A Few Notes
We end this chapter with a few more release-planning thoughts on estimation and risk—and Pareto versus Parkinson.
On Estimation and Risk
Many people think that there is risk attached to missing your estimate. At worst it might be embarrassing; however, the real risk is in missing your delivery dates. It is not important to be able to predict at the level of the story; what is important is predicting at the release level.
Risk also plays a role in prioritizing features. Usually, we prioritize by the business value each feature represents—possibly offset by the cost of creating it. However, sometimes prioritization is affected by the potential cost of delay. For example, let's say we have Feature A and Feature B. Feature A may be twice as important as Feature B, but we need Feature B for a conference coming up in three months. We may actually do Feature B first to ensure its completion before the conference if delaying Feature A is not too costly.
Pareto versus Parkinson
We have heard Lean software development likened to following Pareto's Law: 80 percent of the value comes from 20 percent of the work. In other words, find that 20 percent of features that will provide your customers with 80 percent of their value; then, find the next features that will provide the greatest value to your customers.
The problem with this is that if there is no time-boxing—no end-date—Parkinson's Law may apply: "Work expands so as to fill its time for completion." Parkinson's Law is particularly dangerous when multiple product managers are competing for a team's resources. Manager A is focusing the team on one thing and Manager B is concerned about when she will have the team's availability. You can counteract the effect of Parkinson's Law, by having the team follow Pareto's Law in the shortest amount of time they can. In other words, have the team always focus on building the smallest things as quickly as they can, end to end, while ensuring quality.
Add the most value possible in the least amount of time possible with the right level of quality.