The DBS Bank CEO lost $3 million of his salary because DBS failed so many times the regulator made it personal. By the time the salary clawback happened, the incidents were already history. The failures had already run. The customers had already been locked out. The board had already been briefed, after the fact, by teams who themselves had been briefed after the window to act had closed.
This is the standard failure sequence in every major operational incident: the event starts at the operational layer, builds through the management layer, and arrives at the leadership layer as a fully formed crisis. At each transition, some information is lost, some time passes, and the options available shrink. By the time the CEO is in the room with the full picture, the decision that would have cost £10,000 is no longer on the table. The decision that costs £10 million is.
That is not a communications failure. It is a structural one. And it is the same structure in every organisation that has ever written a post-mortem.
The decision that would have cost £10,000 is no longer on the table. The decision that costs £10 million is.
The Gap Between the Dashboard and the Boardroom
What the dashboard shows
When the Barclays mainframe started degrading, every dashboard showed green. The monitoring systems were working. The alerts were calibrated. The teams watching them had no reason to escalate, because the picture they had showed normal. By the time the picture no longer showed normal, 56% of online payments had already failed and three million customers were locked out. The escalation arrived at the leadership layer as a crisis, not as an early signal, because the early signal never made it out of the operational layer.
56%
of Barclays online payments had already failed before the escalation reached leadership
What the board sees
A board receives information that has been filtered, aggregated and formatted for the level it sits at. That process takes time, and it removes the detail that would tell the board the decision is time-critical. An operational team escalating a degrading mainframe does not escalate with a countdown clock. It escalates with a status report. The board receives the status report and convenes to discuss it. By the time the meeting ends, the countdown clock has finished.
The structural gap
The gap is not between what the CEO knows and what the operations team knows. It is between when the signal arrives and when the leadership picture is updated. Those two things are never synchronised, because the process that translates operational signals into leadership-level decisions was not designed for time-critical information. It was designed for governance. Governance is thorough. It is not fast.
The Decisions That Were Never Made
Aviation
When a navigation anomaly starts building in an affected corridor, the decision to reroute traffic is a straightforward operational call. It costs one delay. It saves the cascade. That decision is available to the operations centre, to the airline dispatch team and, if the signal is visible early enough, to the crew. It requires no board approval. It requires knowing that the anomaly is building, that it is going to cascade, and how long before the cascade point arrives. Without that, the decision is not made. Not because nobody was in charge. Because nobody had the information in time.
Banking
The ECB’s T2 settlement system stopped three trillion euros of daily payments for seven hours. At no point during those seven hours did any leader in any eurozone finance ministry, central bank or major commercial bank have a single picture of how much of the eurozone economy had stopped and when it would restart. Each institution was managing its own exposure. The combined picture did not exist. It was assembled after the fact, for the post-mortem.
€3 trillion
of daily eurozone payments halted for 7 hours with no combined leadership picture
Manufacturing
A production line stops when the component does not arrive. The decision to source from an alternative supplier, to accept a premium logistics cost, to delay a downstream customer delivery rather than stop the line entirely – all of these decisions exist. They are available, at different costs, at different points in the cascade. What closes them is not the absence of authority. It is the absence of a time dimension. Nobody knows they are in hour 2 of a 6-hour window because nobody calculated what the window was.
What Leadership Actually Needs
Leadership does not need more data. It needs fewer decisions that arrive as emergencies. The difference between a decision that arrives in time to be made cheaply and a decision that arrives as a crisis is not the quality of the people involved. It is whether the organisation can calculate how far along the cascade is and how long before the window closes, at the point when acting is still the less expensive option.
A CEO who knows the window has 4 hours left can make a different decision than a CEO who is told the system has already failed. A board that receives a probability-weighted outcome with a time attached to it is in a different position than a board that receives a status report. The OCTA model produces the first kind of information. Most organisations are structured to produce the second.
The structural gap between operational signals and leadership visibility is closeable. It requires mapping the dependency structure, calculating the cascade before it runs, and surfacing the intervention window to the people who have the authority to act on it while it is still open. That is not a governance reform. It is a modelling problem.
The working paper on the OCTA model and its mathematical basis is available here: jameslarsson.pro/research-octa
