September 14, 2007

Estimation - Do You Have the Crystal Ball

Do You Have a Crystal BallEvery project manager or team lead is invariably asked the question – How long will it take? What will it cost? How many FTEs do you need? Anyway you look at the questions – they are all asking for a prediction. And what better way to make predictions than use a crystal ball.

Think about all the myriad reasons why you cannot accurately predict the future – in terms of the project cost or duration. Here are a few …

- You don't know everything
Which is why you assume. You validate your assumptions, list and document your assumptions, but what you don't know – you have to assume. Most plans will have a list of assumptions around the critical areas of the project – scope, resources, durations, technology, dependencies, cost etc. Use assumptions judiciously to track and plan for the known unknowns (Unknown unknowns are dealt under the change section).

- People matter
Believe it or not but people make a huge difference in how projects are executed. The skill set of your team, their morale, attitude etc. all play a huge role in the execution of the project. How will you staff your project, what level of attrition will you face, will the team `gel' together and work in a cohesive fashion – has a huge impact on the project cost and success. Your estimation model should be explicit about staffing requirements and related assumptions.

- Change is inevitable
We say it, live it but don't always plan for it. One way to counter the effect of `change' is by instituting effective contingency management. Is contingency planning a basic component of your planning effort? Do you prioritize and evaluate the probability of change on your project? Are these changes captured as appropriate risks and mitigated with corresponding contingency plans?

- Technology is the easy part
Not always true. In today's dynamic IT environment, technical components are evolving faster than you can deliver projects. Technology consolidation, new products/versions, new integration challenges, new architectures and not enough maturity pose a significant cost component on most projects. Accounting for realistic cost for architecture management, infrastructure management and deployment activities can lead to a better planned effort.

- And what about data
There is a tendency to focus on the application components more than the data components. Not the best strategy since most systems use data as an integral part of the business solution being delivered. Data planning should include data modeling, database management, data conversion, test data management, test id management etc.

These are some of the estimation challenges a PM faces. Focus on these and similar components of project estimation and your dependence on the crystal ball may diminish.

No comments: