Six metrics instead of velocity
The Kite Method does not measure how much gets built, but how fast value lands and whether the line stays taut. Velocity and story points go. Six metrics take their place.
Intent-to-value
Lead time from Kite Brief to use in production.
Rework rate
Share of delivered items that come back within 30 days.
Review load
Items to check per person per day.
Zone mix
Distribution of work across Free, Guided, Paired and Held.
Escape rate
Production incidents, per zone.
Anchor health
Can people explain their work, and how is the team doing (short pulse survey).
Start with a baseline
Always record a baseline at the start. Thresholds differ per team and product, so calibrate them in the first phase. Also keep an eye on AI cost per delivered item, so speed does not quietly become expensive.
Pitfalls
The biggest pitfall is a human who formally holds the line but no longer really grips it. The table below helps you spot that and other pitfalls early.
| Pitfall | How you spot it | What you do |
|---|---|---|
| Rubber-stamping | Approvals get faster, review comments disappear | Fewer items per person, strengthen the Tail, run spot-check audits |
| Vague Frames | Rework rises, agents build the wrong thing | Make the Kite Brief mandatory: no Brief, no flight |
| Everything in Free | Escapes from the Free zone, incidents increase | Recalibrate zones, move borderline cases to a heavier zone |
| Everything in Held | Zone mix does not change, intent-to-value does not fall | Strengthen the Tail and deliberately test items one zone lighter |
| Juniors become bot-sitters | Juniors cannot explain what they deliver | Follow the junior learning path, enforce explain-back |
| Costs out of sight | AI spend grows faster than delivered value | Measure AI cost per item, set a budget per team |
| AI on a broken process | You mainly get faster at the wrong things | Map the flow from intent to value first, then accelerate |