Two restaurants receive the same review on the same night: one star, a cold entrée, a server who vanished. The first owner replies at 9:40 the next morning — apology, fix, an invitation with her name on it. The second owner replies the following Thursday with the same words. The first review is edited upward within a week and the reviewer returns. The second review sits untouched at the top of the profile for a month, doing its work on every customer who reads it.
Same words. Wildly different outcomes. The variable was time, and it is the most underrated variable in reputation management. Businesses obsess over what to say — and the what does matter — but when you say it quietly decides how much the what is worth.
The 6-hour rule, defined
The rule: a negative review gets a public reply within six waking hours of being posted. Not six business days, not "when we do reviews on Fridays." Six hours during the part of the day when the reviewer is awake and still carrying the emotion that made them write.
Six is not magic. It is a ceiling derived from three practical observations: most reviews are posted within hours of the experience, while the details — and the heat — are fresh; most reviewers revisit the platform within a day, usually to see whether anyone responded; and review platforms sort and surface recent activity, so a fast, visible exchange changes what the next hundred readers see. Inside six hours you are having a conversation. Outside it you are leaving a note on a door the customer has already walked away from.
Why speed changes the outcome
Three mechanisms stack, and each one favors the fast responder.
- The reviewer is still reachable. Anger decays into narrative. In the first hours, a reviewer is a person with a problem; a day later they are a person with a story about your business. Apologies work on problems. Stories are much harder to edit — which is why same-day replies are where rating changes overwhelmingly come from.
- Future readers are scoring the exchange. Most people who see a negative review never met the reviewer. They read the thread as a proxy for how they would be treated. A reply timestamped hours after the complaint reads as a business that listens continuously; one timestamped weeks later reads as reputation janitorial work. Same words, different signal.
- Platforms reward recency. Review surfaces — Google's local results most visibly — weight fresh activity, and an active thread keeps your profile's recent history mixed rather than purely negative. A fast response does not hide the bad review; it prevents the bad review from being the last thing said.
What the pattern looks like in the data
We do not publish a controlled study here — review platforms do not offer A/B testing, and any vendor quoting you an exact uplift percentage is extrapolating. What monitoring data across managed local-business accounts shows consistently, and what platform-side research on review engagement supports directionally, is a steep decay curve: replies posted the same day account for the large majority of rating edits and updates, and the probability of any change collapses as response age stretches past a few days.
A reconstructed example from a demo workspace makes the shape concrete. A dental practice with a 4.2 average gets a cluster of three 1-star reviews in a weekend — billing confusion, same root cause. In the slow-response scenario the reviews sit unanswered for nine days: no edits, a measurable dip in profile-driven calls, and the cluster still ranking among the "most relevant" reviews a month later. In the fast-response scenario each gets a specific reply within hours and a direct line to the office manager: two of the three reviews are updated upward within ten days, and the third at least carries a visible, reasonable business answer under it. Simulated data, labeled as such — but the shape matches what any practitioner will recognize from real accounts.
Spike detection: catching the cluster before it sets
Single bad reviews are a tax. Clusters are a threat. Three 1-star reviews in a weekend is rarely three independent accidents; it is one root cause — a new hire, a broken process, a billing change — expressing itself repeatedly, or occasionally a coordinated attack. Either way, the cost is not linear: clusters rank prominently in relevance-sorted review views, and they change the story your profile tells from "occasionally imperfect" to "something is wrong there lately."
Catching a cluster requires noticing a rate, not an event. A single 1-star review against a baseline of two reviews a week is noise; three in forty-eight hours is a spike. This is a solved problem technically — compare current mention volume against a trailing baseline and fire when the deviation crosses a threshold — but it is unsolvable manually, because the human checking reviews once a day is themselves the sampling gap. The cluster that starts Friday at 7pm has a sixteen-hour head start on the Saturday-morning check.
Alerting that actually reaches you
The six-hour ceiling is an alerting problem before it is a writing problem. Drafting a good reply takes ten minutes with the right templates; the other five hours and fifty minutes are usually spent not knowing the review exists. Effective alerting has three properties:
- Threshold-based, not firehose. Every-review pings train you to ignore pings. Alert on what breaches your SLA: ratings below a threshold, spikes above baseline, negative themes repeating.
- Delivered where you already are. An alert that lands in a dashboard you open weekly is an archive, not an alarm. Slack and email reach the pocket; the dashboard is where you go after.
- Evidence attached. A useful alert carries the review text, the source, the velocity chart, and a suggested next step. If the alert makes you hunt for context, the context hunt will happen later — and later is the whole problem.
Building the six-hour habit
If you are running reputation manually, the rule is achievable with structure rather than software:
- Fix two check windows. Morning and late afternoon, every day, calendar-blocked. Two checks cap worst-case discovery at roughly half a day — not six hours, but inside the same day, which is where most of the recovery value lives.
- Pre-write your skeletons. Keep the four core templates one click away. Speed dies in the blank page, not in the typing.
- Name one owner per window. "The team checks reviews" means nobody checks reviews. One named person owns each window, with a named backup.
- Track your median response time. You cannot run a six-hour SLA you do not measure. Log posted-at and replied-at for a month; the median tells you whether you have a speed problem or a discovery problem.
- Instrument when you outgrow the windows. The day you miss a cluster because it started between checks is the day alerting pays for itself. That is the exact moment the monitoring loop stops being optional.
The bottom line
The perfect reply sent next week is a weak reply. The good reply sent today is reputation management.
Six hours is a ceiling, not a target — faster is better, and same-day is the floor that keeps recovery on the table. Measure your median, close the discovery gap, and let speed make your words worth more.