Six weeks into the Program Increment, the forecast board was still green. Then a message landed: the platform team’s auth migration — never listed as a blocker on this release’s dependency map — wouldn’t land until after the committed date, and half of what was left to ship needed it first.
Nobody had lied at PI planning. Nobody had skipped a box on the dependency board. At planning, this team’s confidence vote — SAFe’s fist of five — had landed solidly on four fingers, and that vote was not the weak link. It was expert engineering judgment, an honest read of everything the room could see that day. The problem was never the vote. The problem was everything that happened to the plan after the hands went down: the dependency simply hadn’t been the one anyone was watching, because it lived on a team’s backlog nobody on this release was tracking closely. By the time it surfaced in a status meeting, the slack that could have absorbed it was already gone.
The dependency you mapped is rarely the one that hurts you
Every team maps its known dependencies at planning. The two-team handoff everyone already expects gets reviewed carefully, because everyone is already watching it. Real risk lives somewhere else — in the dependency that looked minor, sat two teams removed, or got filed as a nice-to-have when it was actually load-bearing. That is not a discipline problem. It is what happens when a dependency map gets treated as a diagram drawn once, instead of a live picture of risk — and late dependencies are one of the three things that quietly erode a committed release date, the same way a confidence vote is a snapshot of what the room could see on planning day, not a guarantee of what’s still true in week six.
Hard link, soft link, critical path
Not every dependency deserves the same worry. A hard link means the dependent work literally cannot start until the blocking item is done — miss it, and the delay is direct. A soft link is a sequencing preference: convenient, not load-bearing. Teams routinely misclassify the two, usually because the person doing the classifying doesn’t own the blocking work and underestimates it.
Advanced Release Planning’s dependency-aware forecasting keeps that distinction honest. A Critical Path view shows the exact chain of work currently dictating your date. Cross-release context nodes surface blockers living on another team’s Fix Version instead of leaving them invisible until a status meeting. Cycle detection catches circular dependencies before they distort the picture. When a hard link on that path goes into negative slack, the forecast doesn’t wait for the retro to say so — the probable completion date moves the moment the slack does, measured in days, ranked on a risk heatmap so you know which dependency is worth a conversation this week.
The one that’s genuinely a cross-team problem
Scope creep and capacity changes are both dangerous, but at least they usually live inside a team that can see them coming. Dependencies are different. The risk sits on somebody else’s board, reported through somebody else’s standup, moving at a pace your own burn-up can’t show you. Ask a Scrum Master to name their own team’s biggest scope risk and they’ll usually get it right. Ask them to name the dependency quietly at risk on a team three desks over, and the honest answer is often: I don’t know it’s a problem yet. That is exactly why late dependencies are the erosion source most likely to surface in a status meeting instead of a planning conversation.
Keep the vote live, across every team’s backlog
This is what Advanced Release Planning does with a dependency map: it does not replace the judgment the room used to draw it, and it does not re-run the confidence vote with “better” math. It keeps watching the map after the hands go down — reclassifying risk as links change, tracking the critical path across every contributing team’s Fix Version, and re-forecasting your committed date’s confidence the moment a blocking chain moves.
Instead of a single typed-in release date, you get a committed date carried with its confidence — June 7 at 85%, say — riding on a band that shifts from green toward amber the moment a hard link on the critical path goes negative. The Critical Path view tells you, today, exactly which cross-team link is the reason that number might change. See how dependency-aware forecasting works in the docs, and bring it to your next cross-team sync while there is still slack left to use.
The dependency that blows up a release is almost never the one you flagged. See your critical path on your next Fix Version, free for 30 days. It Runs on Atlassian, so your Jira data — and every team’s — never leaves Atlassian.
1 Comment
Leave your reply.