam dependencies, each team with their milestone schedules, and within teams, there were small working groups, also with milestone schedules.

We set up pipelines for sequential delivery, and gates to frame the pipelines.  We managed the pipelines with Excel.  [Look for more on pipelines in a future post]

We also did resource leveling with Excel.... that's a good thing!  The algorithms in schedule tools can return some silly answers.  I always marvel at Excel--it's truly amazing what you can do quantitatively with that tool!

We successfully delivered business value.  The first block was late--and that was a value bummer--but the next three hit their milestones on time.  We did not successfully earn the intended project EV package of cost-performance-schedule.  It was an ERP project [Oracle business systems] for a multi-billion$ enterprise.  Re EV: ERP says it all!

In any event: Plan with MSProject [or Primavera, or other similar], but manage with milestones, gates, and EV

 Bookmark this on Delicious
Are you on LinkedIn?    Share this article with your network by clicking on the link.