Non-Disparagement Clauses in Service and Employment Contracts: What to Say, What to Avoid, and Why Courts Care

A disgruntled contractor who feels underpaid has a few options when the project wraps. They can invoice for disputed extras, file a complaint with a licensing board, or — if things went particularly sideways — post a detailed public thread describing every mistake your team made from kickoff to close-out. If your service agreement has a confidentiality clause but no non-disparagement clause, that last option is completely open. The two clauses are not the same thing, and the gap between them can cost real money.

Non-disparagement clauses restrict what the parties can say about each other after the relationship ends — or sometimes while it is still running. They appear in service agreements, consulting contracts, employment arrangements, and settlement deals. They are also among the most consistently botched provisions in standard business contracts: too broad, missing key carve-outs, structured in a way that a court will strike down, or absent entirely when the situation called for one. You can use the template library to create a base agreement and add the provisions described here, but the details of those provisions matter as much as their presence.

When the Project Ends and the Talk Begins

The dynamic is predictable. While a contract is running, both sides manage the relationship carefully. Once the engagement ends — whether by natural expiration, early termination, or a payment dispute — the incentives shift. The contractor who felt scope was abused, the client who thought the deliverable missed the mark by three miles, the employee who disagreed with a termination decision: they all have something to say, and the reach of professional networks and review platforms means that something can reach your prospective clients or future hires before you know the post exists.

Non-disparagement clauses exist to manage that post-contract communication. As a matter of basic contract law, they are enforceable in most states as an ordinary promise not to make disparaging statements in exchange for whatever the other party offered — payment, a settlement, or the opportunity to work together in the first place. The complications arise in three specific areas: the definition of "disparaging," the limits that federal labor law places on what employers can extract from employees, and the carve-outs that courts require before they will enforce the clause against someone who made a statement in a legally protected context.

This is precisely where a generic draft clause fails most consistently. A provision that says "neither party shall make disparaging remarks about the other" gives lawyers enough to argue about but leaves the parties fighting over whether a factual critical statement qualifies, whether the obligation survives termination, and whether a Yelp review counts as a "public statement" the clause was meant to cover.

What Non-Disparagement Clauses Cover — and What They Do Not

The core of a non-disparagement clause is a mutual promise: each side agrees not to make false, misleading, or derogatory public statements about the other — including about their business, products, services, employees, or management. That is the working definition courts are most likely to enforce. Understanding where the clause ends is as important as understanding where it begins.

Confidentiality clauses and non-disparagement clauses address different risks. A confidentiality clause restricts the disclosure of specific information — trade secrets, client lists, project details. A non-disparagement clause restricts what can be said publicly about a party's character, conduct, or performance, even using entirely public information. A vendor can violate a non-disparagement clause by writing a factual review describing delays and quality problems without disclosing a single confidential item. Conversely, they could leak confidential pricing data without technically disparaging anyone. The clauses operate independently.

A properly drafted clause typically covers:

• Public statements in any medium — written, oral, on review platforms, in news media, on social networks
• Statements made directly to the other party's clients, vendors, or identified business partners
• Social media posts, professional network articles, and public commentary in any format
• Oral statements to identifiable third parties with an active business relationship to the other side

What the clause does not cover, absent express language added to extend it:

• Statements made in litigation, arbitration, or regulatory agency filings
• Truthful responses to government inquiry or lawful subpoenas
• Privileged communications to an attorney for purposes of legal advice
• Internal employee communications about working conditions protected by the NLRA

The scope of "disparaging" is a critical drafting question. Courts have invalidated clauses that define disparaging to include any negative or critical statement, because that definition would prohibit truthful, factual communication. The safer standard ties "disparaging" to false or misleading statements, or to statements made with the specific intent to harm the other party — not simply to statements that happen to be unflattering.

The Legal Architecture: Contract Law, Labor Law, and State Rules

Three separate legal regimes affect the enforceability of non-disparagement clauses. Getting only one of them right is not enough, and small businesses routinely sign or send agreements that are technically non-compliant with at least one of the three layers.

Contract law is the foundational layer. A non-disparagement clause is enforceable as a contract term if it is supported by consideration, is reasonably specific, and does not violate public policy. If the clause is overly vague or impossibly broad, a court may decline to enforce it — or may reform it to something narrower. Some states permit courts to "blue-pencil" overbroad restrictive clauses; others require them to be struck in their entirety. There is no universal rule, which means the state where your contract is governed matters a great deal before you rely on this provision.

Federal labor law creates a specific overlay for employment relationships. Section 7 of the National Labor Relations Act, 29 U.S.C. § 157, protects employees' rights to engage in "concerted activities" — which the NLRB has consistently interpreted to include discussing working conditions, wages, and employer conduct with coworkers and the public as a protected form of collective communication. A non-disparagement clause that would chill an employee from speaking to a coworker about workplace grievances or discussing working conditions on social media can constitute an unfair labor practice under Section 8(a)(1) of the NLRA, regardless of whether the employee signed the clause voluntarily. This protection applies to most private-sector employees; it does not apply to independent contractors, supervisors, agricultural workers, or domestic workers.

State law adds a third layer that varies significantly by jurisdiction. California prohibits non-disparagement clauses in settlement agreements covering sexual harassment and related employment claims under Code of Civil Procedure § 1001. New York's Silenced No More Act (CPLR § 5003-c, effective 2023) similarly restricts these provisions in settlements involving harassment or discrimination. What is enforceable in Texas may be unenforceable in California with identical language.

Service Contract Language That Has Actually Held Up

For a service contract between a business and a vendor — a marketing agency, IT consultant, accountant, or any other independent contractor — the non-disparagement clause is primarily a reputational protection tool, not a labor law issue. The vendor is not an employee, so Section 7 protection for concerted activity does not apply. The clause is governed by ordinary contract law and the state rules described above.

Courts have consistently enforced mutual non-disparagement clauses in service agreements when the clause defines what "disparaging" means, identifies covered parties (not just the company but its officers and related entities), specifies covered channels (public statements, not private communications), and contains an express carve-out for truthful statements in legal proceedings. The following sample language passes judicial scrutiny in most jurisdictions and can be added to any Service Agreement template:

Non-Disparagement Clause — Service Contract Version:

"12.1 Non-Disparagement. During the term of this Agreement and for twenty-four (24) months following its expiration or termination, neither party shall make, publish, or cause to be published any statement — written, oral, or electronic — that (i) is false or misleading, or (ii) is made with the intent to harm the other party's reputation or business relationships, concerning the other party, its products, services, employees, officers, or business conduct. This restriction applies to statements on public review platforms, professional networks, social media, news media, and direct communications to the other party's identified clients or prospective clients. Nothing in this Section prohibits either party from making truthful statements in connection with any legal proceeding, government agency inquiry, or regulatory filing, or from communicating with retained legal counsel under attorney-client privilege."

What makes this version work: "disparaging" is tied to false or misleading statements and intent to harm — it does not gag parties from making true, factual, good-faith commentary. There is a defined duration (24 months). The covered channels are specified. And the legal/regulatory carve-out is explicit rather than implied. Courts that have struck down non-disparagement clauses in service contracts have almost always done so because one of these four elements was missing.

The NLRA Problem: Why Employee Agreements Require Extra Care

The same non-disparagement language that works in a service agreement between two businesses becomes a legal landmine when applied in an employment agreement. The National Labor Relations Act is the reason.

Section 7 of the NLRA protects employees' right to engage in "concerted protected activity" — which the NLRB has consistently read to include discussing wages, working conditions, and employer conduct with one another and with the public. A non-disparagement clause that would broadly prohibit an employee from saying anything negative about the company — including truthful statements about working conditions made to coworkers — can constitute an unfair labor practice under Section 8(a)(1), regardless of whether the employee voluntarily signed it. This means the mutual assent that normally makes a contract enforceable is not a complete defense in the employment context.

At minimum, any employment non-disparagement clause should include these protections if you want it to survive legal challenge:

• An explicit carve-out preserving NLRA Section 7 rights to discuss wages and working conditions
• A limitation to public statements — not internal communications between employees
• A restriction covering only false or misleading statements, not all negative commentary
• A carve-out for communications with government agencies (NLRB, EEOC, OSHA, and their state equivalents)
• A clear statement that the clause does not prohibit an employee from filing a charge with any agency

The At-Will Employment Agreement template provides a base structure where you can add a properly scoped non-disparagement provision without inadvertently violating federal labor law. The key is in how you define what the employee cannot say — not whether you include the clause at all. A narrowly drawn clause that survives NLRA scrutiny is far more valuable than a sweeping one that gets struck down the first time someone challenges it.

McLaren Macomb (2023) and the Severance Agreement Reckoning

In February 2023, the National Labor Relations Board issued its decision in McLaren Macomb, 372 NLRB No. 58. The case involved a hospital that offered furloughed employees severance agreements containing broad non-disparagement and confidentiality provisions. The NLRB held that merely offering a severance agreement with non-disparagement language broad enough to prohibit employees from discussing working conditions — even with coworkers — constituted an unfair labor practice under the NLRA, regardless of whether any employee actually signed the agreement.

The NLRB General Counsel followed with Memorandum OM 23-17 (March 2023), providing guidance on what employers must carve out. The core message: a lawful non-disparagement clause cannot restrict employees from speaking to each other, to union representatives, or to government agencies about their employment conditions. The following severance agreement carve-out reflects post-McLaren Macomb standards:

Section 7 and Regulatory Carve-Out — Employee Severance Version:

"Nothing in this Agreement shall be construed to prohibit Employee from: (a) filing a charge or complaint with, or participating in any proceeding before, the National Labor Relations Board, Equal Employment Opportunity Commission, Occupational Safety and Health Administration, or any federal, state, or local government agency; (b) engaging in concerted activity protected by Section 7 of the National Labor Relations Act, including discussing wages, hours, or working conditions with current or former coworkers; (c) making any disclosure required by law or compelled by legal process; or (d) providing truthful information to an attorney retained by Employee in connection with any matter. The Company agrees not to seek to enforce any provision of this Agreement in a manner that would interfere with Employee's rights under this paragraph."

A note on current status: with the change in NLRB leadership in 2025, enforcement priorities have shifted toward a more employer-favorable approach. General Counsel Abruzzo was removed in January 2025. As of this writing, the McLaren Macomb decision remains formally in effect as board precedent — it has not been formally overruled — but the enforcement landscape is less predictable than it was in 2023 and 2024. The safest approach is to include the carve-out language regardless, because a clause with proper carve-outs is defensible under any interpretive standard, whether the next board tightens or relaxes the rule.

Mutual vs. One-Sided: Who Stays Quiet Under Which Version

The mutual vs. one-sided distinction matters more in practice than most clients expect. In a service contract, a one-sided clause requiring only the vendor to stay quiet is unusual but not inherently problematic — if the vendor agreed in exchange for adequate compensation, courts generally enforce it. In an employment context, the analysis runs deeper.

A one-sided employment clause — where the employee agrees not to disparage the employer but retains no corresponding protection — raises enforceability concerns even apart from the NLRA. If the clause is interpreted to prohibit Section 7-protected speech, it is unenforceable regardless of what the employee signed. Mutual clauses are structurally easier to defend: if the employer is also agreeing not to disparage the employee (for instance, not to give negative references beyond confirming dates of employment), courts view the provision as a balanced business arrangement rather than an attempt to silence a potential critic.

The comparison below shows how each version operates across service and employment contexts, and which situations favor each approach.

When drafting or negotiating these provisions, consider who carries more reputational risk in the relationship. A vendor who worked on a high-profile project for a recognizable client has more to gain from a mutual clause — a bad reference from that client can damage their ability to win future work. A client who engaged a vendor for sensitive internal work may care more about the vendor's potential public commentary. Neither goal is illegitimate; the structure of the clause should reflect whose risk is actually being managed and what the other party agreed to accept in exchange.

Settlement Agreement Non-Disparagement: A Different Standard

Settlement agreements occupy a specific category. Courts examine them with heightened scrutiny because when two parties settle a dispute — especially a workplace claim — the risk that the weaker party is being pressured into a silence arrangement in exchange for compensation is real. Several state statutes directly address this concern, and they apply regardless of what the agreement's governing law clause says.

California Code of Civil Procedure § 1001 (effective January 2022, amended 2023) prohibits settlement agreements that prevent a sexual harassment, sexual assault, or workplace discrimination claimant from disclosing facts underlying the claim. The provision cannot be waived by contract. Any term purporting to prohibit such disclosure is void as against public policy, even if the settling party was represented by counsel and received substantial consideration.

New York's Silenced No More Act (CPLR § 5003-c, effective 2023) extends similar protections to claims involving harassment or discrimination on the basis of any protected characteristic under the New York State Human Rights Law. It prohibits non-disparagement clauses in settlement agreements that would prevent the claimant from speaking about the alleged conduct. For enforcement purposes, any such clause is treated as void and unenforceable — not just voidable at the claimant's option.

For the Consulting Agreement context — where the relationship is with an independent contractor rather than an employee — these state statutes may not technically apply. But if the claim being settled has any employment-law dimension, such as a misclassification dispute where the "contractor" argues they were actually an employee, the protection may extend to the settlement terms regardless of what the contract called the relationship. Check the nature of the claim, not just the nature of the original agreement, before relying on any non-disparagement provision in a settlement context.

Clauses Courts Consistently Strike Down

These clause structures appear regularly in agreements generated from outdated samples, and courts do not enforce them. Discovering this after you try to invoke the clause is not a good outcome. The patterns are consistent enough that a careful pre-signing review should catch all of them.

The most commonly rejected variations:

• Permanent obligations with no end date. Courts in most states treat perpetual restraints with skepticism. A non-disparagement clause that runs "forever" or "in perpetuity" is disproportionate to any legitimate business interest and particularly vulnerable in states that apply a reasonableness-of-duration analysis to restrictive covenants.
• "Any statement" scope with no false/misleading limitation. A clause prohibiting "any statement — positive or negative — that in any way reflects on the other party" could technically bar a vendor from mentioning that they once worked with you. Courts will not enforce it as written, and several states hold that clauses covering truthful speech are void as against public policy.
• No legal proceeding carve-out. Without an express carve-out for litigation, arbitration, or regulatory proceedings, the party invoking the clause creates the argument that the other side cannot testify truthfully in court. Courts read in a public policy exception in most jurisdictions, but express language avoids the ambiguity entirely.
• Employee clauses with no NLRA protection. After McLaren Macomb, any employee-facing clause that does not expressly preserve Section 7 rights is potentially unlawful on its face. This occured repeatedly in 2023 and 2024 as employers discovered that severance agreements they had used for years were suddenly being challenged on this exact ground.
• Provisions restricting SEC, OSHA, or EEOC communications. Federal whistleblower statutes — including Dodd-Frank, 15 U.S.C. § 78u-6, and the Sarbanes-Oxley Act, 18 U.S.C. § 1514A — prohibit contractual restrictions on reporting to federal agencies. Any clause that can be read to prevent such communications is void to that extent under federal law.

Remedies: What You Can Actually Recover When Someone Talks

Suing someone for violating a non-disparagement clause is harder than it sounds, even when the violation is obvious. The remedy challenges are real, and understanding them before you draft the clause helps you build in the provisions that actually produce consequences.

Injunctions — ordering someone to stop speaking or to delete a post — are the most obvious remedy but among the hardest to get. Courts applying the standard injunctive relief test require showing irreparable harm, likelihood of success on the merits, and a balance of hardships that tips in the movant's favor. For a single social media post that has already been read by hundreds of people, courts often decline injunctive relief on the grounds that the harm has occurred and money can compensate it. An ongoing posting campaign fares better, but even then, a judge ordering a party to delete speech faces First Amendment adjacency arguments that some courts find uncomfortable.

Liquidated damages clauses are a cleaner remedy structure. If the contract specifies a per-violation dollar amount — $5,000 per prohibited public statement, for instance — the court does not need to determine actual damages, which are notoriously difficult to prove for reputational harm. The liquidated damages clause must meet the standard test: the amount must represent a genuine pre-contract estimate of probable harm (not a penalty), and actual damages must be genuinely difficult to calculate at the time of contracting. Reputational harm is a classic case where quantification is hard at signing, which makes LD clauses highly defensible in this context. For a standard employment contract where a departing employee's post-employment commentary is a real risk, a $2,500 to $10,000 per-violation LD provision is a range that courts have accepted as a genuine estimate rather than an unenforceable penalty.

Clawback provisions are increasingly common in employment severance arrangements. If the employee violates the non-disparagement clause, they forfeit some or all of the severance payment received. Courts enforce these when structured as conditions of payment rather than as penalties — the distinction matters because some states limit penalty clauses but do not apply the same analysis to conditional payment provisions. The structure that works: "Employee's right to receive the severance payment described in Section X is conditioned upon Employee's ongoing compliance with the non-disparagement obligation in Section Y. If Employee breaches Section Y, Employer may demand return of all or a portion of any severance amounts already paid, as a debt due and owing from Employee to Employer."

Actual damages for reputational harm — lost clients, reduced revenue, decreased enterprise value — are recoverable in principle but require expert testimony and concrete evidence of causation. They are rarely worth pursuing in isolation unless the financial impact is traceable to a specific statement and a specific client decision. The realistic enforcement toolkit for most small businesses is: retraction demand first, then LD claim if retraction fails, with the clawback mechanism as leverage in the background.

State-Specific Rules: California, New York, and Beyond

California's approach is the most restrictive in the country. Beyond the § 1001 prohibition on settlement non-disparagement for protected claims, California courts apply a searching proportionality analysis to restrictive covenants that goes further than the reasonableness review in most other states. California Business & Professions Code § 16600 generally voids agreements by which anyone is restrained from engaging in a lawful profession, trade, or business. Courts have applied that principle by analogy to non-disparagement clauses that would prohibit a vendor from using truthful information about their work history to market their services. The line between impermissible restraint and permissible reputational protection is genuinely unclear under current California precedent, and California counsel is advisable before relying on these provisions there.

New York's Silenced No More Act went into effect in 2023 and applies to all employers regardless of size. Unlike the federal NLRA, which exempts supervisors, the NYSMA covers all employees. Non-disparagement provisions in settlement agreements involving harassment, discrimination, or retaliation claims are void unless the provision is at the employee's request, the employee has 21 days to consider it, and the employee retains a 7-day revocation window after signing. The practical requirement: if you are settling a New York workplace claim, these procedures are not optional formalities — skipping any of them potentially voids the non-disparagement provision even if the rest of the settlement stands.

Seperate from the statutory overlay, courts in Illinois, Washington, and Minnesota have each issued decisions in the past three years narrowing the enforceability of overbroad non-disparagement provisions in employment severance contexts. The common thread across those decisions is proportionality: a clause limited to false or misleading statements, with appropriate carve-outs, has survived challenge in all three states. A clause covering "any statement reflecting negatively on the employer" has not. The trend line in state courts is clearly toward narrowing what non-disparagement clauses can restrict, not expanding it — which means the careful drafter writes for the most restrictive jurisdiction that might apply, not the most permissive one.

The Carve-Outs You Cannot Afford to Skip

Carve-out provisions are not boilerplate filler — they are the difference between an enforceable clause and one a court voids on public policy grounds. Every non-disparagement provision, in every type of agreement, needs these exceptions explicitly stated. Courts in most jurisdictions will read in some of them as implied exceptions, but relying on implication creates litigation over what the implied exception covers. Writing them out eliminates that argument.

The carve-outs that are non-negotiable in any jurisdiction:

• Legal proceedings. Any truthful statement made in the course of litigation, arbitration, mediation, or a court filing is protected. A party cannot be held in breach for providing honest testimony or filing an accurate complaint.
• Government agency communications. Statements to the NLRB, EEOC, OSHA, SEC, or any other federal or state regulatory agency are protected by federal whistleblower statutes and cannot be restricted by contract, period.
• Attorney-client communications. Privileged communications made to an attorney for purposes of legal advice are protected regardless of whether a claim is ultimately filed.
• Required disclosures. Any disclosure required by law — court orders, subpoenas, mandatory regulatory reporting — is protected.
• NLRA Section 7 activity (employment context only). Internal employee communications about wages and working conditions cannot be reached by a non-disparagement clause without converting it into an unfair labor practice.

The standard create-and-include carve-out for any non-disparagement clause, service or employment:

Universal Carve-Out Language:

"Nothing in this Section shall be construed to prohibit either party from: (a) providing truthful information, testimony, or evidence in any legal proceeding, arbitration, mediation, court filing, or government investigation; (b) filing a charge, complaint, or report with any federal, state, or local government agency, including the NLRB, EEOC, OSHA, or Securities and Exchange Commission, or participating in any proceeding initiated by such an agency; (c) making disclosures required by applicable law or compelled by a valid legal order, subpoena, or court order; (d) communicating with retained legal counsel under attorney-client privilege; or (e) engaging in concerted activity protected by Section 7 of the National Labor Relations Act (employment relationships only). The foregoing exceptions apply regardless of any other provision of this Agreement, and no party shall be deemed in breach of this Section by reason of any communication that falls within one or more of these exceptions."

If you are using a Non-Disclosure Agreement alongside your non-disparagement clause, the two provisions should cross-reference each other to confirm they have consistent scope and definitions. Overlapping provisions that appear to contradict each other on what can and cannot be said are one of the more common drafting artifacts that ends up in a dispute over which clause controls — and that dispute wastes everyone's time and money.

Pre-Signing Checklist: What to Verify Before You Sign or Send

Most problems with non-disparagement clauses are drafting failures that could have been caught in ten minutes of careful review before the agreement went out. Run through this checklist before executing or delivering any agreement with a non-disparagement provision. The last item on the list is the one people skip most often, and it is the one that matters most.

• Scope of "disparaging" defined — does it mean false or misleading statements, or any negative statement? Courts enforce the first; they frequently void the second, and the difference between them determines whether your clause is an enforceable protection or a well-intentioned nullity.
• Duration stated — is there a specific end date? Perpetual obligations invite challenge in most jurisdictions, and a 24-month window is reasonable and standard for most service and employment relationships.
• Covered channels identified — the clause should specify public statements, not reach internal employee communications (which are NLRA-protected in employment contexts).
• All required carve-outs present — legal proceedings, government agencies, attorney communications, compelled disclosures, and (for employment) Section 7 activity.
• Remedies section coordinated — does the agreement specify what happens on breach: retraction demand, liquidated damages amount, clawback mechanism, or injunction right? A clause without remedies is a deterrent on paper only.

Beyond the checklist: a non-disparagement clause is most valuable when it is drafted carefully, not when it is brandished threateningly after someone has posted something you dislike. A court will not enjoin a truthful online review because your agreement contained vague non-disparagement language. The clause's real value is its deterrent effect — a clearly worded, properly scoped provision gives the other party a concrete reason to think twice before posting, and gives you a concrete legal basis if they do not. That deterrent only works when both parties understand exactly what the clause covers. If you have to explain to a court what you meant by it, the deterrent already failed.

Article reviewed by: Sylvia M. (Attorney)