Section 1
The first-hour sequence
Work in this order. Do not skip to writing replies, which is where most owners waste the hour. Minutes 0 to 10: confirm it is an attack, not a real crisis. Before you report anything, rule out a genuine cause. Did a viral complaint, a news mention, or a real service failure just happen? A flood of angry reviews from real, established local accounts describing a real event is not an attack, and reporting it as one will backfire. An attack looks different: new or low-activity accounts, reviews with no specific detail or copy-paste phrasing, raters from outside your service area, a spike with no triggering event you can find. Write down which pattern you see, because that judgment drives everything after. Minutes 10 to 30: capture evidence before it changes. Screenshot every suspicious review with its timestamp, the reviewer's profile, and their review history. Attackers delete accounts and platforms remove reviews, so the evidence you want for an appeal may be gone by tomorrow. Log each one in a simple table. This log is the artifact you will attach when you report the cluster as coordinated, and it is far more persuasive than reporting reviews one at a time. Minutes 30 to 45: report the pattern, not just the reviews. Flag each fake review through your Business Profile, and separately contact Business Profile support to report a coordinated attack. The distinction matters: an individual flag says "this review is bad," while a coordinated-attack report says "these fifteen reviews arrived together from suspicious accounts and here is the log." Lead with the pattern evidence. You are giving a human reviewer the same signal the automated filter is already looking for. Minutes 45 to 60: hold your public response. The instinct is to reply angrily to each one-star. Resist it for now. Public replies do not remove reviews, and a defensive wall of replies can make a temporary spike look like an ongoing dispute, which reads worse to the next real prospect than a quiet cluster the platform later removes. Draft one calm, non-defensive holding reply you can post to the most visible few if needed, and save the individual conversion-focused replies for after the dust settles. The reply strategy is its own discipline, and it belongs to the days after the flood, not the first hour.
Section 2
Why paying or panicking makes it worse
Two first-hour mistakes cost the most. The first is engaging the attacker, especially if a message arrived demanding payment to stop. Responding confirms a live target and often escalates the flood. The second is mass-deleting your own real reviews or gating in a panic, which can trip the very authenticity filters you want on your side. The calm move is boring: document, report the pattern, wait for the platform's systems and support to act on the signal you gave them.
Section 3
What the first hour cannot fix
Speed shapes classification. It does not guarantee removal. Platforms publish almost no per-case success data, appeal criteria are opaque, and some fake reviews survive even a clean report. Treat this sequence as improving your odds and protecting your evidence, not as a switch that reverses the attack. If the flood is large, expect a multi-day resolution and a rating dip while the filters work. The fitness test: You are ready for a review flood if, right now, you know where your Business Profile support flow is, you have a place to log evidence, and you have decided in advance that your first move is to document and report the pattern rather than reply to reviewers. If any of those three is missing, you are improvising during the one hour when improvisation costs the most.