The Power of Prediction
If you want to predict how long it will take to complete a project, one of the key elements you will require is the amount of time each step of the project will take to complete. For example, if you want to predict how long it will take to build a four bed roomed house, then from a simplistic standpoint, you will need to know how long it takes to the following:

• Have an architect draw up the plans
• Obtain council resource consent
• Lay the foundations
• Build the walls and put on the roof
• Do the wiring and plumbing - gibstop the walls

If you have already been through the process of building a flat, a 2 bedroomed and a six bedroomed house, and you recorded the time taken for each of these steps detailed above for each dwelling, then you have the raw data needed to help you predict how long it will take to build the 4 bedroomed house you require.

Now instead of building a house, imagine that you work or own a Web Site development company. When you submit a tender for a job, there is a high probability that the company that provides the best quote, will win the job. BUT, and here's the kicker, over estimate on costs of the job, and you may out price yourself relative to your competition. Under estimate the job, and you may win the contract, but end up losing money because you cannot complete the work in the time allocated. This means that those web design companies that have the ability to predict how much design of a web site will cost will have a distinct competitive edge.

In addition to MetriQ®
Before you can begin making any predictions, you will of course need your raw timing data and MetriQ is the ideal tool for the job, allowing you to accurately time how long it takes to create new images, format your html pages, write the text, and add your php code to talk with a relational database for your shopping cart functionality.

For the next part of the estimation process, we have begun the process of designing a new companion software module for MetriQ that will utilize Case Based Reasoning in an effort to help web companies make meaningful predictions for new and existing web design projects which we hope will be released in 2008.

See also the TukuTuku Benchmarking Project