I Studied 1,000 Google Review Responses. Here’s What the Best Ones Have in Common.

Most of them are talking to the wrong person.

That’s the pattern. I spent time reading through Google review responses — not the reviews, the responses — and the bad ones share one quality. They’re written for the person who left the review. The defensive ones, the template ones, the ones that open with “Dear valued customer” — they treat the exchange as a private conversation between the business and the reviewer.

It isn’t private. When a business responds to a Google review, they’re not writing to one person. They’re writing to the next two hundred people who search for that business, see that exchange, and decide whether to book.

The short version: Businesses that respond to at least 25% of their reviews see a measurable revenue premium over those that don’t respond. The optimal response length is 2 to 3 sentences. Specificity matters more than warmth. And generic, template responses now actively put off more than half of potential customers — not neutral, not slightly less effective, a net negative.

Why 97% is the number you should start with

BrightLocal’s 2026 consumer survey found that 97% of people who read reviews also read business responses to those reviews. If you go back and read that sentence again slowly, it changes what the response section of your Google Business Profile actually is. It’s not a courtesy channel for managing complainers. It’s a piece of content that nearly every potential customer will encounter before deciding to visit.

The response is part of the listing. The question isn’t whether to write one.

Harvard Business School research tracking restaurants over 12 months found that businesses responding to reviews saw an 18% improvement in overall ratings — not because responding changes the reviews themselves, but because engaged businesses tend to improve their operations when they’re reading feedback at scale. The same research found a 35% revenue premium for businesses responding to 25% or more of their reviews versus those that don’t respond at all.

That premium matters because it shows up without any change to the review volume or star rating. Just the act of responding, consistently, moves the needle. The research on reviews and local SEO treats response rate as a ranking signal for the same reason — it’s a behavioral signal that Google reads as an active, well-managed business.

What makes a review response bad

The anatomy of a bad response is consistent enough that you can almost set your watch by it.

It opens with a generic thank-you (“Thank you for your wonderful feedback!”). It restates the reviewer’s experience back to them, sometimes almost word for word (“We’re so glad you enjoyed the pasta and found the service attentive!”). It ends with an invitation to return (“We hope to see you again very soon!”). It contains no information that a potential customer couldn’t have guessed, and it could have been written by literally any business on Earth.

BrightLocal found that over 50% of consumers say these responses actively put them off. Not neutral — a net negative. The response designed to sound professional ends up signaling that nobody actually read the review. Potential customers aren’t looking for warmth. They’re looking for evidence that the business is paying attention.

There’s a secondary problem with template responses: they scale the wrong signal. A business that responds to 200 reviews with variations of “Thank you for visiting, we look forward to seeing you again!” has just told every future customer that engagement is a performance, not a practice.

What the best responses actually look like

Three things separate responses that work from responses that don’t.

Specificity. The best responses reference something concrete from the review — a staff member’s name, a product, a situation that was mentioned. Not “we’re glad you enjoyed your visit” but “glad the team sorted that out on your second visit.” Specificity proves that a person read the review. It’s also the only element that makes the response useful to someone who wasn’t there — it tells potential customers something real about how the business operates.

Brevity. The research consensus puts the optimal response at 2 to 3 sentences. Beyond that, responses start reading as over-explanation. A long response to a 4-star review signals anxiety. A long response to a 5-star review signals either a bot or someone who doesn’t respect the reader’s time. Two sentences that acknowledge the specific review and add one piece of genuine information do more work than ten sentences that don’t.

The right audience. The best responses treat the reviewer as the prompt, not the audience. “We’ve updated the process since then” means almost nothing to the person who already had the bad experience. It means quite a lot to someone deciding whether to book. The best responses I came across were clearly written with that future reader in mind — not defensive, not sycophantic, just informative.

Do negative review responses actually help?

This is where the data gets counterintuitive.

Responding badly to a negative review is worse than not responding at all. A defensive response reads as confirmation of the complaint. “We have no record of this customer” reads as dismissal and often generates more negative attention. Long responses to 1-star reviews tend to make them more visible, not less.

But the Harvard Business School data found something unexpected: restaurants that responded to negative reviews saw a bigger improvement in their overall ratings over 12 months than those that only responded to positive ones. The mechanism seems to be operational — businesses reading and responding to criticism tend to fix things. But there’s also a perception effect. Responding to a 1-star review, briefly and without defensiveness, signals something to every future reader about how the business handles problems. That signal is worth something.

The formula that works: acknowledge the experience without arguing about it (“that’s not what we want for anyone who comes in”), offer a direct path to resolution (“email us at [contact] and we’ll make it right”), stop there. The response isn’t for the person who left the 1-star review. It’s for the person reading it next Tuesday before deciding where to eat.

The AI response problem

Most review management platforms now offer AI-generated responses. Some are better than others. The tell is the same as generic template responses, just faster and more consistent.

AI responses tend to share identical sentence structures across all reviews, recycle phrases like “valued customer” and “wonderful experience,” match the emotional register of the review without adding anything specific, and occasionally respond to a 1-star complaint with a level of warmth that reads as absurd. The result is technically a response but functionally identical to no response — except it now scales the template problem across every review, all at once.

The businesses using AI well in this context aren’t using it to write responses. They’re using it to triage: flagging which reviews need human attention, identifying patterns in feedback, prioritizing the ones that require a real reply. The actual writing stays human. That distinction matters more as AI responses become ubiquitous — the ones that read like a person wrote them stand out more, not less.

There’s also an AI Overviews angle worth noting. Google’s AI now reads review content to describe businesses in search results — the same AI system is also reading the responses. A response that says “we use seasonal ingredients sourced from local farms” adds useful content to the listing in a way that “we hope to see you again soon” doesn’t.

The best response I found was eleven words

A reviewer left a detailed, thoughtful 5-star review specifically naming a staff member and describing what she did to turn around what had started as a difficult experience. The response from the business was:

“Thank you. We’ll make sure she knows you said this.”

Eleven words. It acknowledges the review (specifically — “she”), it makes a concrete promise about what happens next, and it treats the reviewer as someone who did something worth acknowledging rather than an opportunity for marketing copy. Anyone reading that exchange knows exactly what kind of operation this is.

The longest responses in the set ran 180 words or more. They thanked the reviewer multiple times, summarized the review back to them, made several promises about future visits, and ended with multiple ways to stay in touch. Nothing in those responses would change a single reader’s decision to book. The eleven-word response might.

Back to the 97% figure. Almost everyone who reads reviews also reads the responses. That’s either a reason to do this well or a reason to stop doing it poorly — the typical response is doing one of those two things.

The broader research on how consumers process reviews is clear that the response section functions as a second opinion on the business — not on the review. Most businesses treat it as the former. The ones seeing the revenue premium are treating it as the latter.

That’s a different thing to optimize for. But it’s the actual game.