Are NDAs unenforceable when they protect more than trade secrets? The standard answer is no. NDAs can prevent disclosure of contractually-defined “confidential” information that is shared in the course of a confidential relationship, even if it is not technically a trade secret. NDAs can, in other words, go beyond trade secrecy.
NDAs have also not traditionally been treated as contracts in restraint of trade, like noncompetes are. An NDA’s purpose is, ostensibly, just to protect secrets. Similar to trade secret law, NDAs only prevent an employee from disclosing (and using outside authorization) specifically-defined information. They don’t prohibit competition per se. NDAs are thus seen as comparatively “narrow restraints” which, all else being equal, should be preferred to noncompetes.
Or at least that is the common wisdom. Although there is some support for this viewpoint in treatises and judicial dicta, our new article, Beyond Trade Secrecy: Confidentiality Agreements That Act Like Noncompetes, shows that a growing contingent of courts across jurisdictions are finding NDAs in employment agreements to be unenforceable when they reach too far beyond trade secrecy. Even Google’s NDA was recently found unenforceable by a California court, because it did not make sure employees could use or share skills they learned at Google with prospective employers. (That said, the Google opinion is quite extreme, even compared to others we reviewed. See pp. 8-11 of the opinion, Doe v. Google, Inc., Case No. CGC-16-556034 (Cal. Super. Ct., Cty. of San Francisco, Jan. 13, 2022)).
The Federal Trade Commission recently jumped into the deep end of this swimming pool by proposing a rule that would ban noncompete agreements in employment contracts nationwide. More surprising still, the proposed rule bans what the Commission is calling “de facto” noncompetes,” such as a “non-disclosure agreement between an employer and a worker that is written so broadly that it effectively precludes the worker from working in the same field after the conclusion of the worker’s employment with the employer.”
Our article shows that there is a long history in the courts of finding NDAs are unenforceable — and not just in California. The jurisdictional differences abound. It would probably be unwise to write a nondisclosure agreement today without consulting up-to-date statutes, cases, and regulations from the relevant jurisdiction. That said, from reading the case law, a few main problems stand out.
Confidentiality agreements are far more likely to be unenforceable when they:
(1) protect information that does not constitute trade secrets, in particular by protecting public or generally known information, or information that falls within (what a court is likely to perceive as) an employee’s general knowledge skill and experience;
(2) try to cover information that the employee already knew when they started the job or lawfully gained from a third-party source; or
(3) are so excessively broad that they have the effect of a noncompete, even if they are styled as a “nondisclosure” or “confidentiality” provision. For example, this hypothetical language in a NDA would almost certainly be unenforceable: “Anything you learn at the company that is used or usable in the business is confidential and can’t ever be shared or used by you again without our permission, and there are no meaningful exceptions.”
There is also an important empirical component to our article. These agreements are themselves often kept, or even required to be kept, secret. NDA skeptics like Orly Lobel and Sharon Sandeen have noted that a threshold challenge in assessing NDAs is simply finding them. Thanks to Chris Seaman and his “army of RA’s” we have a dataset of 450 confidentiality agreements that were disclosed in trade secret litigation. These 450 contracts give a unique snapshot of what NDAs look like in practice. Some people may find the data points alarming. For example, most of the confidentiality provisions in our dataset cover far more than trade secrets, and around 40% of the agreements in our dataset had no carve-outs at all, even for public information. For people in practice, perhaps you will not be surprised at the breadth of the agreements, which is itself interesting. Either way, I hope you will check out our findings on SSRN.