Reputation after a real mistake: the honest route
Work out the damage before deciding what to do about it. One one star review among 40 moves a 4.5 average to about 4.4, while 74 percent of consumers look at the last three months, so the recovery is a steady rate of new reviews rather than a campaign of explanation. The public reply is for the next reader.
Measure the damage first
The feeling after a bad review is out of proportion to the arithmetic. A 4.5 average across 40 reviews is 180 points. Add one review at 1 star and it becomes 181 across 41, which is 4.41. The threshold that 68 percent of consumers apply is four stars, and you are still well above it.
Where a single incident does real damage is repetition. 82 percent of consumers read AI generated summaries, and those are built from what recurs. One complaint about a missed appointment is noise; four in ten is a theme, and the theme is what gets repeated.
So the honest question after a mistake is not how to remove the review. It is whether the same sentence is about to appear three more times.
The recovery, as a rate
Recovery happens inside the window people read. 74 percent of consumers look for reviews from the last three months and 44 percent at the last month, so what matters is what arrives from now on.
At 3 new reviews a month, the three month window holds 9, and one bad review among nine recent ones reads as an exception. At 0 new reviews a month, the same review is 1 of very few and defines the profile. The difference between those two outcomes is entirely under your control and costs about 12 asks a month.
The immediate part is the reply. 19 percent of consumers expect a response the same day, and a short factual answer becomes part of what the summary can quote. Four sentences: acknowledge, one verifiable fact, what changes, and a direct channel by name.
| Before | After one 1 star review | Still above four stars |
|---|---|---|
| 4.5 across 20 | 4.33 | Yes |
| 4.5 across 40 | 4.41 | Yes |
| 4.5 across 100 | 4.47 | Yes |
| 4.1 across 40 | 4.02 | Just |
| 3.9 across 40 | 3.83 | No, and it was not before |
What not to do after a real mistake
Do not explain at length in public. Length reads as defence, and the reader is judging composure rather than fault.
Do not compensate for the review by asking twenty customers at once. A cluster of ratings arriving days after a complaint is the pattern platforms screen for, and it converts a bad week into a credibility problem.
Questions and answers
How much does one bad review cost?
Should I ask for it to be removed?
How do I recover?
Can I fix it faster with a burst of reviews?
Sources
- BrightLocal, Local Consumer Review Survey 2026 Thresholds, recency and response expectations, AI summary readership
- Section 5b Gesetz gegen den unlauteren Wettbewerb (UWG) Why influencing reviews carries a disclosure duty
