It reached you.
Again.
Log the decisions that reached you but should not have. The same decision arriving three times is not bad luck — it is a missing rule, and this shows you which one.
Open the toolNot an inbox.
A map of missing authority.
Every senior technology leader resolves decisions that a well-designed system would have settled further down. One of them is an incident. The same type recurring is a Decision Architecture gap: authority that was never delegated, a rule that was never written, or a delegate who lacked the context to act. This tool records each escalation as it happens and groups them by type, so the pattern becomes visible before it becomes the operating model.
One escalation is an incident. The same decision arriving three times is the system telling you where a rule is missing.
The decisions that reach a senior technology leader are not random. They cluster — around the vendor exception nobody owns, the access request no policy covers, the release nobody else is allowed to sign off. Individually each one looks like a judgment call worth making. Logged and grouped, they stop looking like judgment calls at all.
Three minutes
per escalation.
Log it while it is fresh
What the decision was, when it escalated, who owned it, who escalated it, who resolved it. Facts, all of them checkable afterwards — which is what makes the count trustworthy months later.
Categorize it, once
Nine decision types, and the category is the grouping key. Picking the closest existing one rather than describing it afresh is what lets three related escalations register as three of the same thing instead of three unique events.
Add your read — optional, and never scored
Where it should have been resolved, why it came to you, and the rule you would write to stop it. Recorded beside the facts and shown in the patterns; never fed into any status.
Watch what repeats
A second escalation of a type is worth noticing. A third is a structural gap, and the tool says so — because by then the evidence is in your own log.
A count per decision type,
and nothing you did not record.
The instrument is deliberately thin. Escalations of one type are counted, and the count alone decides what the tool says — there is no weighting, no severity score, and no attempt to rank one decision against another.
Single
An incident. Recorded so a pattern can be seen if one forms. Nothing is concluded from one escalation, and the tool does not pretend otherwise.
Watch
Twice. Not yet a pattern, but no longer a one-off. Worth knowing about before the third arrives and the decision becomes part of your standing workload.
Structural gap
Three or more. Something in the decision architecture is missing: a rule that was never written, authority that was never delegated, or a delegate without the context to use it. Write the rule.
Your own diagnosis of each escalation — where it should have lived, why it reached you — is recorded and shown alongside the counts, and is never an input to them. The diagnosis is the conclusion this tool exists to help you reach, so it is not allowed to decide anything.
At the moment it happens, not at the end of the quarter. An escalation you reconstruct from memory weeks later loses exactly the details — who owned it, who escalated it — that make the pattern actionable rather than merely visible.
It is worth logging even the escalations you were right to receive. “It correctly reached the CTO” is one of the answers, and a log that only contains the frustrating ones will tell you the system is worse than it is.
Stop resolving the same decision.
Write the rule instead.
Log each decision that reaches you but should not have. Grouped by type, the ones that repeat name the rules your organization is missing — in your own evidence, not in principle.
Open the tool