A crisis team is rarely made up of inexperienced people. That is exactly why it is worth remembering that the mistakes which turn an incident into a catastrophe are not signs of incompetence. They are predictable features of human decision-making under pressure — and that is why crisis management has to be supported by structure rather than by good intentions.

Here are five pitfalls we see again and again, and what can be done about them.

1. Exponential blindness

Evolution taught us to expect change to happen gradually. Fires, pandemics and reputational crises spread exponentially. On 23 February 2020 there were ten registered COVID-19 cases in the United Kingdom. A month later there were more than 12,000, and the first lockdown had arrived.

The same arithmetic applies to a damaging social media post: ten people share it, and by the end of the day thousands or millions have seen it.

The failure: assuming the crisis grows at a speed you can keep up with, and therefore missing the signals that should change your course.

Design around it: Prepare for the worst-case scenario, and make sure no signal is lost. When status changes are logged as they happen, escalation becomes a decision made against a record rather than against recollection.

2. Focus on outcomes

We remember the fires we put out and forget the strategies that failed. The incident commander recalls the calls where pumping more water worked, not the ones where it did not. This bias compounds: every time we get away with a little more, the next decision is a little more daring. Before the Columbia disaster in 2003, NASA had got away with at least 14 earlier shuttle damage events. The threshold for what counted as acceptable damage crept up with every safe return.

The failure: reaching for the solution that worked last time instead of reasoning from the situation in front of you.

Design around it: Action cards are written in advance, in calm conditions, and they present the options rather than a single habitual reflex. They are revision-controlled, so lessons from the last event change what the next team sees.

3. Confirmation bias

Our brains reward information that confirms what we already believe and discount what contradicts it. In a crisis, that means the plan is assumed to be working because the people who wrote it want it to work.

The failure: not being willing to conclude that the plan has failed and must be replaced.

Design around it: Plan for the possibility that the plan must be scrapped — and log the deviation when it happens. A documented deviation is not an admission of failure. It is the evidence that produces a better plan.

4. “More heads are wiser than one”

It is widely assumed that a group makes more balanced decisions than an individual. The research says otherwise: without the right support, information and challenge, a group will accept a greater risk than an individual. Irving Janis described the phenomenon in 1972 as groupthink — an unshakeable belief in the group’s own decisions, sustained by group pressure and self-censorship.

The failure: a tight, cohesive crisis team that cannot hear its own doubts.

Design around it: Deliberately compose the team from different dispositions, and give someone the explicit role of devil’s advocate — the task of questioning and pointing out the weaknesses in other people’s ideas. A person who was not involved in writing the original plan is more likely to propose a different approach when one is needed.

5. Going against the current

Solomon Asch’s line-comparison experiments in the 1950s showed how far people will go to fit in: roughly three out of four participants gave a wrong answer at least once, simply because others in the room had given it first. Put that dynamic into a crisis team under pressure, where the wrong decision could turn a crisis into a disaster, and ask how many people would stand up to their colleagues.

The failure: silence. The person who sees the problem says nothing.

Design around it: Asch also showed the remedy. When just one other participant was instructed to give the correct answer, more of the unwitting participants held the correct answer too. And when participants wrote their answers down privately instead of declaring them in the room, fewer conformed. Technical solutions that let the parties respond from different locations rather than gathering in one room reduce group pressure.

That is a direct argument for a distributed crisis platform. When responders can report status and input from the harbour, the terminal, the bridge or another agency — rather than from a single table where the loudest voice sets the tone — the input is recorded as it was given.

You cannot train the biases away

You can design around them. Plan, exercise, plan, exercise. Hold structured status meetings. Build the team dynamic deliberately. And use a digital crisis tool, so the plan is always at hand and the response is a few clicks away.

Read more about the five pitfalls and the four practices that follow them on our page about crisis management pitfalls.