Not knowing what you already know

73 seconds

Rocket engineer Robert Boisjoly was extremely concerned.

According to his calculations, a structural component that linked segments of the system he worked on was prone to failure if ambient temperatures were too low. This part had been assigned the highest criticality level for mission success. In recent days managers had cancelled launches several times as weather issues and minor equipment failures kept cropping up. Pressure was high to stick to the schedule, but he felt strongly that another delay was necessary.

On the night before the new launch, Boisjoly and his colleagues had an intense conference call with their counterparts where he laid out his reasoning. Notwithstanding the data his team sent and an extended back-and-forth, his warnings ultimately did not penetrate the managerial bureaucracy and convince those in authority to scrub yet again. The launch was still on when morning dawned, despite record cold.

Unlike many modern missions, this one did more than loft payloads into orbit. Its technical aspects were hardly the focus, as the American public was captivated by the presence among the crew of a school teacher, who was about to become the first member of the general public in space.

It was January 28, 1986, and the Space Shuttle Challenger waited on the launch pad at Cape Canaveral on an unusually frigid day in southern Florida. The temperature was below freezing and icicles hung from the launch structure.

The public countdown began at the familiar 10 seconds. With six seconds to go the shuttle’s main engines started and rapidly built up to full thrust. Even then it was not too late to abort, and an intervention by the automated systems or mission control could still have stopped the process. But once the two solid rocket boosters ignited nothing could be done, as their very design makes them impossible to extinguish once they start burning. A puff of black smoke emerged from the failing joint that had so worried Boisjoly, portending what was to come, and the fate of the shuttle was sealed.

Boisjoly watched the launch with trepidation. His initial relief upon safe liftoff would turn to horror 73 seconds later, when the Challenger disintegrated violently in a cloud of vaporized fuel, killing the seven astronauts on board.1

NASA had engineers working for it who knew exactly what the problem was. The specific cause of the disaster was debated in detail just one night prior. Concerns about this component dated back to the early years of the shuttle program, and everyone understood that failure of these rocket boosters would lead to the loss of the crew. Yet information that some parts of the organization knew well did not sway the ultimate decision-makers, with tragic consequences.2

Signal versus noise

NASA’s managers were dealing with one of the most complex machines ever built, made by a slew of contractors and worked on by thousands over many years. Issues raised by any of them could range in seriousness from trivial distraction to a matter of life and death. Sorting through the difference was tough. Specific problems were known in some quarters, but not adequately understood in other places where important decisions were made.

Triaging incoming information is a critical function of leadership. Even if your role is not one where lives hang in the balance, the ability to carry out your organization’s mission depends on the ability to find and listen to the right voices. The proliferation of data promises to accelerate as new sensors and cheaper technologies enfold even more areas into their domain. In an environment where metrics are increasingly tracked, stored, and available for unending analysis, finding what’s important becomes even more difficult. Three principles can help filter the flows, both for those seeking information and those providing it:

• Action must depend on it. Large organizations are fond of dazzling, color-coded dashboards full of metrics of dubious value, most of which are routinely ignored. If you require someone to produce data, there should be a decision of some kind that cannot be adequately made without it. If knowing something has zero effect on a course of action, filter it out. Being intriguing, or novel, or hard to obtain does not automatically make it useful. Can you act on it?
• It should be relevant to the overall effort. Just because you can change direction doesn’t mean that will be helpful. For NASA, the priority needed to be the safety of its astronauts, so any findings with a bearing on their wellbeing is always critical. On launch day, set aside the concerns about public relations, or budgets, or aesthetics.3 Does it matter?
• Information should be true. To the extent possible, what you depend on for decisions should be reliable. Are the methods used to produce it valid? Has the source been vetted? Is it accurate? In the case of Challenger, the team responsible for the boosters had been working on the design for years and was more familiar with its properties than anyone, and they had the data to back it up. Can you trust it?

History repeats itself

17 years after the loss of Challenger, the Space Shuttle Columbia launched on a routine mission, carrying the usual complement of acronym-heavy experiments, along with some school science projects included to capture the interest of students. The flight was unremarkable, save for the fact that shortly after liftoff a stray piece of insulating foam broke free and struck the left wing.

Engineers reviewing tracking videos noticed this impact and raised concerns to their superiors. The integrity of the Space Shuttle’s exterior was of extreme interest, as maintaining thermal protection during the unimaginable heat of re-entry is a core engineering challenge of space flight. Any compromise to the protective outer structure would have serious consequences. As with Challenger, these issues had been well understood and documented for years, with foam strikes a recurrent issue.

With Columbia safely in orbit staff sent a request to other agencies in the U.S. government with access to powerful and classified imaging systems that could inspect the shuttle for damage in space. Had the request been executed, they would have discovered a gaping hole in the wing, triggering a rescue mission that would have eclipsed anything in the history of space exploration for drama, with uncertain outcomes.4

But this request for visual evidence was caught by superiors and squelched as unnecessary before any images could be obtained. On re-entry, superheated plasma entered the wing and slowly destroyed it from the inside, before the entire shuttle disintegrated in the upper atmosphere over the southwestern United States. Once again, seven astronauts’ lives were lost.

The issue leading to failure was known by those working most closely on it. They attempted to raise it in real time, but amidst the noise their perspective never got through. The failures in management that led to this were numerous, but among them was an inability to identify and prioritize valuable information that experts within the system already possessed.

Unknown knowns

The larger an organization becomes, the more likely there are to be centers of competence whose voices are not being effectively heard. Sometimes the solution to your challenges are already somewhere in the hierarchy. No matter what level you occupy, help create the environment that is most likely to surface it. Leaders can adopt several tactics to support this:

• Reward those who tell uncomfortable truths in good faith. No one likes the complainer, or the person who highlights all the reasons why an idea won’t work without any suggestions for making it better. But the bias towards optimism and a solutions can blind you to very real risks, and the bigger and more successful you become the more likely these are to be existential. People who have advanced by always putting the best possible spin on things may be shocked when reality ultimately catches up with them.5 Former U.S. Secretary of State Colin Powell has put it succinctly: “Every organization should tolerate rebels who tell the emperor he has no clothes.”
• Talk regularly to those who do the core work.6 This can be practiced in creative ways. One business leader encouraged anyone leaving his company to put a sticky note on his office door with their reasons on their last day. Another would have progress reviews with consultants that explicitly left partners out of the room, ensuring the junior staff who did the work would speak without needing to appease their own bosses. However this happens, remember the right information does not always filter up.
• Give employees throughout your organization a real stake in the process. Toyota’s legendary automobile assembly process implemented a feature known as the andon cord, a physical toggle which was accessible to any worker to signal for help or even halt the entire assembly line if necessary. On occasion someone could set back the production schedule, but that was outweighed by the ability to catch problems early, before they propagated downstream and become more complicated to fix. For critical things, trust those who do the work, and when those closest to the front line suggest that something is wrong, listen closely.

Amidst a barrage of data, leaders are tasked with making the best decision, finding the next opportunity to pursue, or identifying pitfalls to avoid. It is possible the needed answers are already somewhere in your organization, just waiting to be found.

How are you filtering the information you receive, and how are you ensuring that the most insightful voices are being heard?

---

References

NPR had some of the earliest reporting on Boisjoly’s efforts.

For a gripping account of the final flight of Columbia, the compulsively-readable William Langewiesche’s article in the Atlantic sets the bar.

NASA hosts the extremely detailed Rogers Commission report on the Challenger disaster and the Columbia Accident Investigation Board report.