Cycle commitment accuracy

Cycle commitment accuracy is the percentage of committed work a team finishes by the end of a cycle, measured against what it promised at the start. If a team commits to 30 story points and completes 27 of them to the definition of done, its commitment accuracy is 90 percent. The metric counts only the original commitment, so work added mid-cycle does not paper over a shortfall on what was promised.

The point of the metric is predictability rather than speed. A team that completes exactly what it commits, cycle after cycle, is one a business can plan around. Roadmaps become forecastable, dependencies become reliable, and stakeholders stop padding every estimate to protect themselves. A team that delivers wildly more or less than it commits is hard to plan with, even if it is busy and productive.

Over-delivery is a problem too, not just under-delivery. Consistently finishing far more than committed means the team is sandbagging its commitments, leaving capacity on the table that the business cannot see or plan for. The healthiest reading sits close to 100 percent in both directions: the team commits to what it can do, then does what it commits to.

The calculation compares completed committed work to the committed baseline. The discipline is entirely in defining the baseline and refusing to move it. Once the cycle starts, the commitment is fixed, and everything is measured against that fixed number.

1. 1

   Freeze the commitment at cycle start

   Record the exact scope the team commits to when planning ends. This baseline does not change during the cycle. Any work added later is logged as mid-cycle scope, kept separate from the committed total.

2. 2

   Track completion against the baseline only

   Count completed work that was part of the original commitment. A story that was committed and finished counts. A story added mid-cycle and finished does not count toward accuracy, though it is recorded elsewhere.

3. 3

   Apply the definition of done

   Only fully accepted work counts as completed. A committed story stuck in review at cycle end is not delivered, even if it is nearly there. Partial credit undermines the predictability the metric is meant to measure.

4. 4

   Express as a percentage and trend it

   Divide completed committed work by the committed baseline and multiply by 100. A single cycle is noisy, so trend the figure over six to eight cycles. The average and the spread together tell you how predictable the team really is.

Commitment accuracy is a downstream result of how a team plans and how protected its plan is once it starts. It drops when commitments are guesses, when capacity is overestimated, or when the cycle is constantly interrupted. A metric tree separates these causes so a low reading points to planning, capacity, or protection rather than a general feeling that the team overpromised.

Because over-delivery is as much a signal as under-delivery, the benchmark bands sit around 100 percent rather than climbing toward it. The target is a tight band centred on what was promised, held steadily across cycles. Mature delivery teams aim to land inside a narrow range nearly every time.

Better accuracy comes from honest planning and a protected cycle, not from pressure to hit a number. Commit to less than peak capacity, refine work before committing to it, and keep mid-cycle changes visible and deliberate.

1. 1

   Folding added work into the commitment

   Counting mid-cycle additions toward accuracy lets a team miss its promises while reporting a healthy number. Keep the committed baseline frozen and track added scope separately.

2. 2

   Chasing 100 percent at any cost

   Teams that must hit the target every cycle learn to commit to less than they can do. Treat accuracy as a signal to improve planning, never as a target tied to performance reviews, or it will be gamed.

3. 3

   Reading a single cycle as a trend

   One cycle is noisy. A single miss can be a one-off incident. Judge predictability on the average and spread over six to eight cycles, not on the most recent result.

4. 4

   Ignoring consistent over-delivery

   A team finishing far more than committed looks great on a chart but is hiding capacity the business cannot plan for. Sandbagging is a planning problem, just a quieter one than missing.