Why we stopped optimizing for Google Business Profile reviews

For most of the last decade, "more Google reviews this month than last month" was a defensible operating goal for any local business. Most of our client engagements still have a goal in that family.

We've stopped designing strategies around it. The number of new Google reviews you generate per month is no longer where the leverage is. The leverage moved.

This isn't a contrarian hot take. It's a calibration based on what the data has been telling us for two years.

What changed

Three things at roughly the same time.

1. Google's recommendation surface diversified. The Local Pack and Maps results are still important. They're no longer the only Google surface that matters. AI Overviews, "places" answer panels, and direct LLM-style answers in Search increasingly bypass the Local Pack entirely. Reviews matter to the Local Pack ranking; they matter much less to the AI Overviews surface.

2. Review-spam detection got aggressive, then over-aggressive. Operators we've worked with regularly see legitimate reviews disappear, sometimes weeks after they were posted. Operators who run review-generation campaigns frequently get flagged for "suspicious patterns" even when the reviews are organic. The signal-to-noise ratio of the new-reviews-per-month metric has degraded.

3. The competitive ceiling rose. Ten years ago, going from 8 reviews to 80 reviews was a category-changing move for a local business. Today, going from 800 reviews to 1,200 is a rounding error in a market where the top three competitors all have 2,000+. The marginal value of reviews #1,201-1,500 is negligible.

What we run instead

Three substitutions, in order of leverage.

1. Review depth and entity attribution over review count. A review that names the practitioner, mentions the specific service, and includes a recovery story or an anecdote — that review is worth roughly 10x the value of a generic 5-star with no text. We coach clients to encourage reviews that are specific. The downstream effect: those reviews get extracted as entity-level evidence by both Google's algorithms and the LLMs.

2. Review distribution across surfaces, not just Google. Reviews on Google still matter, but reviews on category-specific platforms (Healthgrades for medical, Avvo for legal, OpenTable for dining, Yelp for some categories) feed directly into the LLM training and retrieval pipelines. A balanced distribution across 3-4 surfaces beats a concentrated lead on Google alone for AI visibility.

3. Direct customer feedback loops over passive review generation. We're moving more of our clients to active feedback collection — short, structured, identifiable — that flows into the operational dashboards and gets resolved before becoming a public review. The reviews that do go public are then disproportionately positive, because the negative ones got handled in the feedback channel.

This is not about gaming reviews. It's about not letting the public review count be your main customer-experience signal when better signals exist.

What about negative reviews?

Three short rules, none of which have changed:

1. Respond to every negative review, in the customer's language, within 24 hours.
2. Take the conversation off-platform with a real contact name and direct phone/email — never "please contact our customer success team."
3. Don't ask the customer to remove the review. Ask them to update it if/when the underlying issue is resolved. Most won't. Some will. Either is fine.

This part of the playbook is unchanged for ten years. It's still right.

What this means for the operator who's been chasing review count

A few things, depending on where you are.

If you're under 50 Google reviews: keep generating reviews aggressively. The first 100 reviews matter a lot. Review count is still a real ranking signal in the Local Pack at low volumes. We have not abandoned the metric for new operators.

If you're over 500 Google reviews: stop optimizing the count. Optimize for depth, distribution, and feedback loops. Your marginal review is no longer worth the effort to generate it.

If you're somewhere in between: split the difference. Maintain the generation cadence at lower intensity, but redirect resources to entity-level structured data, practitioner-specific pages, and category-specific review surfaces.

What to do this week

Pull your last 100 reviews. Count how many name the practitioner, the service, or include a story longer than one sentence. If it's under 30%, your problem isn't review count — it's review depth.

The AI Mention Tracker will also show you whether your reviews are translating into LLM citations. The signal is striking: businesses with deep, attributed reviews show up far more in AI responses than businesses with high counts of shallow ones, controlling for everything else.