Engineering Effectiveness

DORA Metrics Should Be Signals, Not Targets

2026-08-05 · 2 min

DORA metrics are useful because they make parts of the engineering system observable: how often value ships, how long change takes, how often releases hurt, and how quickly we recover.

They become harmful when leaders turn them into personal scorecards or vanity targets divorced from customer outcomes.

Goodhart’s law applies quickly. Teams can raise deployment frequency by shipping smaller meaningless changes, or improve lead time by narrowing the measured window.

The leadership move is to treat each signal as a question. What constraint is this pointing at — clarification, review, test environments, architecture coupling, release policy, ownership?

Pair the metrics with narrative: incidents, escaped defects, developer experience and product outcomes. Numbers without story create false confidence.

Measure systems to improve systems. The moment metrics become surveillance, they stop telling the truth.

Related: Engineering Playbook