Terms  /  Measurement  /  Iteration Velocity
09 · Measurement

Iteration Velocity

The speed at which a human-AI pair can move from idea to tested output.

Iteration Velocity measures end-to-end cycle time: from the moment the user has an idea to the moment that idea is implemented, tested, and verified. High Iteration Velocity means the human-AI pair can try ideas quickly, learn from results, and iterate rapidly. Low Iteration Velocity means each cycle is slow — bogged down by re-onboarding, debugging agent errors, or waiting for retries. Iteration Velocity is the compound metric that reflects everything else working well: good Context Continuity, low Reasoning Tax, high Execution Fidelity, and efficient Selective Recall.

Example
Two developers working on similar features. Developer A (stateless system): idea → 10 min re-onboarding → 5 min coding → 3 min debugging agent error → 5 min re-explaining the fix → 5 min testing = 28 minutes per iteration. Developer B (stateful system with Structured Context): idea → 30 sec context recall → 4 min coding → 2 min testing = 6.5 minutes per iteration. Developer B iterates 4x faster. Over a day of work, that's the difference between trying 8 ideas and trying 30. Iteration Velocity isn't just about speed — it's about how many experiments you can run.

More in Measurement