What belongs in a crisis communication plan — and what's just paperwork

A crisis communication plan sets out who decides, who speaks, and what gets said in the first hour of an incident. The useful version is two pages; the version that gets approved is usually forty.

Ask a communications team whether they have a crisis plan and almost all of them say yes. Ask when it was last opened during an actual incident and the answer is usually a pause. The document exists, it was signed off, and when something happened everyone worked from memory and a group chat.

That is not a discipline failure. It is a design failure. A forty-page plan cannot be used in the ten minutes when it would help, so it isn't.

The six things that get used

1. The escalation threshold

A written trigger: what turns an issue into a crisis, and who must be told when it does. This is the most valuable line in the document and the most commonly missing one.

Without it, the recurring failure is nobody escalating early enough, because nobody was certain it qualified and nobody wanted to be the person who panicked. The organisation loses its first two hours to hesitation. Make the threshold generous and explicit — "if in doubt, escalate; over-escalation carries no penalty" — and put that in writing, because the unwritten version is not believed.

2. Who decides

One named person, with a named deputy, and a rule for what happens when neither is reachable. Not a committee. Committees are how a statement takes four hours.

The decision rights need to be specific: who approves a holding statement, who approves a full statement, and what can go out without legal review. If everything requires counsel, nothing goes out in the first hour, which is a decision in itself — just not one anyone made deliberately.

3. Who speaks

A named spokesperson per scenario type, because the right voice for a safety incident is rarely the right voice for a financial one. Include the deputy, and be honest about who is actually able to do it — seniority and capability are different things, and an incident is a poor time to discover the gap. See media training for the rehearsal side of this.

4. Contact details that work

Mobile numbers, not desk extensions. Personal email as a fallback, because incidents take out corporate mail more often than people expect. Your external counsel, your insurer if a notification obligation exists, and the agency contact if you use one.

This is the section that decays fastest and matters most. A plan whose phone numbers are eighteen months stale is not a plan. Check it quarterly — it takes twenty minutes and it is the highest-return maintenance in the document.

5. Pre-cleared holding statements

Templates for your three or four most plausible scenarios, already reviewed by legal, with blanks for the specifics. This converts a ninety-minute drafting-and-review cycle into about twelve minutes.

The value is not the words — it is that the legal argument has already happened, calmly, at a time when nobody was under deadline. See the first 60 minutes for what belongs in one.

6. Stakeholder order

Who hears first, and through which channel. Staff before media is the rule most often broken and the one whose breach is most reliably punished — employees who learn about their own organisation's crisis from a news alert become a second problem, and often a leaking one.

What is just paperwork

None of the following is worthless, but none of it is consulted mid-incident, and all of it can live in a separate document without loss:

• The statement of communication philosophy. Nobody has ever opened a plan under pressure to be reminded that the organisation values transparency.
• Exhaustive scenario catalogues. Twenty scenarios at two pages each will not be searched at speed. Three or four archetypes with a clear structure generalise better than a catalogue that is always missing the one that actually happened.
• Organisational charts. Useful for planning, useless during. What is needed is a name and a mobile number, not a reporting line.
• Media contact lists. Stale within months, and in practice journalists contact you rather than the reverse.
• Detailed channel-by-channel messaging matrices. These assume time to consult a matrix. There isn't any.

The test for any section: would someone open this while a journalist is holding for an answer? If not, it belongs in the appendix or in a different file altogether.

Two pages, and a plan people can find

The operative plan should fit on two pages: threshold, decision-maker, spokesperson, contacts, templates, stakeholder order. Everything else is background.

A two-page plan that gets opened beats a forty-page plan that gets approved.

Then solve the retrieval problem, which is usually ignored. If the plan lives on the intranet and the incident is an outage, it is unreachable. If it is in a shared drive nobody can name at speed, it is unreachable. Printed in a drawer, on a phone, and in a location everyone on the response list can recite — that is a plan you can actually get to.

Testing it

Reviewing a plan means reading it. Testing it means running it, and only the second one tells you anything. Three tests, in increasing cost:

• The call test. Ring every number in the contact section, unannounced, on a weekday evening. Count how many reach a human. This takes half an hour and reliably finds the worst problem in most plans.
• The retrieval test. Message someone on the response list and ask them to produce the plan in five minutes without asking anyone. If they cannot, the content is irrelevant.
• The simulation. Run the scenario with the actual people, against the clock. This is the only test that surfaces decision-making failures rather than logistical ones — see running a crisis simulation.

And if the incident is likely to touch coverage you may dispute, know in advance which regulatory clocks start. In South Africa that means the Press Code for print and online and the BCCSA for broadcast; other jurisdictions have their own, on their own deadlines.