Case Study 1 — Restructuring the Challenger Memo

A note on accuracy. This case study works with the well-established historical record of the Challenger disaster (the cold-temperature O-ring concern, the night-before teleconference between Morton Thiokol and NASA, the failure to persuade decision-makers, the loss of the shuttle and its seven-person crew on January 28, 1986). The "before" and "after" memos below are reconstructions for teaching, not transcriptions of the actual documents—we are illustrating a structural principle, not quoting the record. We invent no casualty details, no quotations attributed to real people, and no facts beyond the documented concern that cold temperatures degraded O-ring sealing. For the original data presentation, see Edward Tufte's analysis (Further Reading) and Chapter 9.

The situation

The solid rocket boosters were sealed at their joints by rubber O-rings. Evidence from prior flights suggested the O-rings sealed poorly when cold. The forecast for launch morning was far colder than any previous launch. The engineers who understood this had one night, one teleconference, and a set of charts to convince management not to launch. They did not succeed. The information was there. It did not arrive.

This is the purest possible lesson in the chapter's central theme: structure serves the reader. The reader, here, was a group of decision-makers under intense time pressure and organizational momentum. The communication had to put the single decisive fact where such a reader could not miss it. It did the opposite.

A teaching reconstruction of the "before"

Imagine a recommendation built the way technical arguments are usually built—bottom-up, evidence accumulating toward a conclusion at the end, the decisive relationship distributed across dense material:

❌ Before (reconstructed, bottom-up, buried conclusion):

Subject: O-ring Erosion Data — STS Booster Joint Performance Review

This memo reviews field-joint O-ring performance across recent missions. Flight 51-C experienced the most significant erosion observed to date. Several flights have shown blow-by and erosion of the primary O-ring to varying degrees. The resiliency of the O-ring material is a function of temperature; durometer and compression-set testing indicate reduced responsiveness at lower temperatures. Soot and blow-by patterns on recovered hardware have been documented and correlated with joint rotation under pressure. The secondary O-ring provides redundancy under nominal conditions, though its sealing capability under dynamic conditions is less well characterized. Ambient temperature at the time of several incidents has been recorded. Predicted O-ring temperature for the upcoming launch is approximately 29°F. We recommend the joint performance data be considered in the launch decision.

Read that the way a tired manager reads it at 11 p.m. Every clause is technically defensible. And the conclusion that should stop everything—it will be 29°F, the O-rings have never been tested anywhere near that cold, and cold is exactly when they fail—is nowhere prominent. It must be assembled by the reader from a temperature buried in the eighth sentence, a material property in the fourth, and an erosion history in the second. No scanning reader assembles it. The recommendation at the end ("be considered") asks for nothing. The lede is not just buried; it is dismembered and scattered across the page.

Against the chapter's principles, the failures are exact: - Bottom-up, not top-down (§4.2): evidence first, conclusion last—and the conclusion is weak ("be considered"). - No inverted pyramid / no BLUF (§4.3): the one decisive fact is not at the top. - No isolation of the key relationship: temperature-versus-failure is fragmented, never stated as one sentence or shown as one chart. - A non-recommendation: "be considered" tells a decision-maker nothing to do.

The restructured "after"

Now rebuild it for how the reader actually reads—conclusion first, the decisive relationship isolated and unmissable, a clear recommendation:

✅ After (reconstructed, top-down, BLUF):

Subject: RECOMMENDATION — Do Not Launch Below 53°F (O-ring failure risk in cold)

Bottom line: Do not launch tomorrow. Predicted O-ring temperature is ~29°F. The O-rings lose sealing ability as temperature drops, they have shown the worst erosion on our coldest prior flight, and they have never been tested anywhere near 29°F. We recommend no launch below 53°F—the coldest condition with flight data.

The one relationship that drives this: Colder O-ring → less resilient seal → more blow-by and erosion. Our coldest prior flight (53°F) showed the most erosion to date. Tomorrow would be ~24°F colder than that, far outside any condition we have data for. [Figure: a single chart, O-ring temperature on the x-axis vs. erosion/blow-by severity on the y-axis, every prior flight plotted, with the 29°F prediction marked far to the left of all existing data.]

Why the redundancy argument doesn't save us: The secondary O-ring is less characterized under dynamic conditions and cannot be relied on as the primary seal in an untested cold regime.

Recommended action: Delay until the O-ring temperature is at least 53°F. Supporting flight-by-flight data follows below.

What changed structurally: - The bottom line leads (§4.3). The first thing on the page is the recommendation and the decisive fact, in one block a reader cannot miss even while scanning. - The key relationship is isolated (§4.9). Temperature-versus-erosion is stated as one sentence and given one prominent chart, instead of being scattered. (Chapter 9 examines why a single well-designed chart here was the crux.) - The recommendation is actionable. "Do not launch below 53°F" tells the decision-maker exactly what to do and gives a number to argue about—far stronger than "be considered." - Detail is demoted, not deleted (inverted pyramid). The flight-by-flight data still exists, below, for anyone who wants to verify—but it no longer stands between the reader and the conclusion. - The subject line itself carries the bottom line. A reader who reads only the subject already knows the recommendation.

The lesson

We cannot claim a better-structured memo would have changed history—the failure that night was also organizational, cultural, and about power, not only about a document. But the communication lesson is undeniable and it generalizes to every report you will ever write: being correct is not the job; being received is the job. The engineers had the data and the right conclusion. What they lacked was structure that forced the one decisive fact to the top and held it there. When the stakes are high, the test is not "did I include it?" It is: if the reader reads only the first line and the headers, will they reach the one conclusion that matters?

Related: Chapter 4 §4.9 · Case Study 2 · Further Reading