Velocity Tracking for Multi-Team Programs

Multi-Team Velocity

In scaled agile frameworks (SAFe, LeSS, Nexus), multiple teams work on the same product. Tracking velocity across teams requires careful handling.

The Comparison Trap

The first rule of multi-team velocity: never directly compare team velocities. Each team has its own estimation scale, skills, and domain complexity.

Team Alpha:  Velocity 40 (front-end, simple stories)
Team Beta:   Velocity 25 (back-end, complex distributed systems)

Does Alpha produce more? Not necessarily. Their "5 points"
may equal Beta's "3 points" in actual effort.

Normalization Approach

If you must aggregate across teams (e.g., for program-level forecasting), normalize:

Team Alpha: 40 pts / 5 people = 8 pts/person
Team Beta:  25 pts / 4 people = 6.25 pts/person

BUT this is still misleading because point scales differ.

Better Alternatives

1. Use throughput (item count) for cross-team comparison

If all teams work from the same backlog with similar-sized stories, throughput is more comparable than velocity.

2. Track each team's velocity independently

Use each team's own velocity for their own sprint planning. Aggregate only the number of sprints needed (time), not points.

3. Normalize via reference stories

Have all teams estimate the same 3-5 reference stories. This creates a cross-team calibration factor.

Program-Level Forecasting

Remaining program backlog: 500 points (Team Alpha scale)
Team Alpha velocity: 40/sprint
Team Beta velocity:  25/sprint (but 1 Alpha point = 0.8 Beta points)
Adjusted Beta capacity: 25 / 0.8 = 31.25 Alpha-equivalent points

Combined capacity: 40 + 31.25 = 71.25 Alpha-equivalent points/sprint
Sprints needed: 500 / 71.25 = ~7 sprints

Key Principle

Velocity is a team-level metric. At the program level, prefer time-based metrics (sprints to completion) over point-based comparisons.