A trade secret is information that is actually secret, that the owner took reasonable measures to keep secret, and that is economically valuable precisely because it is secret. Those three elements come from 18 U.S.C. § 1839(3) in the federal Defend Trade Secrets Act and from § 1(4) of the Uniform Trade Secrets Act, which nearly every state has adopted. The two definitions are close enough that courts routinely treat them as one test.
Unlike a patent or a registered trademark, nothing is filed and nobody examines the claim. Status is decided after the fact, usually by a judge, in the middle of a lawsuit. That means the question is never “did I get a trade secret” but “can I prove, on this record, that all three elements were true on the day the information walked out.”
The statute’s list is deliberately sprawling: “all forms and types of financial, business, scientific, technical, economic, or engineering information,” including patterns, plans, compilations, program devices, formulas, designs, prototypes, methods, techniques, processes, procedures, and codes. It does not matter whether the information is stored physically, electronically, graphically, or in memory alone.
So the subject matter element almost never decides a case. A recipe qualifies. So does a customer list, a manufacturing tolerance, a pricing algorithm, a supplier’s identity, a source code repository, and an unfiled patent application. Negative know-how counts too, meaning the knowledge that a particular approach does not work. The comments to the UTSA say directly that information has value “from a negative viewpoint” where, for example, the results of lengthy and expensive research showing that a process will not work have real worth to a competitor who would otherwise repeat the spend.
What subject matter cannot include is anything already public. A compilation is the interesting middle case: individual data points can each be public while the assembled, curated whole is still secret, because the value lives in the selection and arrangement rather than the components.
The second element asks whether the owner “has taken reasonable measures to keep such information secret.” The word doing the work is reasonable. Courts do not require a vault. They require effort proportionate to the value of the information and the circumstances of the business.
In practice this means confidentiality agreements with employees and vendors, access controls that limit the secret to people who need it, exit interviews, marked documents, physical security, and network segregation. A five-person startup is judged against what a five-person startup can plausibly do; a pharmaceutical company is not.
This is the element most plaintiffs lose on, and they lose it for a mundane reason. The information genuinely was confidential, but it sat on an open network share, or it was shown to a prospective partner with no NDA, or every employee had access regardless of role. Secrecy in the owner’s head is not a measure. The measures have to exist before the loss, because they are proved with documents created in the ordinary course, not with testimony assembled for litigation.
The third element is the one people skip, and it is the sharpest. The information must derive independent economic value, actual or potential, from not being generally known to, and not being readily ascertainable through proper means by, another person who can obtain economic value from its disclosure or use.
Read that carefully. The value has to come from the secrecy itself. It is not enough that the information is valuable and happens to be confidential. The question is whether a competitor would gain something by learning it, and whether the owner would lose something by that competitor learning it. A document can be sensitive, embarrassing, or expensive to produce and still fail this test if knowing it confers no competitive advantage.
“Readily ascertainable by proper means” is the other half of the same element, and it is a real filter. If a competitor could assemble the same customer list from a trade directory in an afternoon, the list has no independent value from secrecy. If the formulation can be read off the product with routine analytical chemistry, the formulation is ascertainable. Courts ask how much time, cost, and effort proper means would take, not whether the defendant actually used them.
Trade secret law does not prohibit knowing the secret. It prohibits acquiring it improperly, or using or disclosing it in breach of a duty. Independent development and reverse engineering are proper means, expressly so under the UTSA and settled in Supreme Court doctrine. In Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470 (1974), the Court upheld state trade secret law against a preemption challenge in part because it leaves those routes open, so it does not function as a patent-like monopoly.
The boundary of “proper” is not the boundary of “legal,” which surprises people. In E.I. duPont deNemours & Co. v. Christopher, 431 F.2d 1012 (5th Cir. 1970), two photographers were hired by an anonymous client to take aerial photographs of a methanol plant under construction, capturing a secret process that was exposed to view from above while the roof was still unbuilt. The Fifth Circuit held that aerial photography “from whatever altitude” was an improper means, and it pointedly refused to decide whether the flight itself broke any federal aviation rules: “Regardless of whether the flight was legal or illegal in that sense, the espionage was an improper means of discovering DuPont’s trade secret.” Owners must take reasonable precautions, but as the court put it, “an impenetrable fortress is an unreasonable requirement.” Forcing DuPont to roof an unfinished plant would be “an enormous expense to prevent nothing more than a school boy’s trick.”
The single most litigated boundary is the one between an employer’s trade secrets and an employee’s own general skill and knowledge. General skill and knowledge is not protectable, full stop. What a person learns about how to do their job, including judgment, technique, and professional education acquired on someone else’s payroll, is theirs and leaves with them. This is precisely why non-compete disputes and trade secret disputes travel together: an employer that cannot identify a specific secret often reaches for a contract instead.
That leads to the procedural teeth. A plaintiff must identify the claimed secret with reasonable particularity, not gesture at categories. California codifies this in Code of Civil Procedure § 2019.210, which bars discovery until the plaintiff identifies the secret with reasonable particularity, and federal courts apply a comparable expectation. “Our manufacturing processes” is not an identification. Courts insist on specificity because without it there is no way to test whether the thing is generally known, whether reasonable measures covered it, or whether the defendant’s own knowledge predates the employment.
There is nothing to file and nothing to renew. Protection begins the moment the three elements are satisfied and continues indefinitely, which is why the Coca-Cola formula has outlived by decades any patent it could have obtained. Trade secrets are property in a constitutional sense as well: Ruckelshaus v. Monsanto Co., 467 U.S. 986 (1984), held that a trade secret can be property for Takings Clause purposes.
The asymmetry is the cost. A patent survives publication because publication is the deal. A trade secret does not survive disclosure at all. Once the information is generally known, the value from secrecy is gone, the element is unsatisfiable, and no lawsuit brings it back. Damages against whoever leaked it may still be available, but the asset itself is over.
What are the three elements of a trade secret? Under 18 U.S.C. § 1839(3) and the Uniform Trade Secrets Act, information qualifies only if all three are true. First, it is information of essentially any type. Second, the owner took reasonable measures under the circumstances to keep it secret. Third, it derives independent economic value from not being generally known or readily ascertainable by proper means. Fail any one and there is no trade secret to protect.
Can an employee’s general skill and knowledge be a trade secret? No. The skill, experience, and general knowledge a person accumulates on the job belong to that person and travel with them to the next employer. Courts consistently refuse to convert an ordinary professional education into an employer’s property. The line falls between what a worker learned how to do and specific confidential information such as a customer list, a formulation, or pricing data they carry out the door.
Does a trade secret expire? No. There is no registration and no statutory term. A trade secret lasts as long as it stays secret and keeps its independent economic value, which is why Coca-Cola’s formula still has protection more than a century on. The flip side is that it can die instantly. Public disclosure, independent discovery, or lawful reverse engineering ends the protection with no infringement claim.
Is information a trade secret if it can be reverse engineered? Only until someone actually does it. Reverse engineering from a lawfully obtained product is a proper means of acquisition, so anything readily ascertainable by taking the product apart is weakly protected at best. If a competitor could get there quickly with ordinary effort, the information is readily ascertainable and never qualified in the first place.
Going further: How to inventory and audit your trade secrets .
This page is general legal information, not legal advice, and it does not create an attorney-client relationship.
Washington's Supreme Court held that a memorized customer list can be a trade secret. The form of the information is legally irrelevant under the UTSA.
How a chocolate chip cookie recipe guarded by lock and key qualified as a Massachusetts trade secret, and why the secret element decided the case.
The Fifth Circuit held that a trade secret can live in a novel combination of publicly known elements, and that sharing information with a few partners need not destroy its secrecy.
The Fourth Circuit affirms dismissal of a DTSA complaint that defined its trade secrets three ways, holding that sweeping definitions fail the particularity that the statute's secrecy and value elements presuppose.
Judge Posner explained why a container design disclosed in a bid (conceded not to be a trade secret and outside the parties' logistics-only nondisclosure agreement) was free for the recipient to use.
The Seventh Circuit reinstated a jury's trade-secret verdict for a two-man toy startup, holding that economic value and reasonable secrecy measures are fact questions and that an oral confidentiality agreement can suffice.
The Supreme Court held that trade secrets are property protected by the Takings Clause, but that the right exists only so long as the holder guards the secret and holds reasonable expectations of confidentiality.
Judge Easterbrook held that a software company's forty-three-page, undifferentiated description of its medical-billing system failed to identify any trade secret with the specificity the law demands.
The Ninth Circuit held that a buffet chain's everyday recipes lacked novelty and secrecy-derived value, and that its loosely guarded training manuals were never reasonably protected, marking the outer boundary of what 'qualifies' under the Uniform Trade Secrets Act.
