Check this page with an assistantOpens a chat asking it to summarise this article and name the evidence behind each claim.
Claude opens with the prompt on your clipboard: Anthropic does not support prefilled prompts on the web, and we would rather copy it than ship a button that drops it.
The Category Is Selling You Prohibited Practices
Reputation management for restaurants was built on four techniques: incentivising reviews, filtering who gets asked, suppressing negatives, and supplementing the rest. Every one of those now runs into an explicit platform prohibition, a federal rule with civil penalties, or both.
The FTC's rule on consumer reviews and testimonials took effect on 21 October 2024. It prohibits creating, selling, or buying fake reviews, procuring them from insiders, and using unfounded legal threats or intimidation to prevent or remove a negative review. It carries civil penalties per violation, and in December 2025 the Commission issued its first warning letters under it, which ended the question of whether it would be enforced.
The platform layer is separate and, for a restaurant, more immediate. Google's Maps content policy prohibits offering incentives in exchange for posting a review, discouraging or prohibiting negative reviews, selectively soliciting positive ones, and requiring or pressuring users to leave reviews while on the premises or asking that specific content be included.
| Common practice | What prohibits it | Still available to you |
|---|---|---|
| Free dessert or discount for a review | Google policy on incentives | Incentivise private feedback instead |
| Rate us 1–10, then routing by score | Google policy on selective solicitation | Ask everyone, publicly, the same way |
| Server asks guests to name a dish or staff member | Google policy on requesting specific content | Ask for their honest experience, unscripted |
| Tablet at the till, staff waiting while they type | Google policy on on-premises pressure | A card with the bill, acted on later |
| Legal threat to a negative reviewer | FTC consumer reviews rule | Platform flagging where policy is breached |
| Purchased or staff-written reviews | FTC rule and Google policy | Nothing; there is no compliant version |
Google states what is permitted just as plainly, and it is more than nothing: merchants may solicit or encourage content representing a genuine experience, without offering incentives and without attempting to influence the rating or the contents. That is a real licence to ask. It is simply a licence to ask everyone, the same way, for whatever they actually thought.
Gating Is the One Nearly Everybody Still Runs
If a tool asks how likely a guest is to recommend you, then sends high scorers to Google and low scorers to a private form, that is selective solicitation. It is the default configuration of much of the software sold into hospitality, and it is prohibited in terms that leave no interpretive room.
Gating survives because it works in the narrow sense: it demonstrably raises a public average. It also produces a rating that does not describe the restaurant, which matters more than it used to. Guests have learned to read the recent one-star reviews rather than the headline number, and assistants summarising a restaurant read the text of reviews rather than averaging stars, so a manufactured 4.9 increasingly fails at the exact moment it was built for.
The uncomfortable part of removing gating is that your average will fall. That is the point. A 4.4 that reflects the room is a working instrument, and it tells you things a gated 4.9 was constructed specifically to hide.
Reviews Are the Cheapest Operational Data You Will Ever Get
A restaurant receives, free and unprompted, a continuous stream of timestamped reports on its food, service, wait times, noise, and value, written by the people who paid. Almost every operator reads these as reputation to be managed rather than as telemetry to be processed.
The processing is where automation genuinely earns its keep, and it is a different product from the one usually sold. Extract the complaint from each review, classify it, and count it by shift. Three separate mentions of a forty-minute wait, all on Friday, is not a reputation event. It is a kitchen throughput problem visible in your reviews weeks before it shows up in covers, and it is fixable by a rota change rather than by a reply.
This is the analysis restaurants are best placed to benefit from and least likely to run, because the volume that makes it statistically useful is exactly the volume that makes reading them all feel pointless. A dentist gets a handful of reviews a month and can hold them in their head. A busy restaurant gets enough that patterns exist in them, and nobody has time to find those patterns by hand.
Draft the Reply Automatically, Send It Deliberately
Responses are the part of review handling that automation should assist and never own. A generic reply is visibly generic, and a column of identical apologies under your worst reviews does more damage than silence would have.
The workable split is drafting and triage. A system reads the review, identifies what actually went wrong, drafts a reply that names it, and routes anything involving illness, safety, discrimination, or a staff member by name to a manager who writes it personally. Everything else a human approves in seconds rather than composes in minutes.
The test for whether a reply should have been sent is whether it could have been posted under any other restaurant's review. “We're sorry to hear this and have shared your feedback with our team” passes that test, which is why it is worthless. Naming the dish, the night, and what changed is unfakeable, and it is read by every future guest who scrolls that far.
Reviews Only Matter to People Who Found You
Review work is conversion work. It changes the decision of someone already looking at your listing, which means its value is capped by how many people reach that listing and what they were searching for when they did.
Restaurants have an unusual advantage here and rarely press it. The queries that matter are highly constrained and highly local: a cuisine, a neighbourhood, a party size, a dietary requirement, a time. Those constrained queries are also what assistants answer, and they answer them from the text of reviews and profile content rather than from a star average, which is covered in local businesses in AI search.
That connects the two halves directly. Reviews mentioning the gluten-free menu, the early sitting, and the fact you take groups of twelve are the raw material those answers are built from, which is another reason to ask everyone rather than to curate. The rest of the visibility side sits in restaurant SEO, and the mechanics of asking well are in how to get more Google reviews.
Questions Restaurant Owners Ask
- Is it legal to offer a free dessert for a review?
Google's Maps policy prohibits offering incentives such as payment, discounts, or free goods and services in exchange for posting a review, so the practice violates the platform rules regardless of how it is framed. Separately, the FTC's consumer reviews rule reaches reviews that misrepresent themselves, and an incentivised review presented as an ordinary customer opinion is the pattern it addresses. A free dessert for honest feedback delivered to you directly is fine; attaching it to a public review is not.
- What exactly is review gating?
Asking customers how they would rate you first, then sending only the satisfied ones to a public review platform while routing the unhappy ones to a private feedback form. Google's policy prohibits selectively soliciting positive reviews in plain terms. It remains the default configuration of a great many reputation tools sold to restaurants, which is why it needs naming rather than describing.
- Can I ask customers for a review at the table?
You can ask without pressuring the customer or requesting specific content. Platform policies prohibit requiring ratings or reviews while guests are on the premises. A server hovering while a guest types, a tablet by the till, or a script asking for a named dish or staff member fits the prohibited pattern; a card left with the bill does not.
- Can I get a bad review removed?
Only where it violates platform policy, such as being fake, off-topic, or abusive, and the removal request has to go through the platform. What you cannot do is use unfounded legal threats or intimidation to suppress it. The FTC rule addresses review suppression directly, and a threatening letter to a reviewer is now a documented exposure rather than a hardball tactic.
- Do review responses affect visibility?
Google names responding to reviews as part of how it describes managing a profile well, and reviews in aggregate contribute to local prominence. What responses definitely do is change how the review page reads to the next person, who is making a decision on a Friday evening after skimming your three worst reviews and your replies to them. That audience is the reason to respond, whatever the ranking effect.
Primary Sources
- Google Maps: Prohibited and restricted content policy, including incentives and selective solicitation
- FTC: The Consumer Reviews and Testimonials Rule, questions and answers
- FTC: Final rule banning fake reviews and testimonials
- FTC: Soliciting and paying for online reviews, a guide for marketers
- Google Business Profile Help: Improve your local ranking
