Making Difficult Decisions as a Leader

In September 1982, after seven Chicago-area deaths from cyanide-laced Tylenol, James Burke had to decide whether Johnson & Johnson would pull 31 million bottles at a cost later put by Wharton at about $100 million. No leadership seminar prepares you for that. That is making difficult decisions as a leader in the only sense that matters: your name is the one that will be attached to the outcome.

Making difficult decisions as a leader means naming the one assumption carrying the highest stakes and testing it before the organisation commits to the outcome. In my experience, hard decisions come from significance. One weak assumption can do real damage to the Purpose when the stakes are high, while the organisation is still admiring its paperwork.

Burke did the useful thing early. He named the Purpose before the lawyers and PR people could bury it in process, which they would have called prudent. Consumer safety came first, so the recall followed. Tamper-resistant packaging and public warnings made the return to market supportable.

Why making difficult decisions as a leader starts with significance

I do not care how tense the room feels. I care which assumption is holding the weight, and whether the Purpose survives if it breaks. That is where making difficult decisions as a leader becomes practical, because you can finally see what must be tested or contained.

I have watched directors try to lower significance by sounding firmer. The pressure stays exactly where it was, because a louder voice does not strengthen a weak assumption. If you want the wider frame around decision making for leaders, I deal with it there. Here the job is narrower and less glamorous: find the exposed assumption before the machinery, and the people paid to admire it, carry the room.

Making difficult decisions as a leader means naming the assumption

Click to expand

On July 19, 2024, CrowdStrike pushed a faulty content update to Windows machines running Falcon. Microsoft said 8.5 million devices were affected, and CrowdStrike's root cause analysis said a validator defect let the bad content through. The exposed leadership assumption sat inside the release process: validation and deployment would fail safely before customers were hit.

Nobody in that situation needed another speech about courage. They needed the one assumption with the blast radius written in plain English. I see the same problem in any difficult business decision. Test the assumption directly, or change the decision so the assumption cannot hurt you as much.

When I know which assumption is carrying the weight, the work becomes ordinary again. I either test it or change the decision shape. Hard decisions get easier when the decision is no longer naked. That is also the foundation of effective decision making in leadership more broadly, though people prefer the personality version.

Sufficient certainty is not maximum certainty

Abbott's Sturgis infant-formula shutdown in 2022 shows what tidy certainty can turn into. The FDA's own September 2022 review said the recall and production pause were necessary for safety and also worsened the shortage, and a U.S. International Trade Commission report said the Sturgis plant accounted for 20% of U.S. production. That is what maximum-certainty thinking does. The contamination file looks cleaner on paper while parents still have babies to feed.

I have seen the same habit in boards and public bodies for years. One narrow risk gets cleaned up because it can be measured neatly, while the wider Purpose is left bleeding in the corridor. Everyone involved can then call themselves cautious, which is one of the great institutional euphemisms. It usually means nobody wanted to own the trade-off in daylight.

I do not use the phrase "all the information" because it invites cowardice in formal clothes. Committees and consultants are fond of one more paper, because one more paper delays ownership and keeps the machine in work. Meanwhile the person with the signature still cannot say what would count as enough. This is where making difficult decisions as a leader usually goes stale, and where executive decision fatigue starts dressing itself up as prudence.

Change the decision before you try to change yourself

Netflix handled paid sharing better than most leaders handle strategic bets. It started in Chile, Costa Rica, and Peru, watched what customers actually did, and by the second quarter of 2023 had extended the model to more than 100 countries covering over 80% of revenue, according to its shareholder letter. Leaders prefer grand launches because pilots flatter the ego less. I prefer the pilot.

That is what I mean by reducing significance. I would rather change the decision than ask a leader to become a braver specimen five minutes before the vote. Stage the rollout and keep rollback authority, and you have already reduced the danger. In my experience, leaders resist this because a staged decision looks provisional in a board pack, and provisional does not flatter anybody. People sometimes ask for "Grant Purdy's method" as though I keep a secret flourish in the cupboard. I don't. In the Universal Decision-Making Method, these are secondary elements, the safeguards that stop a hard decision turning into a vanity project.

Write the trigger while you are calm

Pilots have a blunt phrase for one common failure, got-to-get-there-itis. The FAA says few pilots are immune to it. The Flight Safety Foundation says that if stabilisation criteria are not met by the required height, the answer is a go-around. The useful point is not the phrase. The trigger was written before pride started bargaining.

Leaders need the same discipline. Before the decision leaves the room, write the signal that will stop it and name who owns that signal. In practice, making difficult decisions as a leader gets easier when that trigger is written before the room starts arguing with reality. I have sat through too many reviews where everyone saw the warning and nobody owned the call. If nobody owns the trigger, the meeting did not decide anything.

---

Grant Purdy is the co-author, with Roger Estall, of Deciding (2020), and the architect of the Universal Decision-Making Method.