17 Agile Estimation

Agile Estimation is the practice of approximating the effort, complexity, or duration required to complete backlog items, providing the information agile teams need to plan realistic iterations, forecast delivery timelines, and support informed prioritization decisions. Rather than pursuing precise, deterministic predictions, agile estimation embraces approximate, relative judgments that are refined continuously through experience, reflecting the recognition that early estimates are inherently uncertain and that this uncertainty diminishes as work is actually undertaken.

The Purpose of Estimation

Supporting Realistic Planning

Estimates give a team the information needed to determine how much work can realistically be committed to within a given iteration, preventing overcommitment that leads to incomplete work and undermines predictable delivery.

Informing Prioritization

Alongside value, effort estimates are a key input to prioritization, since items offering substantial value at relatively low effort often represent particularly attractive candidates for early delivery.

Enabling Forecasting

Aggregated estimates, combined with a team's historical delivery rate, allow forecasting of when larger bodies of work are likely to be completed, providing stakeholders with reasonably grounded expectations without requiring exhaustive, detailed scheduling.

Approaches to Estimation

Relative Estimation

Rather than estimating absolute time or effort directly, many agile teams estimate the relative size of an item compared to other, previously estimated items, since people tend to be more accurate and consistent when comparing relative magnitude than when predicting exact duration.

Story Points

A common unit for relative estimation assigns items a number of points reflecting their overall size, factoring in complexity, effort, and uncertainty together into a single abstract measure rather than a literal time duration.

Time-Based Estimation

Some teams, particularly those working on more predictable or operationally routine work, estimate directly in hours or days, favoring a more concrete unit at the potential cost of the flexibility relative estimation offers for highly uncertain work.

T-Shirt Sizing

For rougher, earlier-stage estimation, some teams use broad categories such as small, medium, or large to quickly gauge relative scale before more detailed estimation becomes necessary as an item nears active development.

Techniques for Generating Estimates

Planning Poker

A collaborative technique in which team members privately select an estimate for an item and reveal their choices simultaneously, discussing significant discrepancies before converging on a shared estimate, helping surface differing assumptions that a single individual's estimate might miss.

Affinity Estimation

Grouping items relative to one another based on perceived similarity in size, without assigning numerical values until the groupings are established, can accelerate estimation for a large batch of backlog items.

Reference Stories

Maintaining a small set of previously completed items as reference points for a given size category helps ground new estimates in concrete, shared experience rather than abstract judgment alone.

Using Estimates in Practice

Velocity Tracking

Measuring the total estimated size of items a team completes per iteration, commonly called velocity, provides a basis for forecasting future capacity, assuming reasonably consistent team composition and working conditions.

Adjusting Estimates Over Time

As a team gains experience with a particular type of work, its estimation accuracy tends to improve, and teams periodically recalibrate their sense of scale based on actual outcomes compared to original estimates.

Limitations and Considerations

Estimates Are Not Commitments

Agile estimation is understood to produce approximations subject to real uncertainty, and treating estimates as fixed, binding commitments risks encouraging inflated padding or discouraging honest acknowledgment of complexity.

Diminishing Value of Excessive Precision

Investing significant additional effort to increase estimation precision beyond what is useful for planning purposes often yields limited practical benefit, since the underlying uncertainty in early-stage work frequently exceeds what greater estimation rigor could resolve.

Team-Specific Calibration

Estimates, particularly relative units such as story points, are calibrated to a specific team's shared understanding and are not directly comparable across different teams without careful normalization.

Agile Estimation provides agile teams with the approximate, continuously refined understanding of effort and complexity needed to plan realistic iterations, prioritize effectively, and forecast delivery, embracing uncertainty as an inherent and manageable feature of estimating work whose full scope only becomes clear through the process of actually doing it.