Skip to content
Tutorials Center
My learning

Topics

Lesson 12 of 29 ยท Flow and Metrics

Velocity, Burndown and Cumulative Flow

Understand velocity, burndown and burnup charts, cumulative flow diagrams and throughput, compute them with Python, and learn how metrics get misused.

Agile teams make decisions in short loops, so they need quick, honest feedback about their own progress. A few simple measures give it to them: velocity (how much the team finishes per Sprint), burndown and burnup charts (how a Sprint or a release is progressing), the cumulative flow diagram (where work piles up) and throughput (how many items finish per period).

These measures are tools for the team and its Product Owner, not a scorecard for managers. This lesson explains what each one shows, how it is computed, and how it gets misused. The runnable examples are Python programs that read small CSV-like tables, compute the numbers and draw simple text charts. They are stand-ins for the charts built into Jira reports, which compute the same numbers from your real issues. You will see story points here, so Estimating with Story Points is a useful prerequisite.

Four measures, four questions

Velocity How much do we finish per Sprint?
Burndown How much is left, day by day?
Cumulative flow Where is work piling up?
Throughput How many items finish per week?

Use them to ask questions, not to hand out grades.

1 / 2

Velocity

Velocity is the amount of work, usually in story points, that a team completed in a Sprint. Only items that fully meet the Definition of Done count. If a team finished 14 points in Sprint 1 and 18 in Sprint 2, its velocity is 14 and then 18.

Used well, velocity helps with forecasting: the average of the last three or so Sprints is a reasonable estimate for the next one, and the spread tells you how reliable that estimate is (see the forecast in Estimating with Story Points). A team whose velocity swings between 10 and 30 is telling you something about its process or its stories.

Velocity does not measure productivity, quality or value. It only describes the team's own history, in the team's own points, so you cannot compare it across teams. A common companion is predictability: the percentage of the committed points that were actually completed. It shows whether the team plans realistically.

Computing velocity

The first program reads Sprint data from a CSV text, then prints velocity, predictability and a few summaries. It uses only the Python standard library, so nothing needs installing.

Python Velocity and predictability from sprint data

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Notice that Sprint 6 had the highest commitment (26) and one of the lowest results (19). That is a typical pattern after a streak of good Sprints: the team (or its manager) raises the target and the plan stops being realistic. The mid-sprint additions column explains part of it: new work arrived while the Sprint was running.

Burndown and burnup

A Sprint burndown chart plots the work remaining against the days of the Sprint. A straight "ideal" line goes from the starting total down to zero at the end. The real line wobbles around it. The shape tells a story:

  • Flat at the start, then a cliff: work was started but not finished, then finished at the last minute. Work is probably too coarse, or finished late.
  • Always above the ideal line: the team is behind, or the plan was too ambitious.
  • Line goes up: work was added or re-estimated upwards.

A burndown chart has one weakness: it mixes progress and scope change. If the team finishes 5 points and a stakeholder adds 5 points the same day, the line stays flat, and nobody can tell why. A burnup chart fixes this by plotting two lines: the total scope (which rises when work is added) and the work completed (which only rises). The gap between them is what is left. Burnup charts are especially good for release-level views, where scope almost always changes.

flowchart LR
  subgraph Burndown
    A["Shows: work remaining"] --> B["Hides: scope changes"]
  end
  subgraph Burnup
    C["Shows: work done"] --> D["Shows: total scope"]
  end

A burnup chart keeps scope changes visible instead of burying them in the remaining line.

Reading a burndown and a cumulative flow table

The second program takes one Sprint's daily data and prints the ideal and actual remaining work with a little bar. Then it prints a cumulative flow table: how many items were in each state at the end of each day.

A cumulative flow diagram (CFD) stacks the number of items in each state (To do, Doing, Done) over time. Each band's thickness at a given day is the number of items in that state. A widening "Doing" band means WIP is growing, which signals a bottleneck. The vertical distance across the "Doing" band is the amount of WIP, and the horizontal distance approximates the cycle time. Parallel, evenly rising bands are the sign of smooth flow.

Python Burndown and cumulative flow in text

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

In the burndown output, the actual line sits above the ideal one for most of the Sprint and, on day 5, scope jumps from 30 to 34. A burndown alone would only show a flat line that day. The scope column reveals the cause. In the text cumulative flow, "." marks items waiting, "o" marks items in progress and "#" marks items done. The o band grows to five or six items and then holds, a steady WIP, while the # band grows by one to three items per day. That shows a healthy flow.

Throughput

Throughput is the number of items finished in a period, for example per week. It counts items, not points, so it needs no estimation at all. Many Kanban teams forecast by throughput: "we finish about 4 items a week, so 20 items need about 5 weeks". Because story sizes vary, throughput works best when the items are small and of roughly similar size. Teams often plot throughput as a histogram and forecast with ranges (the project Release Forecast from Velocity goes deeper).

Throughput, cycle time and WIP are linked by the same relationship you met in Kanban and WIP Limits: lowering WIP shortens cycle time without needing anyone to work faster.

Misusing metrics

Metrics are easy to abuse, and the usual outcome is that the number improves while the real thing gets worse. A few classic traps:

  • Velocity as a target. Once management demands higher velocity, estimates inflate, quality drops and the number stops meaning anything. This is a known effect often summarized as: when a measure becomes a target, it ceases to be a good measure.
  • Comparing teams. Points are team-specific. A velocity of 40 in one team and 20 in another says nothing about which one delivers more.
  • Individual velocity. Software is built by teams. Counting points per person discourages helping, pairing and reviewing.
  • Burndown as a deadline weapon. Treating the ideal line as a promise leads to hiding work and rushing items to "done" without finishing them properly.
  • Ignoring the context. A dip in velocity after a holiday, a new team member or a production incident is not a performance problem.
  • Measuring what's easy, not what matters. Output (points, tickets) is not outcome (users helped, problems solved).

A better habit is to use metrics for conversation. Put the chart on the screen at the Retrospective and ask "what do we see, and what should we try?". Pair it with outcome measures such as customer satisfaction, defects found after release, and time from idea to delivery.

If a metric is used to reward or punish people, expect people to change the metric rather than the work. Keep team metrics inside the team.

Check your understanding

Quick check ยท 5 questions

of 5 answered You scored / 5 ยท perfect!

  1. 1. What does a team's velocity measure?

    Velocity is the amount of completed work (usually story points) per Sprint, counting only items that meet the Definition of Done.
  2. 2. A stakeholder adds 5 points of scope on the same day the team finishes 5 points. What does the burndown chart show for that day?

    Burndown only shows remaining work, so finished work and added work cancel each other out. A burnup chart shows scope and completion separately.
  3. 3. On a cumulative flow diagram, what does a steadily widening 'Doing' band suggest?

    The band's thickness is the WIP. A widening band means items are starting faster than they finish.
  4. 4. Which of these is the best use of velocity?

    Velocity is a planning aid for the team that produced it. Targets, comparisons and individual scores distort the numbers.
  5. 5. Throughput differs from velocity because it...

    Throughput is the number of items finished in a period, so it needs no point estimates. It is popular with Kanban teams.

Summary

  • Velocity is the points a team completed per Sprint. Use recent history and a range to forecast, never as a target.
  • Burndown shows work remaining against time, but hides scope changes; burnup shows scope and completed work separately.
  • A cumulative flow diagram shows where work piles up. A widening band signals growing WIP and a bottleneck.
  • Throughput counts finished items per period and needs no estimates.
  • Metrics are for the team's conversations. Used as targets, comparisons or individual scores, they get gamed and lose meaning.

Next, move from theory to tools with Getting Started with Jira.