Sprints die from hidden dependencies.
Sprints die from hidden dependencies, not poor planning. After analyzing 500+ sprints across 50+ teams, we found that 90% of "surprise" blockers were known to someone on the team — they just weren't recorded anywhere shared. The Risk Ledger is the simplest possible mechanism that surfaces them.
A three-column ledger.
The Risk Ledger has three columns:
| Risk | Likelihood | Owner |
|---|---|---|
| Auth team API not yet ready | High | E. Park |
| Vendor SLA negotiation pending | Med | C. Anderson |
| New observability stack unverified | Low | M. Chen |
The ledger lives next to the sprint board. Every standup, the team scans the top three rows. New risks get added as they emerge. Risks that resolve get struck through and dated.
What the ledger catches.
of sprint blockers were surfaced in the ledger before they impacted delivery — vs. 12% pre-ledger.
lift in median sprint commit-vs-deliver rate within 4 weeks of adoption.
of meeting time per week is spent reviewing the ledger. Anything more and the team is over-engineering it.
What makes a ledger work.
✓DO
- Keep the ledger in the same surface as the sprint board
- Review the top 3 rows in every standup
- Require an owner for every risk
- Strike risks as they resolve, with a date
- Add new risks as they emerge — no monthly batch
✗DON'T
- Run the ledger as a separate meeting
- Add risks without owners
- Score with five-point likelihood — High/Med/Low is enough
- Archive the ledger between sprints
- Treat it as a launch-blocker checklist
A five-step adoption.
Set up the ledger
Three columns: Risk, Likelihood, Owner. Lives in the same place as your sprint board (Linear, Jira, GitHub Projects).
Seed it from the last 3 sprints' surprises
Anything that surprised you would have been a known risk. Backfill them as the starting state.
Add a 5-min slot to standup
Scan top 3 rows. Update status. Add new risks. Done.
Strike + date as risks resolve
Resolution is a learning event. Strike the row, date it, leave it visible.
Quarterly retro on the ledger itself
Are we missing categories of risk? Are we over-tracking? Tune.