Q: Does Earned Value work for software projects?
A: Absolutely, Earned Value Analysis (EVA) is a statically valid reporting approach that can be applied to any endeavour. It compares actual progress and spend against projected progress and spend.
Q: Can you use Earned Value on Agile projects?
A: You can, but I would not recommend it. There are fundamental problems using EVA on agile projects relating to baseline plan quality. Also there are better alternatives available for agile projects.
Earned Value analysis and reporting measures conformance and performance to a baselined plan. So, given that on agile projects we know that our initial plans are likely to change, why track progress against a weak plan?
An excellent reference to Earned Value reporting is “Earned Value Project Management, Second Edition” by Quentin Fleming, Joel M. Koppelman. In it the authors list 3 critical success factors for Earned Value reporting.
1. Quality of the project’s baseline plan. Earned Value is compared against the baseline plan, whether the plan is accurate or not. Therefore, cost ‘overruns’ will occur if the project costs are under-budgeted, and scope creep will occur if the initial project scope hasn’t been adequately defined.
2. Actual Performance against the Approved Baseline Plan. i.e. whether the actual performance tracks to the baseline plan.
3. Management’s Determination to Influence the final results. Final results for a project based on earned value projections can be modified based on management’s commitment to take action as soon as deviations from the plan are observed.
Agile projects fail to meet the first two critical success factors. First, the quality of the original baselined plan is very low from a completeness and scheduling perspective. We could try to use EVA against our continuously evolving plans, but the figures become meaningless if you keep changing the baselined plan. Each time you do so the performance indexes and variances change making tracking and forecasting extremely problematic. Secondly, since our initial baselined plan may be little more than “today’s best guestimate“ critical success factor number 2 “whether the actual performance tracks to the baseline plan” would be more a matter of luck than anything else.
Given these challenges, what should replace EVA on Agile Projects? Are there agile equivalents of Cost Performance Index (CPI), Schedule Performance Index (SPI), Cost and Schedule variances (CV and SV), and Estimate at Completion (EAC)? The good news is that all of these metrics can still be obtained and even better, if you are using Cumulative Flow Diagrams, you are half way there already.
A colleague, Anthony Cabri of Quadrus Development and I wrote the attached Research paper “Agile and Earned Value Reporting” for this year’s Agile2006 conference in Minneapolis. It explains Earned Value reporting, the problems with agile usage and provides alternatives for each of the EVA metrics. I hope you find it useful and welcome your feedback.