Introduction to Part 11: Building a Metrics Programme
A metrics programme is a long-lived service of its own: it needs an audience, a small set of measures, honest tooling, the trust of the people measured, and a path to grow one step at a time.
Parts 2 to 10 describe what to measure. This part describes how to turn those measures into a habit that lasts. Many organizations treat measurement as a project: choose a tool, build a dashboard, announce a set of targets, and move on. A year later the dashboard is stale, the targets have bent behaviour in ways nobody intended, and the people who were measured have learned to distrust the whole exercise. A programme is different. It has owners, a purpose, a review cadence, and a way to retire measures that stopped helping. It is itself a service, and it can be measured and improved like one.
The agile aims apply to the programme as they apply to any other work. Continuous stakeholder involvement means the people who will be measured, and the people who will use the results, help to choose the measures from the start. Continuous improvement means the programme reviews itself and drops what does not earn its cost. Ongoing evolution means the set of measures changes as the services change. A sustainable pace means the programme does not add a reporting burden that crowds out the work it is meant to illuminate. And building around motivated individuals means that measures are offered as help to teams that own their work, not imposed as surveillance (see topic 1.2, Goodhart's law and the psychology of metrics).
Two quiet failures haunt every programme. The first is fear. If people believe the numbers will be used to rank, blame, or cut them, they will manage the numbers instead of the work, and the data will stop being true. The second is sprawl. Every stakeholder wants one more measure, every tool offers fifty more charts, and soon nobody can say which numbers matter. The topics in this part are mostly about avoiding these two failures: keep the set small, keep it owned by the teams, and keep the use of it visible and benign.
The four lenses shape a metrics programme. It should be built in the same continuous delivery spirit as the services it measures. AI features in tools and AI-derived data about people need explicit rules. Automation reduces collection effort and needs the same ownership and checks as any other automation. Analytics practice, with its data quality and uncertainty, is what makes the programme's output worth reading.
Topics in this part
- 11.1 Designing a service metrics dashboard: audiences, a small balanced set, context and trend, and views for teams, services, and leadership.
- 11.2 Tooling landscape: build versus buy, the kinds of platforms, data pipelines, and vendor-neutral selection criteria.
- 11.3 Rolling out metrics without breeding fear: transparency, team ownership, no individual league tables, and a rollout that earns trust.
- 11.4 Maturity model for metrics programmes: the shared five-level scale applied to the programme itself.
- 11.5 An incremental adoption roadmap: a plan for the first 30, 90, and 180 days that starts where you are.
How these topics interrelate
Rollout without fear (see topic 11.3, Rolling out metrics without breeding fear) comes first in importance, even though the dashboard is what people see. The agreements about who owns the data, who sees it, and what it will never be used for decide whether any dashboard will carry true numbers. The dashboard topic (see topic 11.1, Designing a service metrics dashboard) turns those agreements into a design: small sets, context, trend, and views suited to each audience. The tooling topic (see topic 11.2, Tooling landscape) then asks which platforms and pipelines can supply the data without becoming a project larger than the benefit.
The maturity model (see topic 11.4, Maturity model for metrics programmes) gives the programme a way to describe where it is and what the next step could be, using the same scale as every other topic. The roadmap (see topic 11.5, An incremental adoption roadmap) turns that into a plan of small steps, each of which delivers some value to the people measured before the next begins. This is the same continuous delivery logic that the rest of the book recommends for services: ship something useful early, learn from its use, and improve.
Several pairings keep the programme honest. A dashboard of results should be read with a record of how the data was collected and its known gaps. A rollout plan should be read with a promise about what the data will not be used for. A tooling choice should be read with the cost of leaving it. A maturity score should be treated as a conversation starter and never as a target.
The part draws on the foundations in Part 1, especially metrics governance (see topic 1.4, Metrics governance and ownership) and data sources (see topic 1.5, Data sources and instrumentation), and on the improvement and change measures of Part 9 (see topic 9.5, Continual improvement metrics, and topic 9.6, Organizational change and adoption metrics). The appendices in Part 12 supply the checklists, templates, and consolidated references that a programme team will use day to day.