Writing · Operational intelligence

Perishable decisions

Some decisions lose their value the longer you wait. Those are the only ones worth building around.

Plot a decision's value against how long you take to make it. Some decisions barely move: you can sit on one for a quarter and lose nothing. Others fall off a cliff. Each line below is a decision, and the curve is how much of its value survives as the decision gets later. The steeper the curve, the more being slow is costing you.

I call the steep ones perishable, and they're where operational systems earn their keep: a perishable decision made on time is worth a multiple of the same decision made late. A durable one can wait. Most reporting treats them the same, which is exactly the mistake.

VALUE OF THE DECISION TIME UNTIL IT'S MADE → Durable decision Perishable decision ACT-BY WINDOW most of the value still on the table
Cost of delay. Steep curve = perishable; flat curve = durable. Figures illustrative.

Here's the organizational version, made concrete. At a homebuilder, a project's margin lands on a report roughly two months after the point you could still have acted on it. Everything between incurring a cost and reporting on it is a blind window: whatever goes wrong stays invisible until it's already locked in.

Say foundations are budgeted at about 30 units × ~$30k = $900k. Concrete runs 20% over: a $180k overage on a single line. If that only surfaces in the month-end margin report, it's already spent. Caught at 60% poured, roughly half (~$90k) is still recoverable. The decision to act was perishable; the reporting cadence let it rot.

THE BLIND WINDOW · ~2 MONTHS Cost incurred CONCRETE POURED 60% poured — ~$90k still recoverable Margin reported TOO LATE TO ACT
The blind window between incurring a cost and seeing it. Figures illustrative.

So the work is always the same shape. Find the decisions that cost money when they're slow. Fix how they get informed: shorten the blind window until the number arrives while you can still act on it. Prove it by the decision getting faster, not by the dashboard getting built. A project isn't done when it's built; it's done when the decision got faster.

That's the lens I bring to operations: not "what could we measure," but "which decisions are perishable, and how late are we making them."

← Back to writing tyler@tylerdove.me