Systems handle the steady state. This handles the fire.
Your review generation system keeps new Google reviews flowing. Your review-response routine answers the Tuesday three-star politely. Your service-recovery process turns one unhappy customer into a saved job. None of those are incident response. A reputation incident is different: a single event, sudden and public, that threatens to become the story people find when they search your name. The viral complaint video. The one-star pile-on in a local Facebook group after a jobsite dispute. The accident that makes the evening news. The employee whose off-duty behavior gets tagged with your company name.
Incidents punish improvisation. The businesses that survive them cleanly aren't the ones with the best PR instincts in the moment — they're the ones who decided before anything went wrong who triages, who speaks, what gets paused, and what happens after. That's what this playbook is. Build it now, while nothing is on fire.
Severity triage: levels 1 through 3
Not every bad moment is a crisis. Over-responding burns goodwill and amplifies small problems; under-responding lets real ones compound. Triage within the first thirty minutes.
- Level 1 — Localized complaint. One angry post in a community group, one scathing review, one dispute with a customer who is venting but has limited reach. Handle through your normal service-recovery and review-response processes. No incident protocol needed. Monitor for 48 hours in case it escalates.
- Level 2 — Spreading event. A video with real traction, a cluster of new negative reviews in a short window (a review bomb), a thread being shared across multiple local groups, or coverage by a local blogger or news site. Activate the full first-hour protocol below.
- Level 3 — High-stakes incident. A jobsite accident or injury, a safety or licensing issue, an employee incident with legal exposure, or anything involving a regulator, attorney, or insurance carrier. Activate the protocol and involve legal counsel before any substantive public statement. In these cases the goal shifts from reputation protection to accurate, legally sound communication — speed matters less than correctness.
The first-hour protocol
For Level 2 and 3 events, execute these steps in order:
- Pause scheduled marketing. Silence the automated email campaign, the social posts queued for the week, the ads running. An ad promoting your service landing in the same feed or search results as the complaint is self-inflicted damage. Pause everything, don't delete anything, and resume channel by channel once the acute phase passes.
- Verify the facts internally. Call the crew lead, the office manager, the tech on the job. Establish what actually happened, what you know, and what you don't. Respond to verified facts only. Guessing publicly is how a manageable incident becomes a credibility crisis.
- Name one spokesperson. Usually the owner for a local business. Everyone else — employees, subs, family members — refers all inquiries to that person. One voice, one story, zero improvised comments from the field.
- Screenshot and document everything. The posts, the reviews, the timestamps, the video. Platforms remove content; disputes with insurers and attorneys require records.
- Decide: respond publicly or hold. Use the framework below. Don't default to responding just because the comments are loud.
Holding-statement anatomy
A holding statement is a short, non-defensive acknowledgment used when you know something happened but the facts aren't fully established. It works when it has all four of these elements:
- Acknowledgment without admission or argument. "We're aware of the situation being discussed online and take it seriously."
- What you're doing. "We've contacted the customer directly and are reviewing what happened on the job."
- A concrete next step. "We'll share an update once our review is complete."
- An off-ramp to private resolution. "Anyone with information can reach the owner directly at [phone]."
What it never contains: blame, sarcasm, speculation, promises you can't keep, or debate about anyone's character. Two to four sentences. If your draft needs a second paragraph, cut it.
When a public response helps — and when it amplifies
Respond publicly when the criticism is factual, the audience is local (people who could actually hire you are watching), and you have a specific, verifiable action to point to. A calm, specific public response read by a hundred neighbors often ends an incident's growth — the pile-on loses oxygen when it's clear a human owner is handling it.
Stay quiet publicly — while working private channels hard — when the content is rage-bait, the poster is clearly fishing for a public fight, the claims are false and legally actionable, or engagement would push the content in front of a far larger audience than it has now. A local Facebook post with 40 comments dies on its own in a day. The same post with your 2,000-word rebuttal gets screenshot and shared for a month. Respond once, factually, if at all — then move to direct contact.
Never argue with commenters
This rule has no exceptions. You will never win a comment-thread argument, because winning isn't available — only escalation. Spectators don't score the exchange on facts; they score it on tone, and the business always loses the tone fight. State your position once, offer the direct contact, and stop. Delete nothing except content that violates platform rules or contains threats (and document it first). Silence after one good response reads as professionalism; silence after twelve replies reads as a meltdown.
Handling review-platform floods within the rules
A review bomb — a burst of reviews from people who were never customers — is a common Level 2 event for contractors. Handle it inside platform policy, because anything outside it risks your entire profile.
- Flag reviews that violate policy. Google and other platforms prohibit reviews from people with no genuine experience of the business, reviews driven by brigading, and reviews with harassment or personal information. Report each one through the official flag flow with a calm, specific explanation. If a decision comes back as not violating, use the one-time appeal, which accepts up to ten related reviews together — documentation of the pattern is what wins it. Removal is inconsistent, but floods of clearly fake reviews do get taken down.
- Respond to what remains. For reviews you believe are genuine customers, use your standard response framework: acknowledge, take it offline, describe the fix. For fake ones you couldn't get removed, a single calm public note ("We have no record of this job or customer, but we'd like to investigate — please contact the owner at [phone]") signals to readers that the flood is suspect without accusing anyone.
- Ask real customers to keep being real. Continue your ordinary, always-on review generation — the same compliant ask you'd make any week. A bomb of fifteen suspicious one-stars followed by thirty genuine five-stars from verifiable customers is how the profile recovers — at your normal cadence, never accelerated or incentivized in response to the bomb. What you must never do: buy reviews, have employees or family post them, offer incentives for positive reviews, pre-filter unhappy customers away from the platform (review gating), or retaliate against reviewers. Every one of those violates platform terms and can cost you the profile entirely — the one asset an incident can't afford to lose.
Internal escalation and the owner-takes-the-call rule
Two rules keep an incident from fracturing internally. First, the owner takes the call: any media inquiry, any attorney, any angry customer with a public following goes to the owner personally — not to the office manager, not to the tech involved, not to a marketing vendor answering your inbox. For anything touching injury, legal action, licensing, or safety, the owner's first move is to bring in counsel and the insurer before any substantive response — the owner remains the single point of contact, not the drafter of substantive replies. Second, escalate on a defined trigger: any mention of injury, legal action, licensing, or safety goes immediately to counsel and the insurer, before any reply is drafted. Staff should know these two rules cold; most incident damage comes from a well-meaning employee improvising a response nobody authorized.
The 30–90 day rebuild: turn the fix into proof
The acute phase ends in days. The rebuild is what determines whether the incident becomes a footnote or your new identity. Use the following window deliberately:
- Days 1–30: Over-deliver visibly. Complete the remediation promised to the affected customer, then document it. Finish every open job cleanly. This is not the quarter to cut corners.
- Days 30–60: Publish proof, not apologies. If the incident revealed a real gap — training, safety process, communication — implement the fix and talk about the system, not the scandal: "Every jobsite now gets a documented walkthrough with the customer before we leave." Specific process change reads as maturity; repeated apologies read as weakness.
- Days 60–90: Resume full marketing with the story built in. Restart ads and campaigns. Fold the improvement into your sales materials — "here's what happened to us and what we changed" is one of the most trust-building things a local business can say, because every prospect knows things go wrong; what they're buying is how you handle it.
What not to do
The failure modes are consistent enough to list outright: don't post fake reviews or solicit them from anyone without a genuine customer experience. Don't gate reviews by routing unhappy customers to a private form. Don't offer discounts or gifts in exchange for review changes. Don't publicly accuse a reviewer of lying, even when they are — handle disputes through platform reporting and private channels. Don't delete your profile or mass-delete comments. Don't let employees respond, even to defend the company. And don't treat the incident as a marketing opportunity in the moment — the rebuild phase is where the story earns its place in your messaging, not day two.
An incident is a stress test of everything your reputation system built. Businesses with a steady base of genuine reviews, a clean profile, and a visible owner survive these events routinely — the reservoir of trust absorbs the hit. Businesses with thin reputations don't. That's the real argument for building the system before you need it, and the playbook for the day you do.