A crisis communications plan survives contact when it answers three questions before the crisis arrives: who decides, what can be said without further approval, and through which channels. Plans that fail almost always fail on the first question.
The document itself matters far less than the decisions encoded in it. A one-page plan people have rehearsed beats a forty-page plan nobody has opened.
Decision rights, written down and named
In the first hour of a crisis the binding constraint is authority, not information. Someone must be able to say something publicly without convening a committee.
Write this down explicitly:
- Who is the single spokesperson, and who takes that role if they are unreachable, travelling, or personally involved in the incident?
- What can that person say without additional sign-off? This is the most important line in the entire plan.
- What requires escalation, to whom, and within what time?
- Who is explicitly not authorised to speak — stated kindly but unambiguously, because well-meaning staff filling a silence is a common failure mode.
Name individuals, not job titles alone. Titles are ambiguous under pressure; names are not.
Pre-clear the language you will certainly need
You cannot draft the statement for an unknown event. You can draft the sentences every crisis requires, and clear them in advance with whoever would otherwise slow you down.
The reliably reusable pieces are: an acknowledgement that something has happened and you are establishing the facts; a holding line for when you genuinely do not yet know; a statement of concern for anyone affected that does not admit or deny liability; and a commitment to update at a stated time. That last one is underrated — a promise to say more at four o'clock buys hours of goodwill and costs nothing.
Get legal or board approval for these while nobody is panicking. That approval is the thing you will not be able to obtain quickly when you need it.
Own at least one channel outright
A plan that depends entirely on social platforms depends on systems you do not control and cannot appeal quickly. Accounts get restricted, reach collapses, and in some contexts the network itself is switched off.
Maintain something you own: a website page you can update, a mailing list, an SMS list, a relationship with a radio station. Test that you can actually publish to it — many organisations discover during a crisis that the only person with the website password has left.
Where connectivity itself is at risk, the fallbacks in verifying information during internet shutdowns apply to outbound communication as much as to verification.
Rehearse the plan, not the scenario
Elaborate scenario exercises are enjoyable and mostly test imagination. What needs testing is the machinery.
A short, useful drill: at an unannounced moment, send a message to the team saying an incident has occurred. Then measure how long it takes for the right person to be reached, for a holding statement to be approved, and for it to appear on a channel you control. Whatever that number is, it is your real response time. Most organisations are shocked by it the first time.
Run it twice a year. It takes an hour and it finds the broken links — the wrong phone number, the dormant account, the approver on leave — while they are cheap to fix.
Say what is true, including that you do not know
The instinct to withhold until you have the full picture is understandable and usually counterproductive. Silence is read as concealment, and the gap fills with other people's accounts.
Saying "here is what we know, here is what we do not, here is when we will next update" is almost always stronger than saying nothing. It is also the only position you cannot later be caught out on.
Where the crisis involves harm to individuals, the standards in trauma-informed reporting on gender-based violence apply to your own communications too: no identifying detail without informed consent, and no use of a person's experience to demonstrate your responsiveness.
After: measure the response, not the coverage
Post-crisis reviews tend to count mentions and sentiment. More useful questions: did the people directly affected receive accurate information, and when? Was our first statement accurate in hindsight? Where did the delay actually occur?
That last question is the one that improves the next response. We look at honest measurement in measuring campaign reach honestly.
Common questions
How long should a crisis communications plan be?
Short enough to be read in full during an emergency. One page of decisions plus an annex of contacts and pre-cleared language is a workable shape.
Who should hold the plan?
Everyone named in it, in a form reachable without the office network — including on paper. A plan stored only on a system that may itself be affected is not a plan.
How often should it be reviewed?
Whenever a named person changes role, and at minimum twice a year. Staff turnover, not strategy drift, is what usually invalidates these documents.
If you would like this built and rehearsed with your team rather than written and filed, let us talk.