Most leaders meet PR at its worst moment: something has gone wrong, the phone is ringing, and someone asks what "the statement" should say. By then the useful work is already behind you. Public relations done well is mostly quiet, continuous, and finished long before a crisis arrives, and the leaders who understand that treat it as a relationship discipline, not a press-release factory.

The quick version

  • PR is relationship management, not publicity. The job is mutual understanding between an organisation and the "publics" it depends on, customers, staff, regulators, investors, communities, not just coverage.
  • There are four recognised modes of practice (Grunig & Hunt), from one-way hype to genuine two-way dialogue. Most reputational damage comes from operating in the wrong mode for the situation.
  • Reputation is earned slowly and lost fast. In a crisis, the research is blunt: accepting responsibility and getting ahead of the story usually protects reputation better than denial.
  • Communications can't fix what's actually broken. PR's hard limit is that words manage perception of reality; they don't replace it. The best message in the world won't outrun bad behaviour.

The idea in depth: PR is a management function, not a megaphone

The cleanest definition still comes from a counting exercise. In 1976 Rex Harlow, a founder of what became the Public Relations Society of America, gathered 472 separate definitions of public relations and distilled them into one working description: a distinctive management function that helps establish and maintain mutual lines of communication, understanding and cooperation between an organisation and its publics, and serves as an "early warning system" for emerging issues. Two words in there do the heavy lifting. Management, it's a leadership responsibility, not a task you delegate to whoever writes well. And mutual, it's two-directional; the organisation is supposed to be listening, not only broadcasting.

The PRSA's crowd-sourced 2012 update keeps the same spine in plainer language: "Public relations is a strategic communication process that builds mutually beneficial relationships between organizations and their publics." So the practical test is what you count. Stop measuring PR by output, releases sent, mentions tallied, and start measuring it by relationship health with the three or four publics that can actually make or break you. If you can't name those publics and say how each currently sees you, you don't have a communications strategy. You have a press list.

"Publics" is doing real work here. A public isn't an audience you talk at; it's a group whose interests intersect with yours and who can act on you, walk out, file a complaint, change a rule, switch supplier. James Grunig, whose research underpins most modern PR theory, built his field-defining work on the idea that an organisation's effectiveness depends on how well it manages relationships with these interdependent groups, not on how loud it is.

The four models: which conversation are you actually having?

The most useful map of practice is the oldest. In Managing Public Relations (1984), James Grunig and Todd Hunt described four models of how organisations communicate, and the value of the model is diagnostic, it tells you which mode you're stuck in.

flowchart TD
  Start("How does the organisation communicate?") --> Dir("One-way or two-way?")
  Dir --> One(["One-way: we talk, you listen"])
  Dir --> Two(["Two-way: we also listen and adapt"])
  One --> PA(["Press agentry: publicity at any cost, truth optional"])
  One --> PI(["Public information: accurate facts, but still a broadcast"])
  Two --> Asym(["Two-way asymmetrical: we research you, to persuade you"])
  Two --> Sym(["Two-way symmetrical: dialogue, mutual adjustment"])
  Sym --> Best(["Grunig's 'excellent' mode: builds durable trust"])
Grunig & Hunt's four models, read as a decision tree. Most organisations default to one-way modes and only reach for dialogue under pressure. Leaders Loop

Press agentry is pure publicity, attention now, accuracy negotiable. Public information tells the truth but still treats people as recipients: a one-way bulletin. Two-way asymmetrical researches an audience only to persuade them more efficiently, the listening is instrumental. Two-way symmetrical uses dialogue to reach genuine mutual understanding, with the organisation willing to change too. Grunig's later "Excellence" study, a multi-year IABC-funded programme into what distinguishes effective communication departments, argued this symmetrical mode is the marker of excellent practice, because it builds the relationships that survive a bad week.

So before you draft anything, name which model you're actually in and whether it fits the moment. A product recall handled in press-agentry mode ("minimise it, control the optics") reads as evasive and compounds the damage. The same recall handled symmetrically ("here's what happened, here's what we're changing, tell us what you're seeing") rebuilds trust. The failure mode is reaching for the megaphone when the situation is asking for a conversation.

The honest limitation: the symmetrical model is an ideal more than a description. Critics, and Grunig himself, later, concede pure two-way symmetry is rare: organisations have interests they won't negotiate away, and "mutual adjustment" can quietly become a polite cover for sophisticated persuasion. Treat the four models as a mirror for self-awareness, not a moral ladder. The point isn't to always be symmetrical; it's to stop pretending you're having a dialogue when you're running a broadcast.

Reputation under fire: what the crisis research actually says

Reputation is the accumulated judgement your publics hold about you, and it behaves asymmetrically, built in drips, drained in floods. The most evidence-based guide to defending it under pressure is W. Timothy Coombs's Situational Crisis Communication Theory (SCCT), set out in Corporate Reputation Review in 2007. Its core claim is that there's no single "right" crisis response; the response must match the level of responsibility the public attributes to you.

Coombs sorts crises by attributed responsibility, from victim crises (a natural disaster, a malicious rumour, where you're seen as a casualty too) through accidental ones to preventable ones (you knew, or should have, and didn't act). The greater the attributed responsibility, the more accommodative your response has to be: denial can work for a victim crisis, but for a preventable one only visible concern for those harmed, and often a full apology and corrective action, protects reputation. Prior crises or a poor track record make the same event land harder.

flowchart TD
  Q("How responsible will the public hold us?") --> Low("Low: a victim crisis")
  Q --> Mid("Some: an accident")
  Q --> High("High: preventable, we should have acted")
  Low --> R1(["Deny / inform may be enough"])
  Mid --> R2(["Express concern, explain, correct"])
  High --> R3(["Lead with the victims: apologise, fix, compensate"])
  R3 --> Note(["A poor track record forces you further toward accommodation"])
SCCT in one decision: match the accommodation in your response to the responsibility the public attributes to you. Leaders Loop

"As the public increasingly holds you responsible, the only message that protects reputation is one that visibly puts the victims first.", the practical core of SCCT

The practical implication is uncomfortable for most comms instincts. In the first hour, don't argue about wording, diagnose attribution. Ask one question: how responsible will people think we are? The answer dictates the response, not the other way round. Reaching for denial when the public already blames you isn't strategy. It's self-harm with a press release attached.

A second, narrower finding is worth carrying around. "Stealing thunder", disclosing your own bad news before a journalist, regulator or whistleblower does, consistently produces less reputational damage in the research than letting a third party break it. Experimental work going back to Arpan and Roskos-Ewoldsen (2005), summarised in the Handbook of Crisis Communication, finds self-disclosure raises perceived credibility, because volunteering damaging information runs against what people expect of organisations. The rule of thumb that falls out of this: if it's going to come out, you break it. Own the framing, the timing and the first impression rather than reacting to someone else's version of events.

Where this gets shakier: SCCT is grounded largely in experiments and survey scenarios, not the messy reality of a live crisis, and "stealing thunder" can backfire if your motives look manipulative, the effect weakens when people sense you're disclosing for advantage rather than honesty. None of it substitutes for the underlying fix, either. Communications shapes how a problem is perceived; it can't make a real problem disappear. If the product is genuinely unsafe, the only PR strategy that lasts is to make it safe.

A worked example

Picture a mid-sized food brand, call it Harbourline, that discovers a packaging fault: a small batch of jars can crack and shed glass. No one has been hurt yet, but the risk is real and the company spotted it before any customer reported it. (Illustrative figures and scenario, for explanation only.)

The press-agentry instinct is to stay quiet, pull the batch, and hope it passes. SCCT says read the attribution first: this is drifting toward a preventable crisis (an internal quality lapse), where the public will hold Harbourline responsible, so a victim-first response is required, not a minimising one. The "stealing thunder" evidence says disclose before a customer or regulator does.

So Harbourline issues the recall itself, fast and plainly: which batch codes, the specific risk, exactly how to get a refund, and an apology that leads with customer safety rather than corporate inconvenience. It opens a phone line and answers honestly, the two-way symmetrical mode, and publishes what it's changing so the fix is visible. The alternative, a journalist breaking the story after a customer is cut, shifts the narrative from "responsible company caught a fault" to "company that hid a danger," a reputational hole that costs far more to climb out of. The move that looked expensive was the cheap one.

Frequently asked questions

Isn't PR just spin or free advertising?

No, and treating it that way is the most common mistake. Advertising is paid, controlled space you talk at people through; PR is earned credibility built by managing relationships and what third parties (journalists, customers, regulators) say about you. "Spin" is press agentry, one-way persuasion with truth optional, and it's the mode most likely to detonate when scrutiny arrives. Sustainable PR is closer to relationship management than to either.

What's the difference between PR, communications, and reputation?

Communications is the broad activity of conveying messages; public relations is the strategic management of relationships with the specific publics an organisation depends on (a subset of communications, focused on those stakeholders). Reputation is the outcome, the accumulated judgement those publics hold over time. You do communications and PR; reputation is what you're left holding.

How fast should we respond to a crisis?

Fast on acknowledgement, careful on detail. The "stealing thunder" research favours getting ahead of the story, and silence reads as guilt or incompetence. But speed serves accuracy, not the reverse: a quick honest holding line ("we're aware, we're investigating, here's what we know and when we'll update you") beats both silence and a confident statement you later have to retract. Never let the desire to look in control push you into claims you can't stand behind.

Do small organisations need any of this?

Yes, scaled down. You don't need a comms department; you need to know your three or four key publics, keep honest two-way lines open with them, and decide in advance who speaks if something goes wrong. The four-models diagnosis and the SCCT attribution question work for a five-person company as well as a multinational, cheap to apply, expensive to ignore.

How do we measure whether PR is working?

Not by counting mentions. Measure the health of the relationships that matter, trust, sentiment and behaviour among your key publics (do staff stay, do customers advocate, do regulators give you the benefit of the doubt?), and track reputation over time, not campaign by campaign. Industry trust studies such as the annual Edelman Trust Barometer are a useful external benchmark for how much credit your sector starts with.

Related in the Toolkit

Where to go next