Google's review content policy, read end to end

Google’s review content policy is not one document and not one page. The rules a reviewer can break sit on a page called Prohibited & restricted content, which is item 4 of 16 in a help section. Detection is described elsewhere, removals on two further pages, appeals on a fourth. Read the whole section and the striking thing is an absence: the word reply appears nowhere in the prohibited-content page, and the list of content types it governs does not include a reply either.

That is not a technicality, because your reply is moderated against this policy anyway. Google’s Business Profile help says “Google reviews your replies: This makes sure that they follow Google’s content policies.” So a merchant is judged by a document written about contributors.

The section, not the page

The top of the help section is a page titled Maps user-generated content policy, and it states its own job: “This policy outlines the criteria for how to publish UGC on Maps and how we address inappropriate or policy-violating content.” The prohibited-content page is the first half of that sentence made specific. The rest of the section is About our policies, Posting Restrictions, two pages on removed content and restricted feature access, and an appeals page. Four of those five change what you do after a takedown.

One structural detail is worth knowing before you go looking. The 21 categories are not headings in the page’s markup. Each is a collapsible block with its own anchor name, so you can send a colleague or a support agent to one section rather than the top of a long page. Those anchors are in the page source and they are not tidy: Restricted content lives at regulated_goods_and_services, Unclear and Repetitive Content at gibberish_and_repetitive, and Dangerous content at dnagerous, misspelled.

What the policy says it covers

About our policies is the scope clause, and it is one sentence and a list. The sentence is “These policies and standards apply to all Maps user generated content, regardless of whether the content is human-generated or AI-generated content, including:” and then 10 bullets, running from Text Reviews to Metadata by way of ratings, images, videos, captions, hashtags, tags and links.

Read that list for what is not in it. No replies, no business responses, nothing posted by a business through its own profile. Whether a merchant reply is Maps user-generated content at all is the ambiguity this page could settle and does not. What it does settle is the AI question: a review written by a language model is covered by the same rules as one typed by a person, which is worth knowing if you use a tool that writes for you.

Which sections are about you

This is the most useful thing in the document and it is invisible unless you read all 21 sections in a row. Four address merchants directly. The other 17 address contributors and say nothing about what a business may write.

Section	Merchant-specific clause	The words that do it
Fake Engagement	Yes, three bullets	“We do not allow merchants or users to”
Rating Manipulation	Yes, a prohibition and a permission	“We do not allow merchants to”
Personal information	Yes, a permission	“We do allow merchants to post contact information”
Advertising & solicitation	Yes, a split list	“For merchants”
The other 17 sections	No	No merchant, owner or reply is addressed

Rating Manipulation is the densest of the four, adding a sentence that begins “When soliciting reviews, merchants should not” and covers pressure on the premises and dictated content. Adult-themed content is the near miss: it scopes a rule to a business without using the word merchant, but only as a permission about a linked website.

The consequence is precise. If a reply of yours is refused under one of those four, the policy contains language written for you and you can read the bullet you tripped. If it is refused under off-topic, impersonation or misrepresentation, that section holds no sentence addressed to a business at all, and you are left inferring how a rule about reviewers applies to a reply.

How enforcement is described

Not on the rules page. The overview page holds the mechanism: “To detect policy-violating content, our machine-learning algorithms scan contributions for signals of suspicious user activity.” What is found is then “either removed by our automated models or flagged for further review by trained operators and analysts who conduct content evaluations that might be difficult for algorithms to perform alone.”

Two sentences there should change how you read a takedown. Google says “we generally remove full posts (for example, a photo post with a text review attached) even if one part of it is in violation”, and that it may “remove UGC, restrict feature access or suspend Google Accounts that violate our policies globally and indefinitely”.

The sentence that reaches a business most directly is on the feature-access page: “If you have a Business Profile and a Maps profile attached to your Google Account, violations that occur on either profile could cause feature access restrictions across both profiles.” That page describes restrictions as following “multiple or severe violations”, not one.

Appeals, and who they are for

The appeals page is written for the contributor whose content came down, not for the business that reported it. It is candid about eligibility: “At this time, appeals are only available for certain types of content policy violations and Google Accounts.” It also warns that “References to the appeals process in notifications from Google aren’t a confirmation of eligibility.”

Two limits matter. “Content removals can only be appealed once.” And “Deleting or editing violating content after you submit an appeal cancels the appeal.” So it is not a route for a rejected reply, nor the route for getting a review taken down or reporting one as fake, which run through Business Profile tools this policy section never mentions.

What the policy does not cover

Content can go for reasons that are not policy violations. The removals page lists them: an edit that is unverifiable, content with quality issues, a place with posting restrictions, a place ineligible for Maps, and content that violates applicable law. Posting can also be switched off for a whole place or category, and Google’s examples there are not businesses: “Examples of places that may have long-term restrictions include police stations, prisons, oceans, seas, schools, children’s homes and orphanages.”

Media has separate criteria on its own page, with a line the appeals page never repeats: “Low quality media may be removed; quality removals can’t be appealed.”

And there are no numbers in it. No removal rate, no decision time, no volume of enforcement. The only timing Google publishes for a merchant is on another page and is about replies: “Replies usually take up to 10 minutes to review, but sometimes a review can take up to 30 days.” We have not measured any of this ourselves, and we will not publish a rate on somebody else’s account of their own case load.

The short version

The famous page is one item in a 16-page section. The policy names 10 content types and a reply is not one of them. Four of its 21 categories speak to merchants and 17 do not, which is why a rejection under those four is legible and a rejection under the rest is guesswork. Enforcement is machine detection with human review, documented on the overview page rather than the rules page, and it reaches your Business Profile through your Google Account. Appeals are single-shot and belong to contributors, so everything a business does after a bad review, including how to answer it, happens in tools this policy never names.