FM-02: Escalation Inversion
A pattern. The escalation path is open. Walking it is the risk.
Justin R. Greenbaum · The Lexicon · June 2026
A site reliability engineer knows a dependency is fragile and will fail under peak load, but the last person who raised a “theoretical” risk spent two sprints defending it in reviews, so he files it as a backlog ticket nobody prioritizes. A compliance analyst sees a gap that belongs in front of the steering committee, but surfacing it means owning a remediation she has no budget for, so she notes it in a memo and moves on. A regional manager has flagged the same staffing shortfall three quarters running, and each time it came back as a question about her forecasting, so this quarter she stops flagging and absorbs the overtime herself. A product lead can see a launch date slip becoming inevitable six weeks out, but the meeting where that gets said rewards confidence, so the slip surfaces at week one instead, when it is a crisis rather than a heads-up.
None of these people are failing. Each one has correctly priced what escalation costs and decided to carry the problem instead.
This has a name. It is Escalation Inversion.
Here’s the pattern. The escalation path exists on paper, but using it is slow, costly, or reputationally risky, so problems get absorbed at the edge until they either resolve quietly or detonate in public. Escalation volumes stay low. The formal process is documented and clear, while the informal rule everyone has learned is that raising a problem makes you the problem. Leadership reads the low volume as things being under control, which is the exact opposite of what the low volume means.
Silence masquerades as stability. The team with the fewest escalations is called high-performing. The manager who never brings problems up is called low-maintenance. The quarter with no red flags is called clean. The engineer who handles it without noise is the one who gets promoted. The signal the system reads is stability. The condition underneath is a team that has learned raising the problem costs more than carrying it.
What Escalation Inversion gets mistaken for is what makes it durable.
Empowered teams. Strong local ownership. A mature, low-noise culture. Only edge cases escalate. Each is a hero narrative, and each is how the system rewards the behavior that is blinding it. Praising the quiet team removes any reason to ask what it is absorbing. Treating escalation as noise teaches people that signal is unwelcome. Reading a clean dashboard as a healthy operation confuses the absence of reports with the absence of problems.
Hero narratives are how Escalation Inversion survives contact with the risk review.
The pattern recurs and changes costumes. In one organization it shows up as an incident postmortem asking why nobody raised the risk, when three people had raised it informally and watched it go nowhere. In another, as a quarterly risk register that stays green right up to the failure. In a third, as a “no surprises” culture so loud that a surprise becomes the only thing that ever gets through.
The conditions are structural, not behavioral. This is why telling people to escalate more does not interrupt it; it raises the ask without lowering the cost. Adding an escalation channel does not interrupt it; it adds a path without adding a reason to walk it. Replacing the manager resets the relationship and leaves the incentive intact.
What interrupts it is structural. Strip the cost, so raising a problem is reputationally and metrically neutral. Reward the early signal, so the person who flags the risk before the failure is the one who is valued, not the one who absorbed it. Measure what did not escalate, the distance between what the edge knew and what reached the top. Where the incentive cannot be changed quickly, the cleanest move is for a named owner to go ask for the problems directly, and to make the asking safe.
When Escalation Inversion has a name, the options change.
The engineer who has been filing fragile-dependency risks as backlog tickets stops reading his own quiet as good judgment and sees it as a cost the system is charging him to stay safe. The manager above him stops counting low escalation volume as a healthy team and starts asking what the team has decided not to send up. The executive sees that the clean risk review and a low-risk operation are not the same reading, and that one has been standing in for the other. The board sees that the absence of escalations is not evidence the system is sound; it can be evidence the system has gone quiet.
Naming does not fix. Naming changes what can be seen. What can be seen is what can be acted on.
If any of this feels familiar, it has a name and a taxonomy.
The canonical definition of FM-02, including its early warning signals, common misdiagnoses, and recovery conditions, is at dripractice.com/fm/fm-02.
A role-specific view of how the same pattern looks from the infrastructure seat is at dripractice.com/lens/it-and-infrastructure.
A five-minute diagnostic that runs entirely on your device and never leaves it is at dripractice.com/diagnose.
Next in The Lexicon: FM-11, Metric Authority Drift. When the edge goes quiet, leadership leans harder on the dashboard. FM-11 is what happens when the dashboard stops informing the decision and starts being it.
Subscribe to The Lexicon.