In the post-mortem of almost every major operational failure, somewhere in the timeline, there is a point where the fix was cheap. A reroute that would have cost one delay. A supplier swap that would have cost a premium freight bill. A software rollback that would have cost one maintenance window. A quarantine that would have cost one batch.
That point closes silently. Nobody marks it on the timeline while it is happening. It only becomes visible in retrospect, when the cost of the decision that was eventually made is compared to the cost of the decision that was available 4 hours earlier.
The intervention window is not a concept. It is a number. And in most organisations, that number is never calculated until it is too late to use it.
How the window closes by sector
Aviation
Minutes
Heading correction to missed approach to diversion
Banking
Hours
Maintenance window to emergency response to regulator
Pharma
Days
Distribution centre quarantine to voluntary recall to mandatory
Manufacturing
Irreversible
Revenue lost when the line stops is not deferred – it is gone
Aviation: Minutes
The window
When a navigation anomaly starts building in a corridor, the window for a heading correction is measured in minutes. At minute 4, it is a cheap procedural call. At minute 12, the crew is in a missed approach. At minute 20, the aircraft is diverting. Each stage closes the previous option and opens a more expensive one.
The cost
A heading correction costs nothing measurable. A missed approach costs fuel, delay and schedule disruption. A diversion costs between £80,000 and £200,000 plus the network cascade into every flight behind it. In a corridor under active GPS interference, not one aircraft makes that journey. The entire sequence does, simultaneously. The cheap fix was available to every one of those flights, early, and was not taken because nobody knew the window was open.
Banking: Hours
The window
When a core banking system starts degrading, the window for a controlled maintenance intervention is measured in hours. At hour 2, it is a scheduled maintenance decision. At hour 6, it is an emergency response. At hour 12, the regulator is involved and the customer communication has already gone public. Each stage closes options and opens costs that the previous stage did not carry.
The cost
Barclays paid up to £12.5 million in compensation for three days of customer lockout. That figure does not include the internal remediation cost, the regulator engagement cost or the reputational cost priced into subsequent customer decisions. Every dashboard was green while the cascade was building. The window for a controlled fix opened and closed unobserved.
The £12.5 million is the cost of not knowing the window was there.
Pharma: Days
The window
When a cold chain breach is detected, the window for a voluntary quarantine is measured in days. A batch quarantined at the distribution centre costs one batch. A batch that reaches the hospital costs a voluntary recall. A batch that reaches the patient is no longer a logistics problem.
The cost
A voluntary recall averages $10 million. A mandated recall runs two to three times that. The production halt while the audit trail is reconstructed adds weeks of lost output. The cost difference between acting at the distribution centre and acting at the hospital is not 10x. It is the difference between a managed operational problem and a public safety event. The window between those two outcomes is measured in days.
Manufacturing: The Cost That Never Comes Back
The window
When a tier-one supplier begins failing, the window for alternative sourcing is measured in hours. At hour 2, a premium freight reroute is available. At hour 6, the air freight option remains but at four times the cost. At hour 12, the line stops. Once the line stops, the vehicles not built are not built later. They are simply not built. Revenue lost to a stopped production line is not deferred. It is gone.
The cost
Every hour of production stoppage costs between $1 million and $2 million in lost output. Customer penalty clauses add to that. The commercial consequences – broken delivery commitments, revised sales forecasts, downstream dealer inventory problems – add to that again. A delay in one part of the operation does not stay in that part. It moves through the whole company.
The Number Nobody Calculates
In every sector above, the intervention window is a specific number. It is calculable from the dependency structure and the incoming signal. It is not a probability or an estimate or a risk rating. It is a duration, with a cost attached to each stage of the cascade, so the decision to act is made with full knowledge of what waiting costs.
Most organisations do not calculate it. They manage by escalation. Something happens. It is escalated. A decision is made. By the time the process completes, the window has closed and the decision available is no longer the cheap one. The OCTA model calculates the intervention window before the cascade runs, so the cheap decision is still on the table when the information reaches the people who can act on it.
The working paper on the OCTA model is available here: jameslarsson.pro/research-octa
