Your Glassdoor reply is read by the people who still work for you

A one-star on Google is written by someone you will probably never meet again. A one-star on Glassdoor is written by someone who knows your payroll schedule, sat in your meetings, and has colleagues still sitting in them.

That is the whole difference, and almost every piece of advice about responding to employer reviews misses it by treating Glassdoor as Google with a different logo. It is not. Three things are structurally different: the reviewer has a relationship with you that the law protects, the review is nearly unremovable if it describes something that actually happened, and the audience for your reply is mostly people who already work for you.

Get the third one wrong and the reply costs you more than the review did.

Who is actually reading

Owners write Glassdoor replies for candidates. That is the instinct: somebody is deciding whether to apply, the review scared them, so the reply exists to reassure them.

Candidates are maybe a third of the readership and the least consequential third. The people reading most closely are the ones who still work for you, and they are reading for one thing: how you talk about someone who is no longer in the room.

A defensive reply does not read to them as a correction of the record. It reads as a preview. This is how they would describe me. This is what I would be called if I left. You cannot write that sentence out of existence afterwards, and your most mobile employees, the ones with options, are the ones most likely to be checking.

Which produces the single rule worth taking from this piece: before you post an employer-review reply, read it as if you are a current employee of three years. If any sentence would make that person update their opinion of you downward, rewrite it, even when the sentence is true.

The anonymity bargain changed, and it matters

Glassdoor was built on a promise: say what you like, nobody will know it was you. That promise got more complicated in 2024.

After integrating the workplace discussion app Fishbowl, Glassdoor began requiring users to supply a real name, workplace and job title, with the option to hide those details. The company said reviews and ratings stayed anonymous and that the identifying information was for verification. Some users reported finding their names added to their profiles without consent, and Fortune reported in March 2024 on users who found their real names attached to accounts they had used to criticise employers. At least one was told that deleting the account was the only way to remove her name.

Two consequences for you, and they point in opposite directions.

The first is that your reviewer pool got smaller and angrier. People who believe their identity might surface do not write mild reviews. They either say nothing or they wait until they have left and have something worth the risk. If your Glassdoor page went quiet and then produced two furious reviews, that is a plausible mechanism, and it is not a sign that things got worse at exactly that moment.

The second is a temptation, and I want to name it so you can refuse it deliberately. Reduced anonymity makes it likelier you can work out who wrote a review. Do not. Not out of squeamishness, but because the moment you act on that guess, you have converted a bad review into a retaliation claim, and the National Labor Relations Act protects a great deal of what employees say publicly about their working conditions. We went through that exposure in detail in whether employees can review their employer, and the short version is that the law constrains you considerably more than it constrains them.

Removal is not the game

On Google, flagging is a live option, because a review can breach a published policy without being false, and the policy grounds are reasonably broad.

Employer-review platforms work from the opposite presumption. A review describing a genuine first-hand experience of working somewhere is protected even when it is harsh, one-sided and unfair, because that is the product. Glassdoor’s terms note that employers may have the ability to respond and that Glassdoor neither endorses nor takes responsibility for those responses. Responding is the mechanism they give you. Removal is the exception.

What does get removed is narrow and mostly procedural: content that names individuals in ways the policy forbids, reviews from people who never worked there, duplicates, and material that is defamatory rather than merely damning. If your complaint is “this is unfair”, you have no case. If it is “this person was never employed here and we can show it”, you have one.

I would also not spend money on a service promising removals. The ones I have watched either flag things that were going to come down anyway or quietly run a review-solicitation campaign, which brings its own problems, below.

What a reply should actually do

Three jobs, in this order, and the order is the part people get wrong.

One: confirm that a real person read it. Not a template. The fastest way to lose the internal audience is a reply that could have been pasted under any review at any company.

Two: concede the part that is true. There almost always is one, and conceding it costs nothing while buying all your credibility for the sentence that follows.

Three: name what changed, with a date. This is the only part that does real work and the part most replies omit entirely. A current employee can verify it instantly, which means a false claim here is worse than silence.

Here is a real-shaped example. Brightwell Logistics, a 60-person freight brokerage, gets two stars from a former dispatcher:

Management promises a career path and then nothing happens for two years. Raises are decided behind closed doors and nobody will tell you the criteria. Decent people, badly run.

The reply the owner wanted to write, which I talked him out of:

We are sorry you feel this way. We have a documented promotion process available to all employees and we hold annual reviews for every team member. We wish you well in your future endeavours.

Every word defensible. It is still a bad reply, and the reason is the internal audience: every dispatcher still at Brightwell knows whether the criteria are actually published, and the reply just told them the company’s position is that everything is fine. The owner has inadvertently confirmed the review.

What went up instead:

The pay criteria point is fair and it is the one I have heard most. We did not publish how raises were decided, which meant people were guessing, and guessing about pay is corrosive. The bands and the criteria went on the intranet in September and the next review cycle runs against them. The career path part I would push back on slightly: we have moved four dispatchers into planning roles in two years, but we never said that out loud, so I understand why it looked like nothing was happening. Both of those are communication failures and they are mine.

It concedes the sharper criticism, it carries a date a current employee can check, and it disagrees once, with evidence, in a way that does not call the reviewer a liar. The disagreement is what makes the concession credible.

If you want starting structures rather than writing each one cold, replysmith.net keeps a set of templates for reviews that involve employees, and the discipline of writing your tone down before you need it is the same discipline we set out in the six decisions that make up a response voice. The employer-review addition to that list is a seventh decision: who signs. On Glassdoor an owner’s name carries weight a generic “The Brightwell Team” never will, precisely because accountability is the thing under discussion.

Indeed is a different problem wearing similar clothes

Indeed company reviews sit beside your job listings, which changes where they bite. A Glassdoor review is found by someone researching you. An Indeed review is found by someone who was already applying, at the moment they were about to apply.

That makes Indeed reviews disproportionately expensive per review, and it makes the ratio that matters different: on Glassdoor you are managing a reputation, on Indeed you are managing a conversion rate. If your applications-per-view fell off a cliff, read your Indeed reviews before you rewrite the job ad.

The response mechanics are similar enough that the three jobs above transfer intact. The difference is urgency. A six-month-old unanswered Glassdoor review is untidy. A six-month-old unanswered Indeed review is sitting next to a live vacancy you are paying to promote.

The solicitation trap

The standard advice for a bad employer rating is to ask happy employees to post. It works on the arithmetic and it has a failure mode that customer-review solicitation does not.

Asking a customer for a review is a request from a business to a stranger who can ignore it at no cost. Asking an employee is a request from someone who controls their schedule, their pay and their reference. There is no version of that ask that is free of pressure, however carefully you word it, and the people most likely to comply are the people most dependent on you.

Which produces a predictable texture: a cluster of four-and-five-star reviews posted within a fortnight, all vaguely positive, all slightly hollow. Candidates read that pattern immediately. It does not look like a happy company, it looks like a company that asked.

If you are going to do it, do it as a standing part of offboarding and onboarding rather than a campaign triggered by a bad month, make clear that not doing it has no consequence, and never mention it in a one-to-one with someone whose review you are hoping for. The compliance line here is softer than the one governing incentives for customer reviews, but the credibility line is harsher, because your audience is better at spotting it.

The part nobody tells you

Your Glassdoor rating is a lagging indicator of a management problem you can already name.

In ten years of reading these, I have never once seen an employer review surface something the leadership team did not already know. Not once. The dispatcher who wrote about pay criteria had said it in an exit interview. The complaint about the Tuesday supervisor had been in two one-to-ones. The review is not new information, it is old information that became public because the private channel did not work.

So the reply is damage control on a decision made months earlier, and the actual intervention is upstream, in whether your people believe that saying something internally changes anything. That is the same argument we make about customer reviews as operational telemetry rather than public relations, and it is more true here, not less, because the sample is people who knew you well enough to be specific.

Answer the review properly. Then go and find out who told you first.

Sourcing note. The 2024 real-name change and the reporting around it are well documented. Claims circulating about a further policy change in 2026 appear only on unattributed marketing pages and are deliberately not repeated here; check Glassdoor’s own announcements before relying on any specific current rule, and treat the mechanics above as a description of how these platforms have behaved rather than a quotation of present policy. Nothing here is legal advice, and the NLRA exposure in particular is worth real counsel.