John AyersAuthor, writer and consultant| Self EmployedLakeville, Ma, United States
According to reports I have read, 97% of IT projects use Agile. One of the problems I have read about is how you measure performance of a project as it evolves. The most effective way to do it is with earned value management. The problem with Agile is it does not have a baseline and you need one for earned value management.
Given this statement, do you think it is feasible and necessary to implement earned value on a large IT project? Saving Changes...
Using EVM on large IT projects is certainly feasible. Using EVM itself is not necessary.
EVM is the most recognized way, but whether or not it is the best way is an opinion, not a given. Often it is used very ineffectively. The most value of using EVM is identifying variances early in a project. Unfortunately that is usually the most loosely defined in terms of discrete tasks assigned a PV. This means EVM is often a method to track variances through project completion rather than managing the variances. Saving Changes...
I'd challenge the assertion that 97% of IT projects use agile. If they mean a particular agile technique, tool or practice, possibly, but no where near that percentage is following a true adaptive life cycle, especially when you look at infrastructure projects or those involving legacy environments.
EVM and agile are not mutually exclusive. If anything, the practice of calculating sprints to complete based on historical sprint velocity is one form of EVM adapted for such contexts.
Projects following an adaptive lifecycle could also lock cost & schedule early on, allowing scope to be flexible. In such cases, you would definitely have a baseline for those constraints to work with.
Kiron Saving Changes...
Anonymous
EVM is feasible but not very useful in knowledge-work. The idea of measuring productivity against a baseline plan is counter-productive because doing so discourages innovation and collaboration and puts pressure on product quality. Baselining also misses the point since, in contrast to other types of endeavour, the potential value of technology projects so often doesn't have much correlation with actual cost or schedule. Saving Changes...
I feel there is an important distinction to be made with regards to value. EVA assesses whether, or not, you are getting what you paid for according to schedule. This can be important, but is not the emphasis of most agile approaches/frameworks/methodologies, call them what you will. They are more concerned with the delivery of product that can start producing value, preferably before the project is finished.
This is relevant to the conversation. If you really needed to use EVM on a project (contractual obligation, maybe?), could you use it at the MVP or "release" level, instead of the full project? The work, at this level, is more likely to be defined and trackable, with set delivery dates. There may still be some slippage, but your results will be more reliable than trying to use it on something that won't be fully defined any time soon. Saving Changes...
Sergio Luis ConteHelping to create solutions for everyone| Worldwide based OrganizationsBuenos Aires, Argentina
While I can debate about some of your comments let me say that you can find the adaptation of EVM for Agile based projects. In fact, in the last PMI´s EVM Standard as far as I remember you can find about it. Saving Changes...
Community Champion
