Project Estimation and the M.E.W.S.

I was in hospital a while back visiting a sick relative, and whilst they were sleeping I was reading some of the charts on the wall. One in particular caught my eye. It was essentially a fast risk measuring tool. Being a consummate project manager, my first thought was, "I can steal this idea!"

The approach below is based on the MEWS (Modified Early Warning Score) a medical tool used to get an early vision on the heath of an at risk patient, and then customized for a particular client/project.

Each projects estimates should be scored against the below criteria, and the sum collated.

The sum is then matched against the score explained list, with the higher the score meaning the riskier the project.

It is a simple way of ensuring that when estimating projects you are validating the estimates against past experience and experience.

E.W.S Grid

0 1 2 3
The project is new functionality No new functionality. Something we have used elsewhere one. New but well understood in the industry Bleeding edge idea.
The project touches multiple Internal systems Only uses internal Systems Project uses well documented APIs The project uses unfinished 3rd party APIs or well documented systems The project uses unfinished APIs and poorly documented 3rd party systems
Includes new infrastructure No new infrastructure One new server on all environments Multiple new servers on every environment OR 1 new server type Multiple new servers on every environments & new server types (e.g. Messaging)
The project touches third party systems The project is a total greenfield The project works with 1 existing system The project uses 2 external systems The project uses 3 or more systems.

E.W.S Score Explained

Score Rating
0-2 Easy project. Estimates should be OK.
3-4 Medium risk - Consider add 20% contingency
5-7 Medium to High risk - consider 30-40% contingency
8-10 High risk - consider 50% upwards contingency, or separating the project into new work-streams
11+ Very high risk - reconsider doing this project in the current fashion

Please note, the idea of the above tables is that you take your own experiences and build the table around them, not that the above options are a one size fits all solution for estimating risk.