Measuring platform ROI without vanity metrics: DORA + reliability + cost

An enterprise platform team needed to justify investment without relying on vanity metrics such as tickets closed, clusters managed, and pipelines created. This talk shares a 90-day approach that tied platform outcomes to three signals: DORA delivery performance, service reliability through SLOs and error budgets, and cost efficiency through unit cost per workload or service. It covers how the team established credible baselines, instrumented data across CI/CD, incidents, and AWS spend, and ran before-and-after experiments on specific platform capabilities. The session also highlights what did not work and the fixes that made reporting defensible.

Takeaways:

• A simple ROI scorecard template (DORA + SLO + unit cost).
• A baseline-after method that survives leadership scrutiny.
• A stop-list of vanity metrics and better replacements.
• Lessons learned from failed measurements and course corrections.