Using Velocity for Release Planning and Forecasting

Velocity-Based Release Planning

Release planning translates velocity into dates. If you know how fast you are going and how far you need to travel, you can estimate arrival time.

The Basic Formula

Remaining backlog:     200 story points
Average velocity:       25 points/sprint
Sprint length:          2 weeks
───────────────────────────────────
Sprints needed:         200 / 25 = 8 sprints
Estimated duration:     8 x 2 = 16 weeks

Adding Confidence Ranges

Use standard deviation to create optimistic and conservative estimates:

Velocity:  25 +/- 5 (std dev)
Optimistic (30/sprint):  200 / 30 = 6.7 sprints  →  ~14 weeks
Expected  (25/sprint):   200 / 25 = 8.0 sprints   →  16 weeks
Conservative (20/sprint): 200 / 20 = 10 sprints   →  20 weeks

Present all three to stakeholders. The conservative number is typically the one to commit to externally.

Accounting for Unknowns

• Buffer factor: Add 10-20% to the backlog for undiscovered work.
• Team availability: Account for holidays, on-call rotations, and PTO.
• Dependency delays: External dependencies can block entire sprints.

Updating the Forecast

After each sprint, recalculate with updated velocity and remaining backlog. This creates a burn-up chart that converges over time.

Best Practice

Never commit to a specific date based on a single sprint's velocity. Always use at least 3-5 sprints of data and present a range, not a point estimate.