I’ve lost count of the number of times I’ve seen a program that looked fine on paper, felt uncomfortable in practice and then suddenly tipped into crisis. Status stayed green. Plans were intact. Everyone was busy. And yet, if you were close enough to the work, something didn’t add up.
It took me a while to understand why that pattern repeats so reliably. It isn’t because teams are careless or leaders aren’t paying attention. It’s because most delivery governance is built around expectation, while failure emerges in reality — and the gap between the two grows long before it shows up in a report.
What changed for me was not better planning or tighter controls. I’ve learned to watch for these signals because I’ve been the one explaining late surprises, and I don’t want to be in that position again. It was learning what to notice once work was in motion.
Why aligned intent still leads to late surprises
In most organizations I’ve worked in, intent is rarely the problem. The goal is clear. The urgency is shared. Accountability is delegated to capable people who want to do the right thing.
Where things start to drift is in execution.
As work moves across teams, approvals and systems, interpretation and sequencing diverge. Decisions are made locally to keep things moving. Small compromises are absorbed. None of this feels dramatic in the moment. In fact, it often feels responsible.
At the same time, status reporting becomes a blunt instrument. Green is safe. Amber invites questions. Red triggers escalation. That dynamic is rarely explicit, but everyone understands it.
The result is not deception. It’s smoothing.
Teams hold risk a little longer than they should. Uncertainty is translated into confidence. Issues are framed as temporary. By the time status changes, options have already narrowed.
That’s the core problem I’ve learned to look for: not misaligned intent, but misaligned coherence. Everyone is aiming in the same direction, but the system is quietly compensating in ways that status reports are structurally bad at revealing. (For a deeper look at how intelligent organizations can still produce fragile outcomes, see my analysis of intelligence, wisdom and decision-making in enterprise systems and AI.)
The 3 signals I now pay attention to instead of status
I don’t think status reports are useless. I think they’re incomplete.
What I’ve found more helpful is paying attention to a small set of simple signals that show how delivery is actually behaving under load. I don’t worry too much about the precise numbers. I watch the direction and the persistence.
These aren’t metrics to manage people. They’re signals that tell me where to look and what questions to ask.
1. Work starts to wait
The earliest sign, for me, is almost always queues. Work waiting between teams. Tickets sitting “ready” but untouched. Lead times stretching without an obvious blocker.
When this shows up, it usually means demand has outpaced capacity somewhere specific. Not in the abstract, but at a boundary that matters. This mirrors what Goldratt called the constraint in his Theory of Constraints: the system’s throughput is limited by its weakest point, and work accumulates there. No amount of optimism in a plan changes that.
The question I’ve learned to ask at this point is simple: Where is work waiting and why there?
Not who’s slow. Not who needs to try harder. Just where the system is telling us it can’t keep up.
2. Work comes back
The next signal is rework. Items that were “done” returning for clarification. Late changes because something was misunderstood. Downstream teams compensating for gaps upstream.
This doesn’t usually mean people aren’t capable. It usually means assumptions were left implicit for too long.
When I see rework increasing, I stop asking how fast we’re moving and start asking: what did we think was clear that isn’t? Often, the answer sits at an interface or a decision boundary that no one owns cleanly. A pattern that mirrors Conway’s Law: Organizations design systems that reflect their communication structures.
3. Capacity is quietly borrowed
The most subtle signal is unplanned capacity burn. The heroics. The “just this once” reprioritization. The quiet shifting of effort to keep the most visible work moving.
In the short term, this looks like commitment. In the medium term, it’s a warning sign.
When teams consistently borrow capacity to protect today’s delivery, they’re usually consuming tomorrow’s resilience. Other work slips. Recovery options shrink.
The question I’ve learned to ask here is uncomfortable but important: What are we stabilising now at the expense of later?
What matters with all three signals is not the absolute value. It’s the trend. If they show up occasionally and resolve, that’s normal. If they persist, the system is telling you something status never will.
How this changes the governance conversation
Once I started paying attention to these signals, my conversations with leaders changed.
Instead of debating whether a plan was still credible, we talked about where strain was accumulating. Instead of asking teams to justify status, we focused on where intervention might actually help. In practice, the strain almost always shows up where responsibility or meaning crosses teams.
This doesn’t remove risk. It doesn’t replace controls, assurance or mitigation planning. It supports them.
Feedback and risk mitigation are not the same thing. Feedback tells you where risk is emerging. Mitigation is what you choose to do about it. The value of early feedback is that it makes those choices deliberate rather than reactive.
It also reduces the incentive to game status. When the signal you’re looking for is strain rather than confidence, surfacing issues early becomes safer, not riskier.
I’ve found this particularly useful in complex, delegated environments, where the CIO or executive sponsor isn’t making every decision, but is accountable for the conditions under which decisions are made.
Since drafting this piece, I’ve had conversations with practitioners across enterprise IT who’ve recogn ized these patterns immediately. One made a sharp distinction between “erosion” (loud, visible failures everyone sees) and “ossification” (quiet drift hidden behind green metrics while boundaries survive too long, looking like stability while quietly killing the ability to adapt).
Teams will fight to defend a boundary that was designed for conditions three years ago. The governance around it still works. The reports still look clean. But underneath, people route around it. Coordination overhead grows. Shadow processes emerge because the boundary no longer fits what the organization actually needs. And the feedback signals I describe — rework, latency, dependency expansion — are all present. They’re just being read as performance problems instead of structural ones.
The three signals catch both failure modes. Erosion tends to be loud — escalations, rework, visible strain. It shows up in stand-ups and incident reports. Ossification seems to hide behind metrics that still look green, but the tell is usually in what isn’t being measured: How much work crosses the boundary informally (spreadsheet proliferation, side conversations), how long decisions take when they should be local and how often people say “we can’t do that because of this process” for things that used to be straightforward.
There are limits to this approach, and it’s important to be clear about them.
These signals do not predict outcomes. They do not eliminate uncertainty. And the absence of visible strain does not mean there is no risk.
External shocks still happen. Some failures are discontinuous. Judgment still matters.
What they offer is earlier visibility of the gap between expectation and reality, while there is still room to act.
Plans express intent. Status expresses confidence. Delivery expresses constraint.
The sooner you can see where those diverge, the more options you preserve.
That is what I have learned to watch for when delivery really matters.
This article is published as part of the Foundry Expert Contributor Network.
Want to join?
Read More from This Article: Why delivery drift shows up too late, and what I watch instead
Source: News