Glassdoor review removal — how the guidelines actually decide.

Glassdoor is the only review platform where the audience is your future workforce rather than your customers, and that changes what a bad rating costs. A weak profile does not lose you a sale; it lengthens every hire, raises every offer you have to make, and gives every candidate a script of questions for the final interview. Employers feel that pressure and reach for removal, which is where most of them discover that Glassdoor's rules are narrower than they assumed and its evidentiary posture is different from any other review site.

This piece sets out what the platform's community guidelines genuinely remove, what they deliberately protect, how the employer flagging process is actually decided, and what the honest options are when a flag comes back rejected.

The two commitments that shape everything

Glassdoor's entire moderation posture follows from two positions it has held consistently and defended publicly.

The first is anonymity. Reviewers post without their identity being disclosed to the employer, and Glassdoor has resisted attempts to unmask them, including through litigation. This is not a technicality; it is the product. Any approach premised on finding out who wrote a review is fighting the platform's core commitment, and firms that claim they can identify a reviewer are describing something you should not buy — the attempt is expensive, usually unsuccessful, and reputationally far worse than the review would have been.

The second is that Glassdoor does not adjudicate truth. It removes content that breaks its guidelines. It does not run a fact-finding process to decide whether your former employee's account of their time at your company is accurate, and it will not remove a review because you disagree with it. Once you internalise that distinction, the entire strategy changes: you stop arguing that a review is unfair and start identifying whether it breaches a written rule.

What genuinely qualifies for removal

The guideline categories that actually produce removals cluster into a handful of areas.

• Naming individual non-executive employees. Glassdoor prohibits reviews that identify most individual staff members by name. Senior leadership is treated differently, as executives are considered public-facing in their role. In practice this is the single most reliable flag category, because it is a factual test — the name is either in the text or it is not.
• Harassing, discriminatory, threatening, or profane content. Personal attacks, slurs, and language targeting protected characteristics fall outside the guidelines regardless of whether the underlying complaint has merit.
• Confidential or proprietary information. Reviews disclosing trade secrets, unreleased products, client identities, or internal financial detail are removable on that basis alone.
• Personal information. Contact details, addresses, and similar identifying data about anyone are not permitted.
• Reviews from people without first-hand experience. Glassdoor's model rests on reviews from current or former employees. Competitors, rejected applicants writing as though they worked there, and contractors misrepresenting their relationship are all outside it — though establishing this requires you to bring a coherent factual account, usually supported by internal records.
• Duplicate or repeat reviews. Glassdoor limits how often one person may review the same employer, which makes a cluster of near-identical reviews from one departed employee a recognisable pattern rather than a matter of opinion.
• Reviews about the wrong entity. Content describing a different company, an acquired brand handled separately, or a franchise location misattributed to the parent.

A pattern runs through all of these: each is settled by looking at the review text or at a verifiable record, not by weighing two competing narratives. That is the test to apply before you file anything. If your argument requires the reviewer to be wrong, it is not a flag. If it requires only that a name appears, or that a policy the review describes does not exist at your company, it is.

False statements of fact — a narrow but real lane

There is one category that sits between the two. "Management is incompetent" is opinion and it stays. "The company withheld March wages" is a factual assertion that payroll records can address. Glassdoor will consider a flag that identifies a specific factual claim and explains precisely why it is false, with reference to records you hold. What gets closed without action is the generalised complaint that a review is misleading overall. Framing matters more on Glassdoor than on almost any other platform, because the reviewer's anonymity means your evidence is the only evidence in the file.

What does not qualify

The following will not be removed, and pursuing them wastes budget that would do more good elsewhere:

• A truthful, negative account of what it was like to work at your company, however painful.
• Criticism of leadership decisions, strategy, compensation, or culture expressed as the reviewer's view.
• A review from someone who genuinely worked there and left angry.
• Reviews that cluster around a real event — a layoff round, a restructuring, an acquisition — where the volume reflects what actually happened.
• An overall rating you consider unrepresentative. Ratings are aggregates of individual reviews and are not editable.

When the pattern in your reviews describes something real, removal work is the wrong spend and we will say so at the assessment stage. That is a management problem wearing a content problem's clothing, and no amount of flagging fixes it.

How the flagging process actually runs

Employers with a claimed employer profile can flag a review, select a guideline reason, and add supporting context. What happens next is a human review against the guideline you cited — not against the general impression your submission creates. Several practical consequences follow.

File individually, never in bulk. A single submission complaining about the tone of your recent reviews reads as a request to remove criticism and gets closed as one. Ten separate flags, each naming a specific clause and pointing at specific text, are ten separate decisions.

Cite the clause, not the injury. The reviewer who cannot see your internal context needs to be shown the breach in the review's own words. "This review names an individual employee in the third sentence" is a decidable statement. "This review is damaging our recruiting" is not.

Bring records before you file, not after. Where the flag rests on employment status or a factual claim, the supporting material — HR records, payroll dates, contractor agreements — needs to be assembled first. Flags filed on assertion and supplemented later tend to have been decided already.

Accept that re-filing has a ceiling. Where a first pass is rejected, re-filing once against a genuinely different clause is reasonable. Repeatedly re-submitting the same complaint is not, and in practice a decision becomes final after a single re-review.

Realistic timelines

Glassdoor does not publish a service level and outcomes vary with case complexity. The pattern worth planning around: unambiguous guideline breaches, particularly reviews naming individual employees, tend to resolve quickly, because there is nothing to weigh. Contested cases that depend on employment records or disputed factual claims take substantially longer and involve back-and-forth. Anything filed as a general objection to a review's fairness is usually resolved fast and negatively.

The timeline that matters more for a business, though, is the recovery timeline rather than the removal timeline. Even a successful removal changes an aggregate rating only slightly on a profile with any depth of history. Planning around removals alone tends to produce a disappointing quarter.

The rules that apply to you, not just to reviewers

Employers have their own set of obligations, and breaching them is the most common way a Glassdoor situation gets worse rather than better.

You may encourage employees to share their experience. You may not incentivise it, script it, select who is asked, review drafts, or pressure anyone about what they write. The distinction is between opening the door to everyone and steering the outcome. A campaign that asks the whole company, with no reward attached and no visibility into what individuals write, is legitimate. A campaign that asks only the people you expect to be positive is not, and it tends to be visible anyway — a sudden cluster of short, uniformly enthusiastic reviews in one week is a recognisable signature.

Fabricated reviews are a category error, not a shortcut. Writing reviews as though you were an employee, or paying anyone to do so, breaches the guidelines outright, is detectable through the same pattern analysis that catches review-bombing, and creates a far worse story than the reviews you were trying to bury. It also exposes the company to regulatory exposure in several jurisdictions where fake endorsements are treated as deceptive commercial practice.

Retaliation is the other trap. Attempting to identify a reviewer internally, or acting against a suspected author, converts a review problem into an employment law problem with a far higher ceiling of damage. Whatever the review says, the correct internal response is to treat the author as unknown.

"Glassdoor will remove a review that breaks a rule. It will not remove a review that breaks your heart. Nearly every wasted flag we see fails on that distinction alone."

Why the platform's culture matters here

Glassdoor's user base reads employer responses closely, and the response is the part you fully control. A defensive reply that disputes the reviewer's account line by line signals to candidates exactly what the review claimed. A short, specific reply that acknowledges the issue, names what changed, and does not argue reads as a functioning organisation. Candidates are not evaluating whether every review is true; they are evaluating how you behave when criticised in public.

This is also why volume beats surgery. A profile with a steady flow of recent reviews absorbs individual negatives; a profile with eleven reviews, three of them scathing, does not. Building an ongoing, non-selective feedback habit is slower than flagging and considerably more durable.

When removal fails

Assume the flags come back rejected. The remaining routes are real, and they are the ones that most often change the outcome for a business.

Respond publicly, on the record. Every unanswered negative review is a claim standing unopposed. Replies are read.

Rebuild the base compliantly. Open the invitation to everyone, keep it running, and let the aggregate move. This is the only lever that changes the number at the top of the page.

Address what the reviews describe. Where several independent people name the same manager behaviour, the same onboarding gap, or the same compensation issue, the reviews are a free diagnostic. Fixing the cause stops replenishment.

Work the search layer. For most employers, the Glassdoor profile is not the only problem — it is the profile plus wherever else it surfaces. Where the profile itself is holding a prominent position on your branded query, the target becomes the ranking rather than the page, and the same displacement discipline applies as anywhere else.

Treat a coordinated wave as an incident. A sudden burst of reviews after a restructuring or a public dispute is a different problem from a slow accumulation, and it is time-sensitive. Our separate write-up on review-bombing incident response covers evidence preservation and platform escalation in the window where both are still effective.

How Glassdoor differs from the platforms next to it

Employers usually deal with several review surfaces at once, and the rules are not transferable between them. Trustpilot's verification challenge can put the burden on the reviewer to produce proof of a transaction — a mechanism Glassdoor has no equivalent of, since it will not ask a reviewer to prove employment. Google's review policy turns entirely on whether the content matches a prohibited category, with no evidence exchange at all. Yelp's recommendation software adds a filtering layer that decides visibility independently of removal, and cannot be appealed. And where the discussion moves off review sites entirely — employees talking about you on Reddit, for instance — the decision-maker becomes a volunteer moderator applying a community's own rulebook.

The practical implication is that a vendor proposing one process for all of them does not understand any of them. When you are evaluating help, the questions in our buyer's checklist for reputation firms apply directly — particularly the ones about what a firm says it cannot do.

The strategic summary

Glassdoor removals succeed on clause-level arguments and fail on fairness arguments. The most productive categories are the mechanical ones — named individuals, confidential information, harassing language, reviewers without first-hand experience — because they are decided by inspection rather than by belief. Anonymity is not negotiable, truth is not adjudicated, and the aggregate rating moves through volume rather than through subtraction.

Before filing anything, get the profile classified honestly. Our Glassdoor practice starts there because the classification usually reallocates the budget: some reviews clients are certain are removable are protected opinion, and some they had accepted as permanent contain a named employee nobody flagged. The sequence we use — assessment, scope with named targets, individual filings, close-out — is set out on the Glassdoor removal page, including the point at which we tell a client to stop. If you want to see how the flags themselves get written, that is on the same page, along with what we ask HR to supply before anything is submitted.