A programme goes amber in week nine. It reaches the steering committee in week fourteen, red, with a question attached: why wasn’t this flagged earlier?
The engagement survey has an answer ready: low psychological safety. The room agrees quickly. It’s a diagnosis that implicates nobody in it. A workshop is booked, and leaders are asked to model vulnerability.
Trace the route a risk actually takes. The delivery lead sees it first and tells their manager, who owns the delivery date and whose bonus sits on it. So the call on whether the programme is in trouble sits with the one person paid not to make it.
The manager can escalate, which puts a named exposure against their own record, or hold it a fortnight and see if it resolves. Nobody above them has ever said, in advance, that an early amber will be backed. Holding is free. Escalating is expensive every time. There’s a risk function on the chart, but it can’t stop anything, so it has never been worth telling.
None of this is about courage. Everyone in that chain read the incentives correctly. Silence is what the design pays for, and they were paid accurately.
Run the workshop and some people will speak up for a while. The structure is still there when it ends, and so is the price of being the first person through it.
When the same behaviour shows up reliably across different people, the people stop being the interesting variable.
This is the second-last Signal. The final one, in December, goes back to where the series started.
— Signal / unfiltered observations from the field.
#EnterpriseArchitecture #DecisionRights #OrganisationalDesign #RiskGovernance #OrdoAnimi