How Betterplan forecasts
“How long will it take, and what will it cost?” is the question that comes up again and again in every project. The answer is usually inadequate and too laborious. Betterplan assumes two things are enough to make it easier: a backlog that is estimated at all times, and a team that works in iterations. Both are hard to sustain by hand. Betterplan takes that work off you. Stories without an estimate of their own automatically get a provisional value, the median of all estimates (see Automatic estimates), so you don't have to estimate everything, and when you complete an iteration it records a snapshot. This page explains how the forecasts in Progress, the Delivery Timeline and the Estimation Widget arise from that.
The ingredients
| Ingredient | Where it comes from |
|---|---|
| Scope | Sum of the story points of all user stories (see Estimating: the basics) |
| Done | Points of the delivered stories |
| Velocity | Average of the points delivered by the last three completed iterations. Until then, the value from the settings applies |
| Scope growth | By how many points the scope grows per iteration on average, over the same three iterations |
| Iteration length | From the project settings, Team tab |
There is a reason the velocity is an average of the last three iterations: older iterations say less about your team today. The help text in the application still mentions a median in one place; the calculation uses the average.
From velocity to a date
Progress subtracts the growth from the velocity: the team delivers the velocity per iteration, while the scope grows by the growth. The net pace is the difference. The remaining points divided by the net pace, rounded up, give the number of iterations to the goal. The ETA lies that many iteration lengths beyond today's state.
In the QuickCart example: 283 points of scope, 171 done, so 112 remaining points. At a velocity of 21 and a stable scope that is 112 / 21, rounded up 6 iterations. With two-week iterations the ETA is therefore 12 weeks from today (January 8, 2027 in the example data). Scope growth lowers the net pace and pushes the date out. If the scope grows as fast as the velocity or faster, Betterplan shows no forecast but the notice that scope grows faster than delivery. The chart draws at most six forecast points even if the goal is further away: the line states a direction, not a promise about iteration 40.
From velocity to the timeline
The Delivery Timeline calculates in finer detail: a story's duration is its estimate divided by the velocity, times the iteration length. Stories are scheduled by global priority: whatever is furthest along comes first, and within a state the order in the Backlog decides. As many run at once as the team worked on in parallel on average in past iterations. The order in the Storyboard plays no part. Both forecasts rest on the same numbers and should therefore agree.
And the cost?
With budget and cost per iteration (project settings, Team tab) the forecast can be given a price. The Estimation Widget shows the remaining amount, and the timeline can show the end of the budget as a marker.
Forecast for a section
With a filter, Progress and the Estimation Widget recalculate scope, done and remaining points for the visible Stories only, for example for one Release. The velocity stays that of the whole project, so both point out that the estimate may be too optimistic. The Delivery Timeline ignores filters and always shows the whole project.
How reliable is the forecast?
A forecast is only as good as its foundation:
- Few iterations: With fewer than three completed iterations, the velocity has few measurements, and Progress warns you.
- Thin estimates: Stories without an estimate of their own count with the median of the existing estimates. From experience it becomes dependable from around 30 estimated Stories, and the more real estimates there are, the less the number hangs on it. So estimate Stories as they mature, at the latest on the way to Ready (see Automatic estimates). Rapid Estimation extrapolates roughly for the backlog before detail work.
- Spread: Whether a date holds also depends on how evenly stories get finished. The flow metrics show that with the 85th percentile.
Read the forecast as a range, not as a date to the day. The more often you complete iterations cleanly, the better it gets.