Hiding Negative Comments: Moderation or Suppression?

One of our clients is a financial institution, who recently found themselves on the wrong side of new regulations around social media. Not long ago, the team managing one of their social accounts developed a routine of deleting direct messages in one of their social media accounts.

No one thought about the compliance matter at the time. They considered it a way to keep the inbox clear. However, a regulator noticed, and the bank was subsequently issued a formal warning, along with a chance to correct the issue before facing penalties.

Having moderated brand communities for over two decades, I’ve observed many well-intentioned organizations miss changes like this. Suppression of customer feedback has moved from a simple housekeeping matter to a potential compliance risk. In regulated industries, these situations could be a disclosure question, a record-keeping question, or both.

We’ll clarify the boundaries of these rules, identify which ones are truly relevant, and dispel some of the exaggerations surrounding this topic, so you have an accurate understanding of your responsibilities.

A new, important compliance rule

In October 2024, the Federal Trade Commission’s Rule on the Use of Consumer Reviews and Testimonials took effect. Most of the coverage focused on fake reviews, which the rule bans outright. The relevant part for social media and moderation is a section on review suppression.

This rule prohibits two things.

The first is using an unfounded legal threat, a physical threat, intimidation, or a public false accusation to stop someone from posting a review (or to get one taken down). Most brands aren’t doing this.
The second is subtler and relevant to social media teams. If you display reviews, like on your website or promotional material, you must ensure they accurately reflect the aggregate reviews submitted, not some cherry-picked subset of reviews.

The rule then lists what you are allowed to remove. You can withhold a review if it contains trade secrets or confidential information; is defamatory, harassing, abusive, obscene, vulgar or sexually explicit; contains someone else’s personal information; is discriminatory about an intrinsic characteristic; is clearly false or misleading; is reasonably believed to be fake; or is wholly unrelated to what you sell.

Then the tricky thing is that those criteria have to be “applied equally to all reviews submitted without regard to sentiment.” You’re allowed to moderate, but you aren’t allowed to wholesale remove reviews. And, crucially, you’re not allowed to let whether it’s negative be the thing that decides.

A Case Study: Fashion Nova

The FTC took issue with review suppression before this rule was in place. In 2022, it reached a settlement with Fashion Nova for $4.2 million related to a practice that seems shocking in hindsight. Reviews scoring below four stars out of five were removed from the website, presenting a picture of perfect customer satisfaction.

It marked their first case addressing concealed negative reviews and resulted in the FTC distributing $2.4 million in refunds to 148,000 affected customers. What’s striking about Fashion Nova’s case is that it didn’t create fake reviews or bully people into changing their reviews. Instead, it established a simple threshold and allowed a filter to manage the flow of feedback.

The rule defines a consumer review as appearing on a “site or platform dedicated, in whole or in part, to receiving and displaying such evaluations”. A dedicated review page or a marketplace listing clearly fits the description. However, a comment under a product launch post on Instagram is in a grey area and the FTC’s guidance does not explicitly address social comments. Is it okay to hide Facebook or Instagram comments?

Let’s look at how this might apply to your brand:

Fashion Nova faced action under the general prohibition on deception in Section 5 of the FTC Act, long before this review rule came into being. The FTC’s own Q&A on the rule makes the same point regarding another practice, asking only satisfied customers for reviews, noting it “could violate the FTC Act” even though the rule allows it.
Enforcement now carries real financial weight. In December 2025 the FTC sent warning letters to ten companies about potential violations, citing civil penalties of up to $53,088 per violation. Then, in July 2026, it finalized an order against TruHeight concerning fake and incentivized reviews: a $4 million judgment, with $750,000 payable. While that case involved fabrication rather than suppression, it shows that the FTC is paying attention to these review rules.
Comments on your social channels may function as reviews, even if unintended. “Does this actually work?” “Worst purchase I’ve made this year.” These are “reviews” and your customers perceive it as such, despite the lack of formal designation.
A screenshot doesn’t care about legal definitions. People notice when a comment disappears. They often repost it. “This brand deletes criticism” becomes a more compelling story than the original comment and risks going viral.

With all of this in mind, the sensible position is to moderate your comments as if adhering to a sentiment-neutral standard, because it’s the only standard that safeguards you across all four of these scenarios.

For a bank, a customer’s comment might be a notice

Financial services companies face challenges extending far beyond reputation management, stemming from a core issue in how they handle customer communications. Guidance on social media risk was jointly issued in 2013 by the OCC, Federal Reserve, FDIC, NCUA, and CFPB, through the FFIEC. Within that guidance is the observation that customers sometimes use social media to raise issues directly with an institution, citing examples such as error disputes under Regulation E and billing errors under Regulation Z.

Considering this from the perspective of someone managing an inbox, a message like “you charged me twice and I need to fix it” isn’t merely a complaint. Depending on its content and timing, the message could signify the beginning of a dispute, which the institution is legally bound to address. Deleting the message is discarding a formal notice. To be fair, the guidance also clarifies that institutions should properly handle messages received through channels they control, such as direct messages to official accounts.

In fairness, the same guidance is clear that institutions aren’t required to monitor and respond to every communication on the internet. The obligation is to properly manage what arrives in your earned channels, like a DM in your official Facebook account or page.

Coming back to my opening story, this is the situation that our customer had to learn from. The solution required recognizing when a customer communication falls under regulations. These messages need capture, routing, response, and record-keeping.

Be careful with hidden-word lists

Most platforms offer the ability to automatically hide comments containing specific words. The intent is to protect your community from inappropriate content, so sensible teams start with a reasonable filter list. This filter includes slurs, spam phrases, scam links, even a competitor’s promotional code. However, over time, additional words tend to accumulate. After a difficult period, someone might add “scam”, and then a string of others like “refund.”, “Lawsuit.”, “Fraud.”, “Cancel.”, “Worst.”, “BBB.”, and, for a bank, occasionally, “CFPB.”

Each addition might be justified in the moment, for some specific inappropriate comment, but collectively, these rules create a filter that unintentionally hides complaints. So, if you take away one thing from this, I urge you to review your hidden-words and keyword-filter lists on every platform, examining each entry closely to see if it catches abuse or suppresses expressions of unhappiness.

What a defensible standard looks like

This is the framework we use, built directly on the categories the FTC itself treats as legitimate grounds for removal.

Category	Examples	What to do
Remove or hide	Harassment, abuse, slurs, threats; obscene or sexually explicit content; spam, scam links, bots; someone else’s personal information; clearly false factual claims; content wholly unrelated to your brand	Remove under a written policy, applied the same way whether the comment is positive or negative. Log the reason.
Leave it up	Negative reviews of your product; frustration with your service; unflattering comparisons to competitors; criticism of a company decision	Respond if it helps. Don’t remove. This is the category a sentiment filter swallows.
Route, don’t remove	Account problems, billing disputes, refund requests, anything describing a specific error or loss; for healthcare, anything touching a patient’s care	Acknowledge publicly without specifics, move it to a private channel, and hand it to the team that can resolve it. In regulated industries, capture it before anything else happens to it.

A few useful operating rules:

Write the policy down and publish it. Community guidelines are a staple of good social media management and allow your audience to follow rules rather than moods.
Log every removal with a reason code. When someone asks why a comment disappeared (a customer, a journalist or a regulator) “it was abusive, comment removed” is an appropriate answer.
Prefer the reversible action. Hiding is usually reversible while deleting is sometimes permanent, so use the action that is appropriate for the comment.
Audit the filter lists regularly. Monthly is reasonable. Make sure every entry has a reason that’s appropriate (such as abuse, bullying, etc.)
Put a person on it, don’t fully automate. The clearly abusive and the clearly fine take care of themselves via filters and automations. But some comments are ambiguous, even if they come across as angry, they may still be a legitimate dispute.

Where automation fits

Automated tools genuinely excel at handling straightforward issues, like slurs, spam, and phishing links, and many of our clients utilize their own tooling in conjunction with ours. Attempting to manage that volume manually would be an inefficient use of resources. As I mentioned last week, automated moderation fails in two ways. A hidden-words list that has grown unchecked may quietly begin deleting complaints and look like suppression. A simple, cost-effective measure is to have someone review a sample of what your automation has removed each week, giving them the power to reinstate comments and messages as needed.

Where we land

Social media moderation practices can turn into compliance issues, completely unintended. The solution involves establishing clear, written guidelines, ensuring those who apply it understand its purpose, and maintain a record of actions.

That’s the service we provide. Our moderators operate within defined, client-specific guidelines, documenting their removals and the reasoning behind them, and escalating any potential disputes instead of deleting content outright. We monitor accounts constantly, around the clock, every day of the year, a crucial factor when a regulated notification arrives late at night or on a holiday weekend.

For a broader perspective on compliance within regulated social media platforms, consult our guide to social media compliance. Find information about our review and reputation management services at review management. And if you’d like us to review your current filter lists together, please see talk to a human here.

Talk to a human about your brand →

An extension of your team. Real judgment, real brand voice, around the clock.

FAQ

Is it legal to delete negative comments on your business’s social media?

Brands typically manage their own social channel. Removing spam, abusive content, or irrelevant posts is a generally accepted practice. However, problems surface when moderation hinges on sentiment, such as deleting comments simply because they express negativity. The FTC’s Consumer Reviews and Testimonials Rule prohibits misrepresenting displayed reviews as complete when negative feedback has been hidden, and the FTC has taken action against review suppression using their broader authority to prevent deception, most notably with a $4.2 million settlement involving Fashion Nova. While the precise application of this rule to social media comments remains somewhat unclear, a documented moderation policy that remains neutral in sentiment offers a defensible standard. The FTC rule outlines specific, legally permissible reasons for removing reviews, which provides a clearer path for brands to follow.

What reviews can a business legally remove under the FTC rule?

Businesses can withhold reviews under the FTC’s new rule if they contain trade secrets or personal information, or if they’re defamatory or clearly false. They might also remove reviews that are harassing, obscene, discriminatory, or simply unrelated to the products or services offered. However, these standards must apply equally to all reviews, regardless of whether they’re positive or negative.

Does the FTC review rule apply to Facebook and Instagram comments?

Not clearly. The rule defines consumer reviews as those appearing on a website or platform dedicated in whole or in part to receiving and displaying evaluations, and the FTC’s guidance addresses reviews on dedicated websites and platforms. However, the FTC can act against deceptive suppression under Section 5 of the FTC Act regardless of the rule, and comment sections under product posts often function as reviews in practice.

Can a bank delete customer complaints sent through social media?

Handle social media complaints with care. Joint guidance from the OCC, Federal Reserve, FDIC, NCUA, and CFPB notes consumers often use these channels to raise issues like error disputes or billing errors, both of which carry regulatory weight. Deleting social media messages is risky. Institutions should prioritize capturing and routing these messages instead, as it could mean discarding a required notice.

What should be on a social media hidden-words list?

Words and phrases that reliably signal abuse, spam, or fraud (slurs, harassment terms, known scam phrases, and phishing link patterns, for example, alongside bot signatures) shouldn’t appear on a hidden-word list. Words indicating dissatisfaction, like “refund,” “scam,” “lawsuit,” “cancel”, or even the name of a regulator, shouldn’t be included either, as filtering them essentially obscures legitimate complaints rooted in sentiment. Regular audits of these lists are crucial, along with documenting the reasoning behind each entry.