INTRODUCTION 1053 I. SIMILARITY IN TRADE SECRECY 1060 A. Background 1062 B. Substantial Derivation 1064 C. Analogies to Patent Law 1069 D. Analogies to Copyright Law 1073 II. DEFINING SUBSTANTIAL DERIVATION 1077 A. Intermediate and Ultimate Uses 1080 B. Materiality 1086 C. Foreseeable Markets 1091 1. Whose Foresight? 1103 2. Foreseeability as of When? 1105 III. POSSIBLE IMPLEMENTATIONS 1107 A. Prima Facie Case 1108 B. Remedies 1111 C. Affirmative Defense 1113 CONCLUSION 1114 INTRODUCTION
Outright and forthright duplication is a dull and very rare type of infringement. --Graver Tank & Mfg. Co. v. Linde Air Prods. Co. (1) Liability for appropriating intellectual property (IP) usually does not require identical copies. All it takes is a sufficient degree of similarity between the plaintiff's and defendant's products. One of IP policy's core questions is figuring out where to draw that line of sufficiency. (2) How similar, in other words, is too similar?
For almost all IP regimes, an extensive body of case law has attempted answers, and an equally extensive volume of academic commentary has offered critiques. (3) Yet there remains a remarkable exception: trade secrecy. In trade secrecy cases, just as in other IP cases, a defendant's protected information (be it software code, a chemical formula, or a manufacturing process) can diverge from the one that the plaintiff developed. Often, a plaintiff's employee or business associate acquired that information lawfully but then left to start a different venture, modifying it into something new. Adaptation seems to occur at least as often as verbatim duplication. Indeed, a leading treatise in this area notes that "[m]ost misappropriation consists of some rather indirect exploitation of the owner's information" rather than a mere slavish copy and paste. (4)
Given such exploitation's frequency, courts should know how to think about it. Our claim in this Article is that they don't. In theory, some derivative uses are supposed to lie beyond the trade secret owner's control. Courts like to repeat the maxim that only those uses that qualify as "substantial" constitute actionable misappropriation. (5) But no one really seems to know what that term means. Without ever acknowledging it, courts have veered between cursory analogies to copyright law on the one hand and to patent law on the other. This doctrinal scavenging obscures more than it clarifies. The standard for actionable similarity in trade secrecy cases remains, fittingly but frustratingly, a secret.
Unfortunately, the upshot in many decisions is that essentially any use counts as substantial. The case law seldom investigates whether the copied information was a significant part of the plaintiff's entitlement or whether the defendant's use poses any threat of market harm. Instead, the test quickly collapses into a binary question of whether exposure to the secret educated the defendant at all, regardless of what the defendant's final product or process ends up looking like. (6)
That test is hopelessly overbroad. To start, in cases where the defendant acquired the information lawfully rather than through a wrongful act, it seems to give a windfall to owners. IP policy generally tries to avoid restrictions on downstream use that don't confer offsetting benefits to society. But such restrictions are especially pernicious in trade secrecy. In other areas of IP like copyright and patent, second comers can usually design around an upstream owner's entitlement by turning to a substitute. (7) A filmmaker unable to license a song can use a different song; a smartphone producer unable to license a chip can use a different chip. But under courts' current approach to similarity in trade secrecy, it's virtually impossible for a departing employee to find a substitute. You can't erase a secret once you know it. If that secret would provide a boost during R&D--even one that leads to a good or method radically different than the one embodying the plaintiff's secret--the employee's best alternative isn't a substitute component but to abandon the line of research altogether. As one judge recently quipped in a headline-grabbing lawsuit between Uber and Google-spinoff Waymo over driverless car technology, "Is an engineer supposed to get a frontal lobotomy before they go on to the next job?" (8)
The inattention to a similarity standard is all the more surprising given trade secrecy's exponentially increasing stakes. Trade secrecy was once a decentralized product of individual states' common law. It's now a major IP scheme. Almost all states have implemented a version of the Uniform Trade Secrets Act (UTSA), (9) and as of Congress's enactment of the Defend Trade Secrets Act of 2016 (DTSA), plaintiffs can pursue a claim under federal law as well. (10) The DTSA's passage has been called "the most significant expansion of federal law in intellectual property" since the 1940s. (11) The executive branch, for its part, has also been ramping up enforcement of criminal laws against misappropriation, from a few cases a year in the late 1990s to hundreds over the last decade. (12)
Trade secrecy law's growing supply meets an equally growing demand. One 2014 study, cited in the Senate Report accompanying the DTSA, pegged the current cost of trade secret misappropriation at somewhere between one and three percent of the U.S. gross domestic product. (13) Even before the DTSA's enactment, trade secret litigation had been growing rapidly in both state and federal court. (14) A recent analysis found that between 2001 and 2012, the number of trade secret cases adjudicated in federal court grew fourteen percent each year. (15) Consistent with these litigation rates, survey evidence confirms that trade secrecy is an enormously popular form of intellectual property protection among firms. (16) Trade secrets, unlike patents, arise by operation of law and are therefore cheap to acquire, without any government examination necessary. And so long as they remain undisclosed, they can last forever. (17)
Part of the story behind trade secrecy's rise may be a declining value in patent protection for certain inventions. Historically, whenever a particular invention has been eligible for a patent, firms have strategized whether a patent or a trade secret would best allow it to capture the invention's value. Following a series of recent Supreme Court decisions, patent-eligible subject matter has narrowed considerably, particularly for software and business methods. (18) According to some, that narrowing has been nudging firms out of the patent system and toward secrecy. (19)
Whatever the reason, trade secrets are everywhere--and growing. (20) As this body of law continues ascending within firms' IP strategies as well as courts' dockets, its doctrine must figure out what to do in the ubiquitous scenario where a defendant's product isn't exactly like the plaintiff's. As a matter of innovation policy, an employee or business partner who comes into contact with a trade secret and then ceases to work with its owner must know how to continue researching the same problem without incurring liability. The answer cannot be, as it de facto too often is, that these individuals must simply find different problems to work on. A competing firm could try to deal with the issue by setting up a so-called "clean room," walling itself off from the individual who knows the secret. But much of the time, that solution would perversely prevent experts from working on the precise line of research they know best.
There's a better way. Trade secrecy's similarity doctrine is currently asking an incomplete set of questions. It inquires almost exclusively into the defendant's innovation means, instructing factfinders to determine whether the defendant has acquired any useful knowledge from familiarity with the secret. It wrongly skips over an inquiry into the defendant's ends. A more sensible test would consider not only the defendant's benefit from knowing the secret, but also the exploitable asset, whether a product or a process, that the benefit ultimately translates into.
This Article proposes that a defendant shouldn't be liable for using a lawfully acquired secret unless it is both exploiting an asset that incorporates material elements from the owner's secret and is doing so in a market that the plaintiff actually foresaw or, given industry trends, could reasonably have foreseen. Merely relying on a secret as a launching pad for developing a genuinely dissimilar good, or operating in a remote and unanticipatable market, would remain permissible. (21) Under that standard, owners would still remain adequately insulated against competition in their core markets. Downstream users, meanwhile, would gain some more freedom to pursue cumulative innovation. Employees who know secret information but wish to build upon it would not be tethered to that same employer.
There are, however, different ways to commit misappropriation, and our proposal affects only one of them: what it means to use a secret. Misappropriation can also occur if the defendant discloses the secret or improperly acquires it in the first place. For reasons we explain in the Article's final Part, we leave these theories of liability mostly untouched. If defendants disclose the underlying secret in ways likely to destroy its value or employ improper tactics to obtain it in the first place, the way they happen to be using it shouldn't provide them any shield. Even groundbreaking adapters ought to be liable in these cases, just as they would be under current law.
Part I surveys how trade secrecy handles inexact similarity. Perhaps reflecting the lack of any framework internal to trade secret doctrine itself, courts sometimes look to a different branch of IP for guidance on assessing similarity. When they do, patent law is the usual reference point. But a patent's scope is defined ex ante by written claims, and its nonliteral...