Most engineering orgs have too many metrics and too little clarity. A dashboard with forty charts tells you nothing at a glance — which means no one glances.
The teams that run well do the opposite: 5–15 numbers, reviewed every week, each with an owner and a target. Not a dashboard you visit when something already feels wrong - a scorecard you read every week whether things feel fine or not. Each number is green or it isn't. If it's red, it becomes an issue to solve, not a chart to admire.
A starting set for an engineering org:
Delivery vs. commitment - did we ship what we said this week?
Reliability - SLO attainment / open SEV1–2 incidents
Cycle time - idea to production
On-call load - pages per week
One live risk number - security posture, tech-debt, or key-person risk
The discipline matters more than the exact list: same numbers, same time, every week. Consistency is what turns a scorecard into an early-warning system - you see the reliability number drift in week one, not at the quarterly review when it's already a crisis.
And it makes reporting up trivial. When your CEO asks how engineering is doing, you're not scrambling - you're summarizing what you already run: "four of five green; cycle time slipped, here's why, here's the fix."
Pick your five. Put an owner on each. Review them Monday. That's the whole trick.
The Engineering Leadership Operating Cadence kit gives you the scorecard, a quarterly-priorities tracker, an issues discipline, and a weekly meeting agenda - the full operating rhythm. It's at ctofieldkit.com.