Trends in Software Licensing
Intellectual property is the principal asset of many established and emerging technology companies. The licensing of these assets is big business: American companies collected nearly $18 billion in royalties and fees from trade-in intellectual properties in 1991, roughly double the figure five years earlier. The risks of commerce to technology- and brand-dependent businesses are nearly as large. The Software Publishers Association estimates that $7.4 billion worldwide ($1.6 billion in the U.S.) in software was pirated in 1993.
Commerce in rapidly-evolving technology does not always mesh easily with commercial law or the sometimes arcane rules of intellectual property law. Counsel involved in licensing transactions thus struggle to know what licensing terms are necessary and reasonable. Clients may be over- or under-protected, and transaction costs often are unnecessarily high. This article suggests ways in which counsel can approach software licensing arrangements and reviews some practices that do and do not work.
Defining Objectives & Mapping Strategy
A licensor must first define its objectives and then develop an appropriate licensing strategy. For some, licensing is a means of funding research and development. Licensing can meet customer demand by tapping the marketing and distribution capabilities of others, enable companies to explore market opportunities at relatively little risk, and can even be used to co-opt infringers. Depending on the licensor’s goals, a license may stand on its own or be but one component of a strategic partnering arrangement, an OEM or private label agreement, a joint venture agreement or other relationship. The terms of the license agreement, the selection of markets, and the screening of prospective licensees should be driven by these considerations.
Is a License Necessary?
A license is necessary if the prospective licensee uses a product, process or service in a way that, in the absence of a license, is reserved by law to the owner of the know-how, patent, copyright or trademark inhering in the product, process or service.
A licensor must consider first what intellectual property rights are at issue in a proposed transaction. Surprisingly, many companies do not know what intellectual property rights they hold. They often hold more intangible rights than they realize, in part because different intellectual property interests can inhere in a single good, service, or process. Software programs may consist of both copyrighted material and trade secrets, and the courts increasingly recognize that some software is patentable.
The prospective licensor must next determine which particular exclusive rights need to be licensed. The Patent Act vests in the patentee the right to prevent others from making, using or selling the patented invention or process. Copyright owners hold a bundle of exclusive rights, including the right to make copies, to make modifications to the copyrighted work, to distribute the work, to publicly display the work, and to perform the work publicly. A prospective licensee should be authorized to exercise only those rights that it will use.
Purchasers do not always need a license to make use of another’s intellectual property. For example, the Copyright Act contains a number of statutory limitations providing that specified uses are not infringement, of which the most widely known is the nebulous fair use defense. Such exceptions are available far less often than most users believe.
There are few truly essential terms in licenses of intellectual property, since generally the law reserves to the owner of intellectual property all rights that are not expressly granted. The omission of some terms, however, can result in the loss of the licensor’s asset.
For example, obviously a trade secret is protectable only if it is secret, or at least the subject of reasonable efforts under the circumstances to maintain its secrecy. It always is uncertain whether the law will imply a confidential relationship from the relationship of the parties, so having a written confidentiality agreement is critical. In licensing confidential know-how, a licensor binds the licensee not to disclose the knowledge, and also to take precautions to preserve the trade secret. Such measures may include limiting access to the secret, maintaining logs and other records of use, and requiring employees to sign confidentiality agreements specifically acknowledging obligations to the licensor.
Confidentiality agreements protect material not clearly protected by other forms of intellectual property. For example, no patent rights attach to the subject matter of a pending patent application; until a patent is issued, it must be covered by a trade secret license. Software programs may include structure, organization and data that might not be entitled to copyright protection, but often can be claimed as trade secrets.
Misappropriation law permits one to reverse engineer another’s trade secret and then use it. Where a vendor wants to retain control over the structure of its software or hardware, it must prohibit reverse engineering in the license. Recent decisions have held that such prohibitions may not reach so far as prevent licensees from taking some steps necessary to legitimate business, such as determining how one product can be made to operate with another. It also is essential to define the territorial scope of the license grant. Rights in intellectual property are territorial. A U.S. patent generally confers no right to exclude in other countries; similarly, a trademark used in this country generally cannot be enforced overseas, where registration in the local country usually is a prerequisite to ownership. Copyright interests also are territorial, although since the U.S. acceded to the Berne Copyright Convention in 1989 U.S. copyright owners now enjoy automatic protection under the copyright laws of most major countries. As a rule, a licensor should not authorize the use of its asset in a jurisdiction until it has secured ownership in that jurisdiction.
Defining the Scope of the Grant
The beauty of intellectual property licensing is that it is enormously flexible. The licensor is free to parcel licensed rights to others as it wishes. It may license on an exclusive or nonexclusive basis, with the former typically featuring higher royalty rates as well as performance criteria and other obligations. It may license rights only for a particular application (or field of use), or for some but not all products incorporating the intellectual property right. It may or may not allow sublicensing. Depending on the circumstances, a licensor of software may limit authorized use to a particular site, to a particular computer, to a specified number of concurrent users on a network, or even a specified number of use sessions. A patent licensor can even reasonably restrict the use of a patented product to even a single-use. Through such selective grants, the licensor controls its marketing strategy, protects its intellectual property, and maximizes its return.
The curse of intellectual property licensing is that the license must set forth all of the foregoing terms and limitations in writing. The law recognizes oral and implied licenses of intellectual property, but unwritten grants always put the licensor’s intangible asset at risk, and invite dispute as to the scope and duration of the grant.
A software license typically lists both authorized and prohibited uses of the programs. Most software licensees have no right to make modifications or distribute copies and may copy only for their own use. Licensors should conspicuously remind licensees that they retain ownership of the software, and the licensee owns at most the device on which it is stored.
Notably, the U.S. District Court for the Eastern District of Virginia is one of only two courts to have expressly ruled that copying includes loading a program into random access memory. Licensors must determine what types of copying will occur (e.g. storing from RAM to hard disk, or to a server), then decide whether to authorize those types of uses.
Value-added resellers and licensees using software to manage manufacturing, inventory or other operations often want the authority to modify or service the software, and thus to have access to the licensor’s human-readable source code. Most licensors guard their source code zealously, as loss of control of the source code can mean loss of control of the product. Where business reasons truly necessitate some disclosure of source code, the licensor is entitled to limit access strictly (often through a professional escrow agent) and to require the licensee to employ stringent confidentiality controls.
Software licensors and licensees frequently dispute the latter’s right to receive improvements. The most common solution in negotiated agreements involving high-value software is that the licensor provides enhancements — minor improvements — at no additional charge, but may require an additional license (and fee) for upgrades — major improvements and next-generation programs. Defining the difference between the two is a vexing problem, but the distinction usually turns at least in part on additional functionality. Licensees generally expect more in the way of upgrades when they purchase customized software on which their businesses may become heavily dependent.
Where the licensee contracts for a right to make modifications or improvements, the licensor often demands some form of grant back (whether by assignment or exclusive or nonexclusive license) of such improvements.
Risk Allocation: Warranties and Disclaimers.
Software licenses usually offer little or no warranty, and are replete with disclaimers and limitations. The rationale is that no software is error-free, that the licensor has no control over the hardware and other software with which its software is used, and that the cost of repair often will exceed the unit cost of the software. At the same time, licensees naturally expect some assurance that what they buy will work. The tension is particularly acute where the licensee’s business depends on the ability of newly-developed software to perform process functions; yet, precisely because the new program is customized and integrated with an existing mix of hardware and software (which the licensor did not select and cannot control), the licensor cannot reasonably guarantee that performance specifications will be met. Demonstration copies, acceptance processes, and fixed-term service commitments are the best available solutions.
Software usually is treated as a good subject to Article 2 of the Uniform Commercial Code, including its warranty of title and its implied warranties of merchantability and fitness for a particular use. Remarkably, in their quest for brevity, many licensors will offer very restrictive express warranties but fail to conspicuously disclaim implied warranties as is required by U.C.C. section 2-316(2). Similarly, licensors too often omit merger clauses, presumably as surplus legalese, thereby opening the door to the argument that every sales representation constitutes an express warranty.
Software licensed for personal, family or household use is subject to the Magnuson-Moss Warranty Act, 15 U.S.C. section 2301 et. seq., and a licensor of such products offering any express warranty may not disclaim all implied warranties of merchantability and fitness for a particular purpose. It may limit implied warranties to the duration of its express warranty, however. Many licensors, therefore, offer only a very short-term express warranty covering only the functionality of the diskette on which the software is distributed.
Warranties should not be drafted so restrictively and limitations so broadly that whatever warranty exists is found to be unconscionable or to fail of its essential purpose. Where such is found to be the case, the disclaimers and limitations are ineffective and the licensee may avail itself of the full panoply of remedies available under the U.C.C.
Licensors often cannot fully warrant ownership and title. Software programs may incorporate software developed by others; patented devices or processes may exploit other patented technology. Thus, a licensee normally must be content with a warranty of the licensor’s right to grant the license, plus a warranty that the licensee’s use of the software in accordance with the terms of the license will not infringe third-party rights.
The warranty of title should be coupled with an indemnification clause by which the licensor holds harmless against infringement claims arising not only from the use of its software, but also any material contributed by a contractor or third-party source. Such terms squarely place the risk of infringement on the immediate licensor.
Forms of License
Especially with low-cost, off-the-shelf software, the realities of commercial distribution and cost factors preclude obtaining signed licenses. Shrink-wrap licenses long have been the solution-of-choice for most mass marketers of software, despite fears that consumers do not read the agreements and, as such, there is no real contractual assent. No decision has ruled as a matter of law that shrink-wrap licenses are unenforceable. Still, the decisions indicate that such contracts will not be enforceable if the contract is not made available, and conspicuously so, at the early stages of a transaction. Similarly, the merger clause in a shrink-wrap agreement may not be effective in such circumstances, opening the door to treating sales representations as warranties.
New technology may render shrink-wrap licenses obsolete. Many vendors of mass-market software now require purchasers to accept licenses by registering on-line, or by entering keystrokes to accept terms as they are displayed on a screen, before the program will run.
Policing Infringement & Licensor Self-Help
Most licensors assume responsibility for enforcing and defending the intellectual property rights they license. Indeed, nonexclusive patent and copyright licensees generally lack standing to enforce the rights they use under license. The licensor typically excludes from its obligations to defend and indemnify claims arising from material added by the licensee or even novel applications of the licensed goods. For example, an otherwise non-infringing device could infringe a third-party patent if the licensee, unbeknownst to the licensor, applies the device to an unusual application that is covered by the third-party patent.
Licensors have enormous difficulty detecting unauthorized use, much less bearing the legal expense of bringing infringement actions. Self-help mechanisms can be invaluable complements to a well-crafted license. Some manufacturers spend as much as $50 million per year to mark and verify the authenticity of their goods, as by applying holograms to packaging, encoding electronic fingerprints on labels, and adding unique designs and other markers on their products.
Self-help devices are increasingly popular among vendors of customized and other high-end software products. These measures can be remarkably unobtrusive. Programs can be laced with code that monitors attempts to copy or that identifies the source of any unauthorized copy. Several suppliers offer chips and plug-in devices that allow access only to authorized users.
Also effective are programs that a licensor can activate to disable the licensed program in the event the licensee breaches its obligations under the license. The licensor thus can retrieve its property without the expense and disruption of litigation — using the electronic equivalent of a repo man exercising rights under a security agreement. Such devices have been held to be a remedy available to licensors. These devices should be used with great care, however, and only in cases where a breach is unequivocal and there is no prospect of cure. A licensor assumes significant legal if deactivating the software will substantially disrupt the licensee’s business.
Avoiding Antitrust and Misuse Problems
In recent years, efforts to enforce intellectual property rights through infringement litigation often have been met with counterclaims that the plaintiff’s licensing practices violate the antitrust law and/or defends that the plaintiff misused its patent or copyright by attempting to inequitably extend the licensor’s patent or copyright monopoly.
The law is very unsettled as to which license terms offend the Sherman Act or constitute misuse. Further, the costs and risks in responding to such challenges can be enormous. At a minimum, such counterclaims and defenses substantially complicate and raise the costs of enforcement, converting a straightforward infringement claim into complex antitrust litigation. When successful, an antitrust counterclaim can expose a licensor to an award of treble damages and fees. A successful misuse defense renders all of the plaintiff’s licenses unenforceable (and the defendant’s infringement unremedied) until the plaintiff purges itself of all offending conduct.
Several court decisions dating from the days of more aggressive antitrust enforcement held that various licensing practices violate the antitrust law. Recent case law and public enforcement policy have been more lenient. In April 1995, the Antitrust Division of the Justice Department issued its Antitrust Guidelines for the Licensing of Intellectual Property, which is generally accommodating of intellectual property licenses. Yet, they are of only modest relevance to most licensors. They relate only to public enforcement policy, are heavily qualified, and largely beg the critical question of whether a licensor has market power in a relevant market, an essential element of many antitrust offenses.
The antitrust counterclaim that arises most frequently in the course of infringement litigation is that the licensor has illegally conditioned the availability of needed intellectual property on the purchase of other goods or services. In Eastman Kodak Co. v. Image Technical Services, Inc., the Supreme Court held that a genuine issue of material fact existed to support the claim of ITS, an independent service organization (ISO), that Kodak engaged in an illegal tie-in by refusing to sell proprietary Kodak copier parts to customers who obtained copier service from an ISO rather than Kodak. While Kodak did not have the market power in the market for the sale of copiers, the Court found it possible that Kodak might be unlawfully wielding market power in the aftermarket for its own brand of copiers, since its customers were then locked into Kodak equipment.
The theory of Kodak has major business and legal implications for licensors. Software vendors can earn substantial sums on maintenance contracts and upgrades, so they have a strong economic incentive to insist that licensees not use ISOs for such services. Initially, it would seem that copyright law permits the vendor simply to refuse to license customers or ISOs to use diagnostic or operating software for purposes of providing maintenance services. Kodak suggests this could be an unlawful extension of intellectual property rights. Further, if a vendor may exercise monopoly power in the aftermarket for its own brand of goods because customers are locked in, even small companies can run afoul of the antitrust laws.
So far, most courts have held that such practices do not violate the Sherman Act. In Service & Training, Inc. v. Data General Corp, for example, the Fourth Circuit held that Data General was entitled to unilaterally refuse to license its MV/ADEX diagnostic software to ISO competitors, and did not illegally tie access to the software to those who purchased service from Data General.
Licensors should take care in formulating maintenance and service policies. A licensor adopting such policies should forestall the emergence of a separate demand for independent maintenance services or service software by applying the policy from its first introduction of software and by doing so uniformly across all licensees.
A violation of the antitrust laws can constitute patent or copyright misuse, but the equitable misuse doctrine is broader. As with any equitable concept, misuse is a highly contextual doctrine. Examples of practices that have been held to constitute the affirmative defense of patent misuse (which is often asserted but seldom found) include:
- Mandatory package licenses, in which a licensee is required to accept a license to a bundle of patents, regardless of its need;
- Certain grant-back clauses, especially where the grant-back requirement imposed on a direct competitor is exclusive and applies to innovations unrelated to the licensed patent;
- Covenants not to contest a licensor’s patents, which are unenforceable;
- Patent pools in which competitors cross-license their patents on an exclusive basis, effectively forming a cartel and/or allocating markets; and
- Requiring the payment of royalties after the expiration of the patent.
A 1990 Fourth Circuit decision recently resurrected the long-dormant parallel defense of copyright misuse. In Lasercomb America, Inc. v. Reynolds, the Fourth Circuit held that Lasercomb, a licensor of a CAD/CAM program for creating machinery that creates steel rule die, misused its limited copyright monopoly by inserting in its ninety-nine year software license a non-compete provision prohibiting the licensee, during the term of the license and for one year thereafter, from writing, developing or selling competing software. The court reasoned that the non-compete clause interfered with the encouragement of individual effort and creativity, a policy inherent in the Copyright Act. Misuse occurs, it concluded, if the copyright is being used in a manner violative of the public policy embodied in the grant of a copyright. Reynolds and his employer thus avoided liability even though their infringement had been plainly established, and even though it had never signed the offending license.
Defendants subsequently have argued that any number of licensing practices violate a public policy embodied in the grant of a copyright. While most courts have declined to find misuse on the facts before them, the decisions suggest that the following practices might constitute copyright misuse:
- a requirement that a broadcaster accept a blanket license to a complete collection of musical works, wherein the licensee has no interest in most;
- failing to disclose in a copyright application that the copyrighted work was derived in part from another work; and
- prohibiting the reverse engineering of software, at least when the prohibition precludes a licensee from learning uncopyrightable elements of a computer program in order to perform necessary tasks, such as developing interoperability protocols.
To minimize the risk of a successful misuse defense, licensors should allow for limited-term licenses (perhaps subject to automatic renewal), and expressly authorize the licensee to terminate at will. Non-compete clauses may still be permissible if tied to the trade secrets in a software program rather than its copyrights. Similarly, prohibitions on reverse engineering should be tied to the protection of trade secrets, and necessary exceptions should be included. It may be prudent to require licensees to acknowledge in the license that restrictive terms the licensor, and do not unreasonably impair the creative or other work of the licensee.
The contents of this publication are intended for general information only and should not be construed as legal advice or a legal opinion on specific facts and circumstances. Copyright 2020.