I take it EVMS is an acronym for Earned Value Management Software, or something similar?
The organisation that I work for uses InControl. You can enter progress, as percent complete, at the deliverable or WBS-element levels and it calculates all the earned values and performance indicies. It pretty good once you get to know your way around the software.
Primavera has a bunch of papers on it posted to the website. Basically as I understand it, you assign costs and revenues to your activites and as you update your percent complete, you can determine your profitability or lack there of and compare it to a base line schedule. Not a bad idea but not the rocket science it is made out to be.
Agreed with DRC1 that earned value management is really pretty simple. It can easily be done on a spreadsheet.
Inputs:
Budget at Completion, BAC
Actual Cost, AC
Planned Value, PV
Percentage Complete, PC
Estimate to Complete, ETC
Outputs:
Earned Value, EV = BAC x PC
Cost Variance, CV = EV - AC
Percent Cost Variance, CV% = (CV/EV)%
Cost Performance Index, CPI = EV/AC
Schedule Variance, SV = EV - PV
Percent Schedule Variance, SV%, = (SV/PV)%
Schedule Performance Index, SPI = EV/PV
Estimate at Completion, EAC = AC + ETC
Variance at Completion, VAC = BAC - EAC
Percent Variance at Completion, VAC% = (VAC/BAC)%
To Complete Performance Indices
- To achieve BAC
TCPI = (BAC - EV)/(BAC - AC)
- To achieve EAC
TCPI = (BAC - EV)/(EAC - AC)
The estimate to complete can also be calculated independently:
- Future performance influenced by past cost performance
ETC = (BAC - EV)/CPI
- Future performance influenced by past cost performance (80%) and schedule performance (20%)
ETC = (BAC - EV)/(0.8*CPI + 0.2*SPI)
- Future performance influenced by previous three CPI results
ETC = (BAC - EV)/(CPI1 + CPI2 + CPI3)/3
- Future performance influenced by past cost and schedule performance
ETC = (BAC - EV)/(CPI*SPI)
One shortfall of EV is that it assumes that costs proceed lineraly with % complete, i.e te cost to perform 20% of the work is half of the costs to complete 40% 0f the work and a quarter of the cost to complete 80% of the work. In actuallity (at least in construction) it is rarely linear. Generally the first 20-25% involves some setup, learning curve and similar problems. The middle (hopefully)goes well and the last 20% falls off due to having fewer units to employ resources against, cleanup and dimantling and corrective work on previous units. In the end, the average should be close to the actual production rate. However, projecting costs to complete off the 20% completion mark may give a very conservative answer.
DRC1, I respectfully disagree. Earned Value doesn't assume that costs are incurred linearly on a project. Instead, Earned Value is a performance measure for comparison against the Actual Cost and the Planned Value. The Planned Value is almost always plotted (time-v-cost) as an S-curve, with a flatter gradient at commencement and completion, precisely for the reasons you describe above.
I stand corrected. I guess if you took the time to use an proper curve for the planned value, it would be pretty slick. I have only seen it used with a linear proportion for planned value and hence my lack of enthusiasim. However, if you were to look at the Planed Value predicted in a realstic manner, I bet that would give interesting (and useful) results.
As a matter of further study if you use MS Project or P3 you can look at resource loading the schedule and then querying the progress. The baseline schedule will generate the planned value curve and you can plot the earned value curve against it to confirm winning or losing.
