handling a pr problem.
the first hour decides the size of it. acknowledge fast, resolve slowly — not the other way round.
pause scheduled content, acknowledge within the hour, investigate properly, and publish one clear update. most damage comes from silence or defensiveness, not from the original problem.
most social media crises are small events that grew because of how they were handled. a complaint, a mistake, a badly judged post, an incident in a branch — none of which are fatal, and all of which escalate when a brand goes quiet, argues, or deletes.
the useful distinction is between acknowledging and resolving. acknowledgement has to be fast; resolution takes as long as it takes, and confusing the two produces either a hasty promise you cannot keep or a silence that reads as indifference.
it is also worth being honest about scale before reacting. a handful of angry comments is not a crisis, and treating it as one gives it a significance the audience had not assigned to it.
— 01the first hour.
three things immediately: pause all scheduled content, work out what actually happened, and post a short acknowledgement that does not yet claim to have the answer.
pausing the schedule matters more than it sounds. a cheerful promotional post landing in the middle of a complaint thread is the image that gets screenshotted and shared, and it converts a contained problem into evidence of a brand that is not listening.
the acknowledgement should be brief and human: that you have seen it, that you are looking into it properly, and when you will say more. no explanation yet, no defence, no legal register. a sentence written by a person beats a paragraph written by a committee. if the incident involves anything legal or regulatory, that sentence still works — acknowledging that you are looking into something admits nothing, and the person handling the inbox should be authorised to send it without waiting for approval.
and take it to a private channel where an individual is involved. resolve the specific case directly, then address the general point publicly if there is one. arguing case detail in a comment thread never ends well for the business. one public line saying you have messaged them directly is enough for everyone else watching.
— 02what not to do.
do not delete the comments or the post unless the content is abusive or unlawful. deletion is always noticed, always screenshotted first, and converts a complaint into a cover-up, which is a much larger story.
do not defend before you know. a confident denial that turns out to be wrong is the single most expensive move available, and it removes any credibility from the correction that follows. if a fact is still unverified an hour in, say that rather than filling the gap with an assumption.
and do not over-produce the response. a designed statement graphic in brand colours reads as a corporate exercise; plain text or a short unedited video from a named person reads as accountability. the register is the message. the same applies to language — corporate phrasing about regretting any inconvenience reads as an attempt to say nothing, which the audience recognises immediately.
— 03the update, and then the resolution.
publish one clear update when you actually know something: what happened, why, what you are changing, and what happens to anyone affected. specificity is what closes a story — vagueness invites people to keep asking. one substantial update also stops the drip of partial answers that keeps a story alive for a week.
own the part that is yours without over-apologising for the rest. an honest, narrow admission is more credible than a broad emotional apology that avoids the specific failure, and audiences are unusually good at spotting the difference.
then stop talking about it. a series of follow-up posts keeps the incident current and reaches people who never saw the original. resume normal content deliberately, without a transitional post about lessons learned. if something structural genuinely changed, mention it months later in the ordinary course of things, where it reads as practice rather than reputation management.
— 04prepare before you need it.
the response quality depends almost entirely on what exists in advance: who has authority to post during a crisis, who must be told, a holding statement structure, and an agreed threshold for what counts as a crisis rather than an ordinary complaint.
that last one prevents the more common failure, which is treating routine criticism as an emergency. most negative comments need a normal reply from whoever handles the inbox, and escalating them creates internal noise and slows the responses that matter. a workable threshold is volume plus reach: several unrelated people raising the same thing, or one complaint travelling beyond your own audience.
write it down in a page, not a manual, and make sure the person likely to see it first at nine on a friday evening knows what to do. that is the practical part of community management nobody budgets for until it is needed — and it is worth rehearsing once, because a plan nobody has read is not a plan. thirty minutes walking through two plausible scenarios is enough to expose the gaps.