Tech companies and startups are familiar with protecting their inventions with patents, and their secret formulas, source code and algorithms as trade secrets.

But they may not be aware of another powerful form of IP protection in California: “Negative trade secrets” are intended to protect a company’s secret know-how gained from extensive research investment about what does not work.

Consider Thomas Edison’s quote about his lightbulb experiments: “I haven’t failed, I’ve just found 10,000 ways that won’t work.” Imagine that Edison’s assistant quit and was hired by a competitor. The former assistant’s “negative know-how” from Edison’s 10,000 failed attempts would allow his new employer to start on attempt 10,001.

But while a trade secret is a company’s intellectual property, an employee’s general knowledge, skill and experience acquired in his or her former employment is not. Where does one draw this line? Does Edison’s former assistant really have to retry all 10,000 prior failures that he knows won’t work?

In a high-profile intellectual property case regarding self-driving technology (Waymo v. Uber), Judge William Alsup asked rhetorically, “Is an engineer really supposed to get a frontal lobotomy before they go to the next job?”

The answer to this question is obviously no, but companies have other ways to protect this information. Companies and employees should bear in mind some general best practices when protecting and navigating around negative trade secrets.

Define the breadth of negative trade secrets clearly

Courts sometimes scrutinize the breadth of alleged negative trade secrets to determine if they prevent others from competing in a particular field altogether. The broader the effect of the trade secret is and the greater its preemptive effect, the more likely a court will refuse to recognize that secret.

In one case, a court found that “Plaintiff’s designation of ‘technical know-how’ regarding what does and does not work in [ … ] digital media management software is simply too nebulous a category of information to qualify for trade secret protection.” The court criticized the plaintiff for failing to “identify any specific design routes,” and instead seeking to prevent the defendants from designing any software at all.

Any company seeking to protect this type of IP should sufficiently narrow the negative trade secret’s breadth so it doesn’t overlap with an entire field or industry.

Documentation can be used protectively

Negative trade secret claims most often succeed when a company can identify specific documents or data that included negative knowledge and were taken.

Such specificity is likely what allowed Genentech’s claims to go forward in a recent pharmaceutical case against JHL. Genentech included specific allegations that the defendants “downloaded and provided to JHL hundreds of confidential Genentech documents filled with proprietary negative know-how.”

JHL argued that its protocols differed from Genentech’s, but the court said this did not foreclose JHL’s possible use of Genentech’s negative trade secrets. This negative know-how “would confer JHL the benefit of steering clear of fruitless development pathways, thereby saving precious time and resources.”

So if a pharmaceutical manufacturer can identify data that was taken, which contained failed formulas, those failed formulas could be protectable negative trade secrets.

For software companies, claims for negative know-how misappropriation may require specific examples of the failed code that was taken.

Negative trade secrets can be the flip side of positive secrets

Companies should bear in mind that courts sometimes enforce a negative trade secret as the flip side of a positive trade secret.