Do Employer Reviews Affect Brand Search?

A different question from software reviews

We have already looked at review platforms once, in do review sites affect what AI recommends. That guide is about B2B software buying: G2, Capterra, TrustRadius, Trustpilot, and whether a product's rating changes a “best CRM for logistics” answer. The buyer there is a purchasing committee, the entity reviewed is a product, and the review sites are selling lead generation to the vendors they rank.

This is a different market with different physics. The reviewer is a current or former employee, not a customer. The entity reviewed is the company as an employer, which is legally and reputationally a distinct thing from the company as a supplier. The reader is usually a job candidate, sometimes a journalist, occasionally an investor. And the search query is almost always the brand name on its own — not a category query — which is why this belongs next to owning page one when someone Googles you rather than next to a software comparison.

One thing does carry over unchanged: nobody has measured the effect. There is no published study in either market showing that a rating change moves a ranking, a citation or a hiring outcome.

The rich result Google built for this

The most useful evidence that employer reviews occupy brand search is that Google engineered a feature on the assumption that they do.

EmployerAggregateRating is a documented structured data type with its own page in Google's Search Central reference. The opening is plain: “If your site publishes user-generated ratings about hiring organizations, add EmployerAggregateRating structured data to your site.” The stated benefit is candidly commercial: it “can provide job seekers with ratings about a hiring organization to help them choose a job” and “also offers prominent brand placement in the enriched job search experience on Google”. The required property itemReviewed “must point to a schema.org/Organization that represents the company being rated”.

Read that from the employer's side. Google has a documented mechanism by which a third party's rating of your company is attached to your company's entity and given prominent placement in a Google experience. The employer is not a party to it. The review-snippet page even routes traffic there explicitly: “Does your site provide reviews about other employers? Use EmployerAggregateRating structured data.”

Then come the guidelines, which are the part every in-house team should read before commissioning a “show our Glassdoor rating on the careers page” project. Google's content guidelines for the type: “Users must be able to post their own ratings on your site and your site must host those user ratings.” “The number of ratings must reflect actual ratings that users provide.” “The aggregate score must be accurately derived from the provided ratings.” A technical guideline adds that the markup must describe “a specific hiring organization, not … a category or a list of items”, and gives “top 10 best places to work” as an example of what does not qualify.

And the self-serving rule on the review-snippet page closes the door properly: “If the entity that's being reviewed controls the reviews about itself, their pages that use LocalBusiness or any other type of Organization structured data are ineligible for star review feature. For example, a review about entity A is placed on the website of entity A, either directly in their structured data or through an embedded third-party widget (for example, Google Business reviews or Facebook reviews widget).” The same section says “don't aggregate reviews or ratings from other websites” and “ratings must be sourced directly from users”.

So the position is symmetrical and worth stating as a rule of thumb: the platforms can be eligible for star treatment on your brand name, and you cannot be eligible for it on your own. Displaying your rating on your careers page as ordinary content is entirely fine and often sensible. Marking it up to try to win stars is a guidelines violation with manual action attached. The broader question of what markup does and does not buy is covered in does schema markup affect citation and does article markup help coverage rank.

Do assistants cite employer review pages?

On the published evidence, not often enough to appear anywhere it would be counted — and the two platforms have made opposite technical choices about whether it can happen at all.

Start with citations. No assistant operator publishes citation shares by domain, so this rests on vendor observation, and the largest published dataset is Profound's: 680 million citations across ChatGPT, Google AI Overviews and Perplexity from August 2024 to June 2025, published 5 June 2025 and updated August 2025. Profound sells answer-engine optimisation software and has an obvious interest in the finding that citations concentrate on a short list of influenceable domains, and the numbers are reported here with that stake named. Review destinations do appear in every platform's top ten — G2 fourth in ChatGPT at 1.1%, Gartner third in Perplexity at 1.0%, Yelp fourth in Perplexity at 0.8%, Gartner fifth in Google AI Overviews at 0.7% — and Glassdoor and Indeed appear in none of them. Because the study reports only each platform's top ten, their shares sit below the reporting threshold and are unknown rather than zero. The general pattern of which domains get reached for is the subject of which sources AI assistants cite.

Now the crawler files, which are the more interesting document, because both platforms have written their reasoning into the comments.

Glassdoor's robots.txt, read on 26 September 2026, contains a section headed ####Partially Block High-Quality AI/LLM Bots. Twelve user agents are listed together — GPTBot, Google-Extended, Amazonbot, anthropic-ai, GoogleOther, Claude-Web, ClaudeBot, Perplexity, Cohere, cohere-ai, Applebot-Extended and Google-CloudVertexBot — and given Disallow: / followed by four narrow allowances: /blog/, /Award/, /About/ and /employers/. Company review pages live at /Reviews/ and are not in that list, so for these agents the reviews are off limits and the marketing pages are not. A second section, #### Block Low-Quality AI/LLM Bots, gives CCBot, Omgilibot, Omgili, FacebookBot, Bytespider, Diffbot, Youbot, FriendlyCrawler and img2dataset a bare Disallow: /. Twenty-one named agents, none allowed near a review.

The gap is the striking part. None of the agents that assistants use to fetch pages at answer time — ChatGPT-User, OAI-SearchBot, Claude-User, Claude-SearchBot, PerplexityBot, Perplexity-User, DuckAssistBot, Gemini-Deep-Research, Grok-DeepSearch — is named anywhere in Glassdoor's file. Unnamed agents fall to User-agent: *, which permits /Reviews/. Whether by policy or by not having revised the list, Glassdoor's file is aimed at training corpora and leaves live retrieval permitted. There is a further ambiguity: the token Glassdoor writes is Perplexity, which is neither of the two product tokens Perplexity documents, and under RFC 9309 a crawler matches the group whose product token matches its own — so that line may bind nothing at all. We flag it rather than resolve it.

Indeed's file makes the same distinction deliberately and lands on the opposite answer. One group is headed # Rules for traditional search engines and for retrieval/grounding bots. and lists Googlebot, Google-Extended, ChatGPT-User, OAI-SearchBot, Claude-User, Claude-SearchBot, PerplexityBot, Perplexity-User, DuckAssistBot, Gemini-Deep-Research, xAI-Grok, Grok-DeepSearch, YouBot, PetalBot, Meta-WebIndexer, Meta-ExternalFetcher, Google-Agent, Google-CloudVertexBot, Slurp, DuckDuckBot and YahooJP together, with Allow: / and a long list of path exclusions. A second group, # Rules for Foundational (Training) bots., covers GPTBot, CCBot, anthropic-ai, ClaudeBot, Meta-ExternalAgent, DeepSeekBot, GrokBot, AI2Bot, PanguBot, Timpibot, Bytespider, Diffbot and others — and gives them the same path rules rather than a site-wide disallow. Company pages under /cmp/ are permitted for both groups; what is disallowed is the machinery around them, including /cmp/*/write-review, /cmp/*/people, /cmp/*/analytics and /cmp/*/company-questions.

Two cautions before anyone builds a strategy on that contrast. First, a robots allowance is not access. Our automated requests to Indeed company pages under /cmp/ returned HTTP 403 with an Indeed “Security Check” page, and every request to Glassdoor — including its own guidelines — returned a “Humans only” interstitial. Edge defences and robots directives are separate systems, and they disagree here. Second, none of this tells us what is already inside a model's weights, only what can be fetched and cited now. The methods and their limits are set out in which AI crawlers publishers name and how to verify a crawler is real.

One structural difference from the software market is worth flagging. As we noted in the B2B guide, G2 licenses its reviews to OpenAI directly, which is a route into an assistant that no crawler file governs. We found no published licensing arrangement of that kind for Glassdoor or Indeed, and absence of a public announcement is not evidence of absence — the landscape of disclosed deals is tracked in which publishers have AI licensing deals.

What the platforms say, and why we cannot quote it

This section is shorter than it should be, and the reason is itself a finding.

Glassdoor's rules for reviewers and for employers — its Community Guidelines, its policy on employer responses, its stated position on removal requests — are the documents that would settle most of the practical questions in this guide. We could not read any of them by automated request on 26 September 2026. Every attempt, at both the help-centre article and the on-site guidelines page, returned an interstitial headed “Humans only”, explaining in seven languages that Glassdoor “has been built on the contributions of real employees and job seekers” and uses “advanced security systems to keep our site safe”. Separately, Glassdoor's own robots.txt disallows /about/guidelines/ to every crawler, including Googlebot. The platform's rules are, by design, not machine-readable.

Indeed's employer support pages behaved the same way: HTTP 403 with an “Additional Verification Required” security check, and the URL most often cited for its company-page guidelines returned a 404.

We are therefore not quoting platform policy in this guide, and readers who need it should open those pages in a browser. Three things can still be said from documents we did read. Glassdoor's robots file reserves /employers/ for the AI crawlers it otherwise blocks, which tells you which side of the marketplace its published material is aimed at. Indeed's robots file disallows /cmp/*/write-review and /cmp/*/analytics to every crawler, which is ordinary hygiene around a form and a dashboard. And both platforms are the sole arbiter of their own content: an employer's only routes are the response field the platform provides and the removal process the platform defines. The general problem of asking a publisher to take something down is treated in can a placement be removed, and it applies here with less leverage, because there is no editor.

Gating, solicitation and the FTC rule

The only rulebook we could read in full is the one that binds the employer rather than the platform.

The Federal Trade Commission announced its final Rule on the Use of Consumer Reviews and Testimonials on 14 August 2024, effective 60 days after publication in the Federal Register. In the Commission's own summary, the rule addresses reviews that misrepresent “that they are by someone who does not exist, such as AI-generated fake reviews, or who did not have actual experience with the business or its products or services”. It “prohibits businesses from providing compensation or other incentives conditioned on the writing of consumer reviews expressing a particular sentiment, either positive or negative”. It prohibits insider reviews that “fail to clearly and conspicuously disclose the giver's material connection to the business”. It prohibits a business from “misrepresenting that a website or entity it controls provides independent reviews or opinions” about a category that includes its own offerings. And on suppression, it prohibits “using unfounded or groundless legal threats, physical threats, intimidation, or certain false public accusations to prevent or remove a negative consumer review”.

Every one of those provisions is framed around a consumer review. An employee's account of working somewhere is not obviously a consumer's evaluation of a product or service, and the answer turns on definitions in 16 CFR part 465 that we could not read at source: eCFR redirected our request to an access-unblocking page on 26 September 2026, so we have the Commission's summary of the rule and not the rule's own definitions. We are not going to assert that employment reviews are in scope, and we are not going to assert that they are out. The honest position is that the rule is a clear statement of what the FTC considers deceptive conduct around reviews generally, it is a sensible floor for behaviour, and its precise application to employer-review platforms has not been settled in anything we could read. The site's wider treatment of what the FTC does require, where the ground is firmer, is in what the FTC requires for paid placement.

What follows from that in practice is narrower than the compliance question. Three common employer tactics sit on the wrong side of the line drawn by the rule's logic even if the rule itself does not reach them. Surveying staff privately and then asking only the satisfied ones to post publicly is review gating: the published set no longer reflects the surveyed set, and the count is engineered. Offering anything of value for a review — a prize draw entry, a gift card, a half-day — is conditioning a review on compensation, and every major review platform prohibits it in its own terms whether or not a regulator does. And responding to a negative review with a legal threat is the precise behaviour the suppression provision names. The reputational asymmetry is severe: a screenshot of an employer threatening a reviewer travels further and lasts longer than the review did.

There is also a plain legal hazard that has nothing to do with review rules. In the United States, discussing pay and working conditions with other employees is protected concerted activity under the National Labor Relations Act, and policies or retaliation aimed at people who post about working conditions can create exposure under labour law rather than advertising law. We are flagging the existence of that exposure, not advising on it; it is an employment-counsel question and we are not employment counsel.

What an employer can legitimately do

Given all of the above, the list of genuinely available moves is short, unglamorous and mostly not about search.

Fix the thing being described. A rating is a lagging indicator of a management problem, and it is the only reputation signal in this field that responds to the underlying reality rather than to a communications programme. Nothing else on this list will outperform it.

Respond, on the platform, in public, as yourself. An employer response is the one piece of employer-authored text that sits on the review page itself, which means it is the one piece that travels with the review into a search result or an assistant's answer. Write it for the candidate reading it in three years, not for the reviewer. Acknowledge, state what changed, do not argue about facts you cannot establish in public, and never hint at knowing who wrote it.

Publish your own crawlable material about working there. If a brand query returns three pages you do not control, the fix is not to attack those pages; it is to make sure there is something of yours worth returning. A careers page that says what the job is actually like, a published pay-and-benefits page, an engineering or operations blog with real names on it, a short document on how hiring works. This is the same logic as owning page one when someone Googles you and review and reputation signals, applied to the employment brand specifically. It is also the only route by which your own words become citable, because your own site is the only domain in this story that does not block the crawlers.

Display the rating if you want to, but do not mark it up. Google's self-serving guideline is explicit that a rating about you, hosted on your site, directly or through an embedded widget, makes those pages ineligible for the star feature. Quoting your score in prose on a careers page is normal. Wrapping it in AggregateRating is a guidelines violation.

And do not buy a rating campaign. Gating, incentives and takedown threats are the three things the FTC rule names and the platforms police, they are detectable by the platforms and by ex-employees, and the downside case — a story about a company manipulating its employee reviews — is far worse than the rating it was meant to fix.

One last measurement note, because the sales pitch usually arrives with a chart. If someone offers to improve your employer rating and quotes a figure for the effect on brand search, ask for the study. We looked for one and there is none: no published research measures whether a change in employer rating moves brand search volume, the composition of a brand-query result page, or what an assistant says when asked about the company. The related question of whether coverage moves brand search at all is examined in does coverage change brand search, and the honest answer there is also narrower than the pitch.

Methodology and limitations

This guide is built from platform documentation, live technical reads and one regulator's own summary of its rule. The structured-data findings are from Google Search Central's Employer aggregate rating (EmployerAggregateRating) structured data and Review snippet (Review, AggregateRating) structured data pages, both read on 26 September 2026; all quoted conditions and the self-serving guideline are taken verbatim from those pages. The crawler findings come from direct reads of https://www.glassdoor.com/robots.txt and https://www.indeed.com/robots.txt on the same day, with agents counted by the group they fall in and the section comments quoted as written. The citation shares are Profound's, as reported in Nick Lafferty's AI Platform Citation Patterns, published 5 June 2025 and updated August 2025, covering 680 million citations from August 2024 to June 2025; Profound sells answer-engine optimisation software, its prompt set is unpublished, and only each platform's top ten domains are reported. The FTC material is the Commission's press release of 14 August 2024 announcing the final rule. Seven limitations should be read alongside the findings. First, we did not run a ranking test: we have no measurement of how often employer review pages actually appear for brand queries, only Google's documented feature and its own description of “prominent brand placement”, and anyone quoting a percentage for that is not citing published research. Second, no published study measures whether a rating change affects brand search volume, brand-query composition or assistant output, so the causal claim in this market is unmeasured rather than disputed. Third, we could not read either platform's review policies: Glassdoor returned a “Humans only” interstitial to every automated request and disallows its own guidelines path to all crawlers, and Indeed's employer support pages returned HTTP 403. No platform policy is quoted here as a result, and that is a gap a reader with a browser can close and we could not. Fourth, robots.txt reads are point-in-time, bind only compliant agents, and say nothing about training corpora already assembled; both files are actively maintained and the counts will date. Fifth, Glassdoor's file names the token Perplexity, which matches neither product token Perplexity publishes, so whether that line binds anything is genuinely unclear under RFC 9309 and we have not resolved it. Sixth, the FTC rule's scope for employment reviews is unresolved in this guide because eCFR redirected our request for 16 CFR part 465, leaving us with the Commission's summary rather than the rule's definitions. Seventh, this is United States practice; employer-review platforms operate under different defamation, data-protection and labour regimes elsewhere, and we have not surveyed them.

Conclusion

Employer reviews occupy brand search by design, not by accident. Google documents a structured data type whose stated purpose is to attach a third party's rating of your company to your company's entity and give it “prominent brand placement”, and the same documentation makes the employer ineligible to use it about itself. What happens downstream of search is less settled: Glassdoor blocks 21 named AI crawlers from its review pages while naming none of the agents that assistants use at answer time, Indeed permits both groups explicitly, both refuse our automated requests anyway, and neither reaches any platform's published top ten most-cited domains. The measurable claims here are all about mechanism; the causal claim that an employer-reputation programme moves brand search has never been published by anyone. Which leaves an employer with the boring list: fix what the reviews describe, answer them in public in a way that ages well, publish your own account of working there on a domain that does not block crawlers, and stay well clear of gating, incentives and threats — not because a regulator has ruled on employment reviews, but because getting caught at any of the three is a bigger story than the rating.

Frequently Asked Questions

Sources

1. documentation Google Search Central, Employer aggregate rating (EmployerAggregateRating) structured data. Purpose of the type; the “prominent brand placement in the enriched job search experience” claim; required itemReviewed Organization; technical and content guidelines. Read 26 September 2026.
2. documentation Google Search Central, Review snippet (Review, AggregateRating) structured data. Routing of employer reviews to EmployerAggregateRating; the self-serving reviews guideline; prohibitions on aggregating ratings from other sites and on editor-curated ratings. Read 26 September 2026.
3. live technical read glassdoor.com/robots.txt. The “Partially Block High-Quality AI/LLM Bots” group of 12 agents with Disallow: / and allowances for /blog/, /Award/, /About/ and /employers/; the “Block Low-Quality AI/LLM Bots” group of nine; the site-wide disallow of /about/guidelines/; the absence of any live retrieval agent. Read 26 September 2026.
4. live technical read indeed.com/robots.txt. The “Rules for traditional search engines and for retrieval/grounding bots” group and the “Rules for Foundational (Training) bots” group; path exclusions including /cmp/*/write-review, /cmp/*/people, /cmp/*/analytics and /cmp/*/company-questions. Read 26 September 2026.
5. vendor Lafferty, N., 2025. AI Platform Citation Patterns: How ChatGPT, Google AI Overviews, and Perplexity Source Information. Profound, 5 June 2025, updated August 2025. 680 million citations observed August 2024 to June 2025; only each platform's top ten domains are reported. Profound sells answer-engine optimisation software.
6. regulator Federal Trade Commission, Federal Trade Commission Announces Final Rule Banning Fake Reviews and Testimonials, 14 August 2024. The Commission's own summary of the prohibited practices, including review suppression by threat or intimidation; effective 60 days after Federal Register publication. Read 26 September 2026.
7. blocked source Glassdoor, Community Guidelines (help.glassdoor.com) and Guidelines (glassdoor.com/about/guidelines/). Every automated request on 26 September 2026 returned a “Humans only” interstitial; the on-site path is additionally disallowed to all crawlers in Glassdoor's robots.txt. No Glassdoor policy text is quoted in this guide.
8. blocked source Indeed, employer support pages for company pages and reviews (indeed.com/support/employers/...). Returned HTTP 403 with an “Additional Verification Required” security check on 26 September 2026; the commonly cited company-page guidelines URL returned HTTP 404. No Indeed policy text is quoted in this guide.
9. blocked source Electronic Code of Federal Regulations, 16 CFR Part 465 — Rule on the Use of Consumer Reviews and Testimonials. Our request to ecfr.gov on 26 September 2026 was redirected to an access-unblocking page, so the rule's definitions — including whether an employment review is a “consumer review” — are not quoted here.
10. standard Koster, M., Illyes, G., Zeller, H. and Sassman, L., 2022. RFC 9309: Robots Exclusion Protocol. Product-token matching, which is why Glassdoor's Perplexity line may not bind PerplexityBot or Perplexity-User. Read 26 September 2026.