4.4 OLD DATA-LOOK IN THE BIN

---

4.4 OLD DATA—LOOK IN THE BIN

Now if all of this is good advice and you take it on board should you expect to see an improvement in estimation effectiveness? Could you prove it?

You may remember earlier that I said there is at least a twenty-year-plus history of software development in some organizations. You would expect to be able to get hold of some data from these organizations. Think how likely it would be to find a similar project to the one you have to estimate today if you had twenty years worth of data to fall back on. You will be lucky if that data is available. Well, okay, ten years worth would be good enough; in fact five years worth would probably be better as it is probably closer to what we do today.

You may like to go into a project team and ask them if they have data, even just estimation data, from their last project. If the answer is yes then you are in luck! Many teams do not even keep the past estimates for their current projects.

Keeping data about itself and its work is not something the IT industry does well, which is ironic when you consider what "IT" stands for. If you wish to do something practical then you should insist that your projects start to keep some basic data about themselves. It does not have to be a sophisticated or complete system of measurement. Raw data such as size of project, estimated cost, planned duration and the updates to these figures together with the actuals can be a good starting point. Get really clever by asking the project manager to summarize his or her feelings about the project. What made this one different from other projects that the manager had been involved in? Keep this summary short, no more than one page, and you have a very useful source of information even if it is somewhat subjective.

To summarize, there are certain basic principles that appear to apply to the estimation process no matter what environment you consider. These principles are:

• The first estimate is likely to be an underestimate. Act accordingly.

• Estimates are estimates, so use bounds and not absolutes.

• Cost is estimated, duration is planned.

• Estimates are dynamic; re-estimate when milestones are completed and additional information becomes available or when the requirements change.

• The best estimation technique is to use an expert, if you can find one.

• The estimate is the responsibility of the estimator, not the tools or techniques.

• Keep previous estimates.

---