Every major failure begins with someone saying, "We've known about that for years."
The Tempered Signal Sunday Edition June 21
"A deal that died and an exchange that went dark this week share the same failure..."
When a risk belongs to everyone, it is owned by no one, and the bill arrives on a schedule no one tracked.
Signal Score: 8.6/10 — Two systems failed the same way: a known risk with no named owner reached its expiry date before leadership did.
OPENING SIGNAL
A deal that died and an exchange that went dark this week share the same failure. A known risk was visible to anyone with access, assigned to the system in general, and therefore owned by no one in particular. When it expired, it billed.
WHY THIS, WHY NOW
The pattern that surfaced this week
On Tuesday, Estee Lauder and Puig confirmed they had terminated a 40 billion dollar combination over a change-of-control clause on Charlotte Tilbury, embedded in a 2020 agreement both sides could read, that surfaced as a deal-ending liability only after two months of momentum had already compounded.
On Thursday, Coinbase disclosed that its matching engine had been running in a single AWS availability zone by design for years. When a chiller failed in Northern Virginia, the venue went dark for seven hours. The architectural tradeoff was not hidden. Every senior engineer on the platform knew it. The decision to carry it had simply never migrated to the level authorized to re-price it.
Why it can’t wait
Both organizations are paying now for decisions that were rational when made and never revisited when the stakes changed.
The clause was legal and signed in 2020. The single-zone architecture was a legitimate latency optimization. Neither was wrong at the time. Both became wrong when the context shifted, and neither had a system that required anyone to notice. The cost of another week without this clarity is that the standing tradeoffs in your own organization continue aging toward the same expiry date, invisible by design.
What shifts if leaders see this clearly
The behavior that changes is the review cadence on standing decisions. Not the decisions themselves. The mechanism that ensures a tradeoff made at one set of stakes gets re-priced when the stakes change. That mechanism does not exist by default. It has to be built deliberately, and it can be built before Friday.
THE DEEP DIVE
The Tilbury clause and the Coinbase matching engine are not two stories. They are one pattern at two scales.
In the Estee Lauder case, the change-of-control provision was written into the Puig-Tilbury agreement in 2020.
It was not ambiguous.
It was not buried.
It governed what would happen to Charlotte Tilbury’s valuation if Puig entered a combination. Both sides to the March 2026 talks could read it. People close to the discussions told Reuters it was the single largest barrier to closing.
What it was not was owned. No individual was accountable for pricing that clause before strategic appetite accelerated past it.
By the time accountability concentrated, the only available decision was to walk. Estee shares rose 11.5% on the termination. Puig fell 15%. The market priced the outcome correctly. The deal teams had not priced the trigger.
In the Coinbase case, the matching engine ran in a single availability zone for co-location latency. That decision made sense when the venue was a spot exchange competing on microsecond execution. It made a different kind of sense when the venue became a CFTC-regulated derivatives exchange whose outage would require a regulatory disclosure.
Most of Coinbase’s other systems were multi-AZ. The one that generates the revenue was not.
Every senior engineer on the platform knew the tradeoff. The decision to carry it simply lived inside engineering, reviewed on an engineering cadence, and never migrated to risk, finance, or the board when the regulatory surface area expanded.
Armstrong made a clean public statement within hours of the outage. The honest load is the next 90 days, when the press cycle ends, earnings recover, the roadmap fills, and re-architecting the engine costs engineering quarters, breaks co-location contracts, and requires leadership to write down that the original call was wrong.
The mechanism that connects both cases is the same. A rational decision, made at a specific set of stakes, survived past the moment the stakes changed. It survived not because anyone re-examined it and decided to hold.
It survived because no system existed to require re-examination.
The clause aged quietly inside a 2020 agreement. The architecture aged quietly inside engineering’s sprint cycle. Neither was a secret. Both expired on a schedule that nobody tracked, because ownership of the risk was structural and therefore nominal, which is to say it belonged to nobody at all.
THE SYSTEM PATTERN
What the pattern is
A known risk with no named owner ages toward its expiry date on a schedule the organization never tracks.
Where it repeats
In M&A due diligence, where contractual triggers on target assets are reviewed by deal counsel, modeled by finance, and owned by neither.
In technology architecture, where tradeoffs made under one regulatory regime persist unchanged into the next.
In supplier agreements, where change-of-control and force majeure clauses are negotiated at signing and never re-priced when concentration risk accumulates.
In any governance structure where a standing decision lives inside one function and the function that would authorize re-examining it is never invited in.
Where leaders miss it
The exact moment the signal compresses is when momentum replaces accountability as the organizing logic of the process.
In the deal, it was the moment two months of board socialization and financing conversations advanced while the clause question remained open.
In the Coinbase case, it was every budget cycle where the matching engine architecture was affirmed by silence because no one outside engineering was authorized to ask whether the original rationale still applied.
Leaders do not miss these signals because they are hidden. They miss them because the cost of slowing momentum to verify them is visible, and the cost of not verifying them is invisible until it is not.
NORMAN’S LAW
The law in play
Norman’s Law: when external pressure exceeds internal regulation, instability follows.
The Norman Gap here is the distance between how fast the context around a standing decision changes and how fast the organization’s review mechanism can price that change.
In both cases this week, the gap was not a week or a month. It was years. The pressure was always legible. The regulation, the mechanism for re-pricing it, did not exist.
What it predicts
A leader whose organization has no formal mechanism for migrating standing decisions to the level authorized to re-price them will discover the cost of that gap at the worst possible moment. Not when it is most convenient to absorb. At the moment external pressure converts a latent risk into a realized one.
What it demands
One thing: a named owner for every standing tradeoff whose failure produces a regulatory disclosure, a trading halt, a deal termination, or a customer-facing outage. Not a committee. A name. Assigned in advance, not in the aftermath.
MOS ARCHITECTURE
The operating system correction is a standing tradeoff registry.
Not a risk register, which is a static document.
A registry with an expiry mechanism. Every tradeoff that produces material consequence if it fails gets a named owner, a re-price trigger, and a review date that arrives before the expiry does.
The trigger is not engineering opinion.
It is external event: regulatory surface area expands, revenue concentration on the asset crosses a threshold, or the counterparty profile changes. When the trigger fires, the tradeoff migrates automatically to the level authorized to re-price it. That level includes risk and finance, not just the function that made the original call.
THE INNER OPERATING SYSTEM (IOS)
The internal load is the discipline to slow the exciting process long enough to surface the unglamorous question.
A leader under deal pressure or turnaround pressure or roadmap pressure is pulled toward the win that reframes the narrative. That pull is real and not irrational. The IOS correction is not to suppress the appetite. It is to hold the structural question open while the strategic one accelerates, and to refuse to let the answer to the second serve as the answer to the first.
The clause that could kill the deal and the architecture that could black out the exchange deserve a named owner before momentum decides they do not.
THE MONDAY QUESTION
Which standing tradeoff in your organization has the highest cost of failure, no named owner, and has not been formally re-priced since the context around it changed?
IF YOU DO ONE THING THIS WEEK
Build the list.
Before Friday, identify the three standing decisions in your operation whose failure would produce a regulatory disclosure, a trading halt, a deal break, or a material customer consequence.
For each one, write a single name, not a team, not a function. The name of the person currently authorized to re-price it. If that name does not exist, you have found the gap. Fill it this week, before external pressure does it for you.
The urgency is not that these risks are about to fire. The urgency is that you currently have no mechanism to know when they will.
SIGNAL SCORE
Sunday Score: 8.6/10 — Two regulated systems failed the same governance mechanism in the same week. The pattern is not sector-specific.
Week Average: 8.3/10 — Ownership Diffusion and architectural governance failure across two structurally different industries at the same moment is a convergence signal, not a coincidence.
FINAL SIGNAL
A risk that belongs to the deal, the system, or the function belongs to no one, and no one is tracking when it expires.
CTA
Send this to the one leader running a deal or an operation where the most expensive tradeoff still has no name attached to it.
WHAT THE TEMPERED SIGNAL REVIEWED THIS WEEK
Reuters, PYMNTS, Business of Fashion, and Associated Press reporting on Estee Lauder and Puig termination, May 19 to 22, 2026. Brian Armstrong public statement on X, May 8, 2026. Rob Witoff, Coinbase Head of Platform, engineering note, May 8, 2026. Coinbase Q1 2026 earnings, May 7, 2026. AWS US-EAST-1 service status, May 7 to 8, 2026. Reuters, CoinDesk, Benzinga
And if you want the full training system - REGULATE is on Amazon.