
Measuring code health success and DORA

AI Coding meets Code Health 🪄 — with Stuart Caborn
Topic 4 of 8
Published Apr 17, 2026 · Clip 22:34–29:27 (6:53)
Stuart describes how code quality was improved: no hard OKRs, but explanation, research data, friendly competition on leaderboards, and celebration of success. Quarterly sentiment and hard hotspot measures moved in agreement, building credibility. The company also uses DORA metrics, with deployment frequency as the productivity measure—growing 20-30% year on year—though he is careful to note confounding factors. He highlights recent research showing LLMs are more successful and spend fewer tokens on higher-quality code, making good code cheaper, faster, and more reliable for agents too.