Metrics Workshop

Purpose

The Metrics Workshop helps teams move from intuition to evidence by selecting a small set of meaningful metrics, defining how they will be measured, and setting up a dashboard. Good metrics drive better decisions; bad metrics drive bad behaviour.

At a glance

How to run

1. Identify goals (10 min) — What questions do we need metrics to answer? What decisions will they inform?
2. Brainstorm candidate metrics (10 min) — Each person proposes metrics. Write each on a card.
3. Evaluate metrics (15 min) — For each candidate, ask: Is it actionable? Does it measure an outcome, not just activity? Can we collect the data? Can it be gamed?
4. Select 3–5 key metrics (10 min) — Choose the vital few. Include at least one leading indicator and one lagging indicator.
5. Define measurement (10 min) — For each metric: data source, collection method, frequency, owner, target.
6. Design the dashboard (5 min) — Sketch the dashboard layout. Decide the tool (Grafana, Datadog, spreadsheet).

Materials needed

• Team goals and OKRs
• Candidate metric cards
• Metric definition template (name, definition, source, frequency, target)
• Dashboard tool or sketch template

Pitfalls & common mistakes

• Too many metrics — More than 5–7 creates dashboard blindness. Focus on the vital few.
• Vanity metrics — Total sign-ups is a vanity metric; active users is actionable.
• No owner — Each metric needs a person responsible for keeping it accurate.
• Metrics drive bad behaviour — If you measure lines of code, people write verbose code. Choose metrics that align incentives.

Inclusion & accessibility

• Define metrics in plain language everyone can understand.
• Make the dashboard accessible (colour-blind friendly, screen-reader compatible).
• Include non-engineering team members in metric selection.

Variations

• GQM (Goal-Question-Metric) — Start from goals, derive questions, then select metrics.
• Pirate metrics (AARRR) — Acquisition, Activation, Retention, Revenue, Referral for product teams.
• DORA metrics — Deployment frequency, lead time, change failure rate, MTTR for engineering teams.

When NOT to use this

• If the team has no goals or strategy — define goals first, then metrics.
• If the data infrastructure does not exist — build instrumentation before the workshop.

Source & further reading

• Croll, A. & Yoskovitz, B. (2013). *Lean Analytics*. O'Reilly.
• Forsgren, N. et al. (2018). *Accelerate*. IT Revolution.