Throughput Measurement

Throughput Measurement is the practice of counting the number of discrete work items a team completes within a given period, without regard to the relative size or estimated effort of those items, providing a delivery indicator based on item counts rather than the story-point totals used in velocity measurement. It offers an alternative, and in many contexts complementary, way of tracking a team's delivery pace, one that some teams prefer specifically because it avoids the estimation step that velocity depends on.

What Throughput Measures

A Simple Count of Completed Items

Throughput is calculated by counting how many work items, regardless of their individual size or complexity, a team finished within a defined time window, such as a week or an iteration, giving a straightforward numerical measure of delivery pace that does not require any prior estimation of relative effort.

Independent of Estimation Practices

Because throughput relies only on counting completed items rather than summing size estimates, it can be tracked even by teams that do not practice relative sizing at all, or that wish to reduce the overhead associated with estimating every item before it is worked on.

Throughput Compared to Velocity

Different Units, Different Assumptions

Where velocity assumes a consistent, team-calibrated relationship between story points and effort, throughput makes no such assumption, treating every completed item as a single unit regardless of its actual underlying size, which is simpler to calculate but can obscure meaningful differences between a very small item and a very large one.

Reliability Depends on Item Size Consistency

Throughput tends to be a more reliable indicator when the team's work items are reasonably similar in size to one another, since a period in which the team happened to complete several unusually large items would show a throughput count similar to a period of many small items, despite representing very different amounts of actual effort.

Using Throughput for Planning and Forecasting

A Simpler Basis for Probabilistic Forecasting

Throughput data is particularly well suited to a specific style of probabilistic forecasting that treats each historical period's completed count as a sample of what future periods might produce, an approach that avoids the need for any explicit size estimation and is discussed further in dedicated forecasting practices within this same topic area.

Supporting Iteration Planning Without Estimation Overhead

Teams that plan iterations based on throughput simply ask how many items, on average, they have completed in recent comparable periods, using that historical count directly to gauge how many new items might reasonably be taken on, bypassing the story-point-based capacity calculation that velocity-based planning requires.

Calculating Throughput

Basic Period Throughput

The most direct calculation counts completed items within a single defined period.

Average Throughput Across Multiple Periods

As with velocity, a rolling average across several recent periods smooths out the natural variation of any single period, providing a more stable reference for planning purposes.

A Throughput History Visualization

Ensuring Throughput Remains a Meaningful Indicator

Watching for Systematic Changes in Item Size

Because throughput does not account for size, a team that begins deliberately breaking work into smaller items, or conversely starts tackling larger items than before, will see its throughput count shift even without any real change in overall delivery capacity, making it important to watch for such shifts in typical item size when interpreting a trend.

Applying the Same Data Quality Standards as Any Metric

Throughput depends on the same reliable, consistent underlying data discussed under Measurement Data Collection and Measurement Data Quality, since an inconsistently defined notion of "completed" would distort item counts just as it would distort any other flow-based metric.

Common Pitfalls

Ignoring Item Size Drift Over Time

Comparing throughput across periods where typical item size has meaningfully changed produces a misleading impression of the team's actual delivery trend, since the count reflects the number of items rather than the amount of work they represent.

Mixing Throughput and Velocity Inconsistently

Using throughput for some planning decisions and velocity for others within the same team, without a clear and consistent rationale, can create confusion and make it harder to build a coherent historical record for forecasting purposes.

Treating a Single Period's Count as Representative

Basing a planning decision on the throughput of only the most recent period, rather than an average across several, exposes the team to the same risk of reacting to ordinary short-term variation that affects any metric evaluated without sufficient historical context.