Five gaps that show up everywhere.
After 85 teams ran the diagnostic, five gaps appeared again and again — and closing any single one moved the cadence. Closing all five doubled it. This article walks through the five, in order of impact.
The five gaps, ranked.
TBL · 01 · THE FIVE GAPS · MEDIAN CADENCE LIFTN = 85 TEAMS · 12 WEEKS
| # | Gap | Cadence Lift | Mechanism |
|---|---|---|---|
| 1 | No async decision log | +38% | Decision docs as a unit of work |
| 2 | PR review > 12 hr | +31% | Reviewer rotation + 2-hr SLA |
| 3 | No flag-by-default policy | +27% | All changes ship behind a flag |
| 4 | Sprints end mid-week | +21% | End sprints on Tuesday; deploy windows align |
| 5 | Retro skipped > 1 in 3 | +14% | Calendar lock + outcome doc |
What full adoption produces.
Cadence (deploys / wk)4×↑
median deploy frequency among teams that closed 4 of 5 gaps within 12 weeks.
Lead time (commit → prod)−54%↓
reduction in median lead time. Shorter lead time mostly came from PR + flag fixes.
Engineer NPS+22↑
lift in engineer satisfaction. Faster cycles aren't more pressure — they're less frustration.
What closes the gaps.
✓DO
- Adopt async decision logging within 2 weeks
- Set a 2-hour PR review SLA, enforced via reviewer rotation
- Enable flag-by-default in CI lints
- Move sprint boundaries to mid-week to avoid Friday deploys
- Lock retros on the calendar; require an outcome doc
✗DON'T
- Run any process change without measuring baseline first
- Try to fix all five gaps at once
- Layer 'process improvements' on top of broken cadence
- Treat retros as optional
- Skip the retrospective on the process itself
See your gaps in 5 minutes.
Run the diagnostic to see which of the five gaps your team has, and which to close first. Most teams have all five — but the order of closure matters. Start with #1 and #2 simultaneously; the rest follow.