Being Right Too Late

Last week’s column made the case that information enables action but doesn’t compel it. The natural extension of that argument: even when action follows, when it follows determines whether it helps.

There’s a category of failure the industry rarely names. Not wrong decisions. Not missing information. Right decisions made after the window had closed.

The analysis was correct. The conclusion was sound. The action was the right one. But the change order came after the sub had already mobilized. The coordination issue got flagged after the pour. The schedule risk was named after the critical path had already shifted.

Being right is necessary. Being right when it still matters is the actual work.

Intelligence Has a Shelf Life

Military operations recognize this explicitly.

Intelligence products are time-stamped for a reason. The same assessment that enables a decisive advantage at the right moment is often just confirmation — or worse, a distraction — once that moment has passed. The value of intelligence decays. Sometimes slowly, sometimes instantly. A commander who acts on D-2 intel at D+1 may find that the conditions named in the assessment no longer reflect the situation on the ground.

Timekeeping isn’t a formality in that environment. The window of maximum leverage is a real operational concept — the interval during which a specific action produces the greatest effect relative to its cost. Act before it opens and you’re premature. Act after it closes and you’re paying full cost for partial benefit. The work of intelligence isn’t just to surface the right assessment. It’s to surface it inside the window.

That framing changes how decisions get evaluated. A decision isn’t just right or wrong. It’s right-and-timely, right-and-late, wrong-and-timely, or wrong-and-late. Each of those produces a different outcome. Being right doesn’t guarantee a good one.

How Construction Measures Decisions

Construction often measures decision quality by outcome. Was the call right? Did it hold up under scrutiny? Did the team that made it look good in retrospect?

What rarely gets measured is decision timing. Whether the call came inside the window of maximum leverage goes mostly untracked. A change order approved the day after a pour is often evaluated on the same axis as one approved the week before — was it valid? — rather than on the axis that actually mattered, which is whether it was still able to prevent the cost it was meant to prevent.

That measurement gap has consequences. It means the industry doesn’t easily learn from timing failures. When a project absorbs a cost that a correct decision failed to prevent, the postmortem tends to ask whether the decision was right rather than whether the detection system surfaced it in time. The lesson stays invisible.

The Correction Cost Curve

The reason timing matters so much in construction is that the cost of the same action changes as the project progresses.

An issue identified in design requires a drawing revision. The same issue identified during construction requires a change order, potential rework, and ripple effects on the schedule. Identified after closeout, it often requires litigation. The issue doesn’t change. The cost of addressing it increases by an order of magnitude at each stage.

That cost curve means the window of maximum leverage closes early. Most project issues have a point — often in design or early construction — at which they can be addressed with minimal disruption. Past that point, the disruption cost starts compounding. The action is still correct. The leverage is already gone.

Detection infrastructure determines which side of that curve an issue surfaces on. A team that finds a coordination conflict in the model before mobilization and a team that finds the same conflict during installation both found it. One found it in time. The window of maximum leverage was already closed for the other before anyone knew it existed.

Why Late Detection Is Structural

Most late detection in construction isn’t a product of inattention. It’s a product of where the detection infrastructure is pointed.

The default detection loop in construction is reactive. Issues surface when a trade encounters them during work. That means the detection event tends to occur at the same moment as the impact event — when the crew arrives and finds the conditions aren’t what they needed. By the time the issue is documented, escalated, and reaches someone with the authority to address it, the window has already closed.

The teams pulling ahead have built detection infrastructure that runs ahead of execution. Constraint logging that surfaces issues before the activity starts rather than during. Model-based coordination that catches conflicts before mobilization. Lookahead reviews that identify missing conditions with enough lead time to close them. The point in each case is the same: surface the issue when something can still be done, not when the crew is already standing in front of it.

What the Teams Pulling Ahead Are Doing

Three patterns keep showing up.

They track when issues surface relative to the decision window. Not just whether a problem was identified. Whether it was identified inside the window where the lowest-cost response was still available. Teams that track this metric tend to improve it — because making detection timing visible makes it manageable.

They build detection infrastructure that finds problems earlier. The architecture of when issues surface is a design choice, not a given. Lookaheads, constraint logs, model coordination, and pre-activity verification rhythms all move detection upstream. The goal in each case is the same: earlier signal, more leverage, lower cost.

They treat late-but-correct as a failure mode worth measuring. When a right decision arrived after the window had closed, the question they ask is why the detection was late — not just whether the analysis was correct. That question points at the system rather than the individual, which is where the fix lives.

That’s not an analysis problem at the individual level. It’s an infrastructure one — the same shift this series keeps returning to. Build the system that surfaces issues inside the decision window, and being right starts to matter at the right moment.

Being right too late is a distinct failure mode from being wrong. The industry often tracks one of them.

---

Field Notes No. 07 — Information Enables Action. It Doesn’t Compel It. — covered the gap between clean information and action. No. 09 lands next Wednesday.