How to: Use Estimation in Agility

This feature is available in all editions

Agile Insight

In Agile development, planning occurs at the release, sprint, and day levels. For the planning to be effective, the planner must have enough information about the items being planned to make a good decision. On the other hand, getting the information takes time and effort, so having more information than necessary is a waste of time. To support planning, two types of estimates are generally used: feature estimates and task estimates. To learn more, see:

Feature Estimation

The feature estimate is an estimation of the feature in size. Size estimates can be created quickly by the software development team and provide enough accuracy for the product owner to prioritize features and create a release plan.

With size estimates, new features are compared to previously estimated features. Story points are the most popular convention for units of this estimate. When looking at a group of features, a software team selects a small feature as being a single story point and then estimates the rest of the features in relation to that feature. A four-point feature should be twice as big as a two-point feature and four times as big as a one-point feature.

In Agility, the feature estimate is recorded in the Estimate field of a story/backlog item or defect.

• See Agile Feature Estimation for more information.

Task Estimation

Most teams identify the tasks that will be required to implement a feature as part of their sprint/iteration planning meeting and assign an estimate in hours to the task. The purpose of this estimate is to give the team insight into how much work they are committing to in the sprint. During the sprint, burning down the tasks also provides visibility into the progress of the team for both the team and external stakeholders.

In Girish, the task estimate is recorded in the Detail Estimate field of the task. Additionally, all of the task estimates for a feature are rolled up into the Detail Estimate field of the feature.

• See Iteration Planning (to come)