1. Define velocity in plain language
Velocity is not just the number of tests shipped. It is how quickly ideas move from hypothesis to decision. Anchor your dashboard on three measures: throughput (tests shipped), cycle time (days from brief to decision), and decision quality (how many tests lead to a clear next step).
Write down what each measure means in one sentence. That definition keeps the dashboard honest and prevents endless metric creep.
2. Build a lightweight experiment log
You cannot measure velocity without a source of truth. Keep a simple experiment log in a spreadsheet or Airtable with these fields:
- Experiment name, owner, start date, end date.
- Status (planned, running, analyzed, shipped).
- Decision (ship, iterate, stop) and why.
This log becomes the data source for the dashboard. If it is not updated weekly, the dashboard will be wrong.
3. Visualize the three core metrics
Use simple visuals that show momentum at a glance:
- Throughput: bar chart of tests completed per sprint or month.
- Cycle time: line chart of median days from start to decision.
- Decision quality: stacked bar showing ship vs. iterate vs. stop.
Keep the time window to the last 6 to 8 sprints. Old history makes the dashboard feel stale.
4. Add a friction tracker
Velocity drops when teams wait on instrumentation, design, or legal reviews. Track friction explicitly with a simple list of blockers. Add a count of experiments delayed by each blocker category so leadership sees where to invest.
This is often more useful than debating whether win rate is high or low.
5. Make velocity reviews a habit
Review the dashboard in a predictable ritual, such as the first 10 minutes of the weekly growth sync. Ask three questions:
- Did throughput match our target?
- What slowed cycle time this week?
- Which decisions unlocked roadmap changes?
Document one action per review. Without actions, velocity dashboards become vanity metrics.
Use velocity to protect quality
Velocity is a balance. If throughput rises and decision quality drops, slow down to improve instrumentation or hypothesis quality. A good dashboard makes those tradeoffs explicit before you over-optimize for speed.