Original Post Date: Friday, August 13, 2010

If you want to read an interesting article on EVM – check out ‘The Three Deadly Sins of EVM’  by Mike Mullaly.  In it he reflects some of my personal feelings about EVM but he does this much more eloquently than ‘it’s a crock’.  OK – while I have actually said that out loud – it’s probably a little too strong.  I do think that EVM may be a good tool to have in the toolbox – it’s just not the project panacea that so many make it out to be.  And it really does require organizational awareness of what is and is not being measured and what is meant by ‘value’.
 

I remember many years back taking a project management course where EVM was taught.  Although the concepts are simple enough – I did not get how the principles being taught could have been applied to any real (software) project that I ever worked on.  Actual work planned at the beginning of a project is rarely actual work that ends up getting done.  And how precisely do we measure doneness?  As far as I can tell few software features are ever truly done - there are tons of ways you can make them better. So it really comes down to being done enough. So a feature may be ‘done’ in the eyes of the project manager but not in the eyes of the customer or stakeholder. As you progress in the project, new features may alter the doneness of existing features.  And where does quality fit into this ‘value’?  

Well run, tightly scoped projects with little scope creep and excellent project and change management practices are less likely to suffer from the ambiguities outlined above.  (How many of us work on one of these every day?)  These of course are the projects least likely to run out of control in the first place.  The things that make a project risky are the things that make it easier to misread what EVM is telling you, creating an allusion of control for a project which has little.  It is unwise to rely on EVM results as the only measure of a project's health, rather, one should heed Millaly's suggestion that blind adherance to earned value could cause your project to crash and burn.