GMU Playbook: Striking the balance with intellectual property

Mason GovCon’s sixth play for software-intensive systems seeks to protect the government’s interests while reducing its need for intellectual property. When government adopts commercial technology, vendor lock can be avoided by using best practices for interface design as a criterion for competitive survival. Most needs for data rights can be deferred to major production contracts.

Read the Draft Playbook, May 13, 2021

Background

The government’s need for adopting commercial technologies to military use cases has never been stronger. Back in 1960, DoD funded one-third of global R&D. Today, the figure is just 3 percent. This is reflected in the fact that eight of the ten USD R&E modernization priorities are dominated by commercial technologies including cyber, quantum, space, biotech, microelectronics, networking, AI/ML, and autonomy. DoD continues to lead in hypersonics and directed energy, but more than ever government must have tools for being a “fast follower” rather than a “first mover.”

A simultaneous trend has been a shift in sources of economic value from tangible to intangible capital. No longer is a company’s competitive advantage found in its physical plant, equipment, and inventory. Instead, the competition is increasingly dominated by intangibles such as software, data, and product design. Back in 1975, these intangible factors only explained 17 percent of the S&P 500’s valuation, the rest being found in tangible assets and earnings. By 2020 the intangibles accounted for 90 percent.

This explains why industry is “terrified of giving up IP rights” to government for meager revenue opportunities. Responses range from “our IP is our company” to “VCs don’t want to touch a DoD contract because the IP might get out to the competition.” IP is an important contributor in the fall of new entrants to defense contracting from 15,000 in 2010 to just 4,000 in 2019. If DoD wants the best commercial firms solving military problems, it will have to respect their primary asset – intellectual property.

At the same time, DoD must retain its unique needs for data rights. For example, most Web 2.0 firms like Facebook, Amazon, Apple, and Google have exclusive ownership of all the data their users generate. Those photos you post on social media don’t belong to you. By contrast, the DoD must have rights in critical mission data that contribute to the planning and execution of a war. Operational data is an enterprise capability, and that means accessibility to other defense organizations and even vendors.

IP Policies

Like many other aspects of acquisition, IP policies have seen a pendulum of reform. The recent standup of the DoD IP cadre and issuance of DoDI 5010.44 marks the latest swing towards increased guidance on technical and software data rights. The last major period started in 1984 with the enactment 10 USC 2320 and culminated in the 1995 rule change to the DFARS that reflects today’s regulation. In that time, the Packard Commission Report found that “suppliers have become alarmed by DoD’s increasingly vigorous pursuit of unlimited rights in technical data to be used in fostering competition.”

It seems the pursuit of IP gets exhausted due to the strains on contracting and the vendors outlasting or out-lawyering the government. By 2001, Congress held hearings on how IP policies were holding back innovation. DoD’s guide on IP management stressed commerciality, specifically negotiated rights, and using performance-based acquisition to obviate the need for data rights.

IP issues laid dormant in the 2000s. The post-9/11 tranche of defense programs did not sufficiently plan for IP. After they went through a cycle of Nunn-McCurdy breaches between 2007 to 2013, the vendor-locked programs transitioned to the field. Defense officials found themselves unable to drive competition, perform organic maintenance, and access data. The Army led the charge with a Data & Data Rights Guide in 2015, followed by the Section 813 Panel initiated in the FY 2016 NDAA.

There exists a fundamental tension in IP guidance. DFARS 227.7103 and 227.7203 for non-commercial items directs DoD to only acquire the technical data and software data rights “necessary to satisfy agency needs.” Usually this means deferring data rights acquisition to the latest point possible. By doing so, however, DoD loses leverage in negotiations. Therefore, guidance also requires lifecycle IP planning very early in program development, causing defense officials to protect themselves from uncertainty by requiring too many rights.

Application

The government has two major concerns with IP: avoidance of vendor lock and promotion of interoperability. In both of these cases, modular open systems architecture (MOSA) provides a theoretical remedy. A program could deconstruct the system into the relevant modules, create a systems map labeled according to whether it was government-funded, privately-funded, or used mixed funding. Using the open standards, government could “plug-and-play” modules, thereby improving competition and lessening the need for data rights. The use of open standards would also create enterprise efficiencies by fostering communication between systems and the sharing of components.

The problem, as we discussed before, is that there is not enough information available at the development RFP stage to pre-specify standards. It can take a long time to reach a global agreement on a standard, and even then, it’s hard to roll out across a diverse set of systems. The Air Force has used the Link 16 data link for decades and it still isn’t adopted across the inventory. Locking in new standards, moreover, will not help existing programs that make up the vast majority of the defense capabilities.

MOSA has met some success over the decades — such as the Navy’s Acoutic Rapid COTS Insertion effort, the Army’s Ground Common Infrastructure Architecture, and the Air Force’s UCI/OMS — but remains elusive as indicated by renewed emphasis in law and guidance. Government officials must often resort to acquiring intellectual property to protect themselves where MOSA could not.

Fortunately, commercial companies have made a great deal of progress breaking down large, tightly integrated software systems. This has been achieved using architecture best practices in interface design, messaging syntax, documentation, and protocols for discovery. By contrast, defense contractors often don’t know their interfaces, requiring months to reverse engineer.

With documentation in hand, DARPA has also created a suite of tools including STITCHES which can create a translation between interfaces, supporting ad hoc interoperability. This means DoD programs do not have to slow down until a global consensus on standards in enforced, but can quickly replace modules or systems on demand. As former deputy director of DARPA Dan Patt explained in plain language:

“Maybe I shouldn’t think about interoperability as trying to achieve a universal language. I shouldn’t try to define Esperanto, and force everybody to talk Esperanto. Maybe I should do on-demand translation — the Google translate equivalent… It might seem less elegant, but it’s a lot more practical. And STITCHES is the equivalent of that, but for systems.”

Mason GovCon’s sixth play for software-intensive systems seeks to protect the government’s interests while reducing its need for intellectual property. Vendor lock can be avoided by using best practices for interface design as a criterion for progression through agile development. Ordering of data rights for the interfaces and operational data can be obtained when there is more information but before competition ends.

Recommendations

Discussions with the acquisition community and additional literature research have led to the conclusion that IP practices could follow two general and perhaps overlapping tracks: one where government is the “First Mover” and another for when it is a “Fast Follower.”

“First Mover” Strategies. When government is a “first mover,” it is pulling commercial development along. Companies aren’t bringing a lot of self-funded investment to the plate because the core technologies are not dual-use. In these cases, government should make clear its intent for acquiring more extensive data rights according to guidance. This is particularly true for government-led development such as occurs in defense labs and software factories.

“Fast Follower” Strategies. As more of defense systems are built from commercial dual-use technologies, the mixed funding rule leading to government purpose rights is overly restrictive. A company could have funded 99 percent of the technology and face losing their IP if the government invests just 1 percent. Long discussions might then occur on what is meant by a “readily segregable” work element (DFARS 27.408).

When the government is acting as a “fast follower” – i.e., market research finds significant commercial capabilities that can be tailored to the defense mission – it should avoid acquiring data rights until its ready to move forward with a major contract associated with a program of record.

Treat Nontraditionals Differently. A chief complaint is of startups and commercial companies is that their self-funded IP is at greater risk than the taxpayer-funded IP investment made by traditional primes. As one small business association representative remarked: “… government practice attempts to acquire intellectual property and fails to do so in most large procurements, but does acquire intellectual property when contracting with small, innovative high-technology firms or outside commercial firms.”

One of the main reasons for this outcome is that in “first mover” strategies, traditional contractors compete for a program of record through a winner-take-all development contract. Because it takes two or more years to award, primes know the requirements and commit independent R&D funds. The government often finds itself on a tight schedule to obligate funds, choosing to move forward with IP needs unresolved.

Nontraditionals often engage with government on smaller-dollar contracts that are not tied to programs of record. They are constrained by capital and have few experts on government contracts, meaning the government has more advantage in negotiating data rights. These companies fear that a contract for experimentation will be traded for their IP, which will be handed out to competitors in the solicitation for the real program of record.

Government’s hands are not tied when contracting with nontraditionals. Flexible IP is one of the primary reasons Congress gave DoD Other Transactions authority when research and prototyping is required. Since contracts with any nontraditional business unit can use commercial item procedures (DFARS 212.102), government can follow FAR 12 procedures and commercial data rights and licensing terms (DFARS 227.7102 and 227.7202). 10 USC 2320 also favors specially negotiated licenses.

Communicating IP needs. Contract solicitations should not include blanket requirements for government purpose rights, but rather specify the government’s objectives for data rights. For example, the government may need the ability to swap an application with a competitor. The contractor could propose adopting open standards or it could provide data rights to custom interface documentation. For another example, the government may need to store data generated in military operations on a government specified cloud solution and share that data with other vendors. Hardware examples include the ability to perform organic maintenance and repair. These statements of use cases in solicitations allow the contractors to propose tailored solution that may be greater than limited/restricted rights but less than government purpose rights.

Source Selection. Most “first mover” strategies will use FAR 15 source selection procedures that are cumbersome. This includes the value adjusted total evaluated price process for determining best value. It requires adjusting the offeror’s price based on the “value” of data rights included. Yet this method is fraught with difficulties in terms of valuation methods.

By contrast, “fast follower” strategies should veer towards contract vehicles exempt from these source selection procedures, including FAR 16.505, FAR 8.4, FAR 13, and CSOs (see more in our Modular Contracting play). Comparative evaluations require no ratings and provides a high degree of flexibility in evaluation and selection. Written evaluations can be streamlined and use on-the-spot evaluations. See the DHS PIL Bootcamp for more information on these approaches (Techniques 5 and 8).

Timing Data Rights. Today’s “first mover” policies are geared toward early identification of IP needs but deferring delivery of the data and rights. Guidance suggests adding separately priced contract line item numbers to the contract. Competition in the development contract award incentivizes vendors to propose and reasonably price data rights, which the government can later exercise at its discretion. This practice works best for once-in-a-generation system that, after the development contract is awarded, loses competition.

A core idea of this playbook, particularly for software-intensive systems, is modularizing large programs and iterating quickly. This means early contract awards are no assurance of large volume ordering, but more likely initial experiments. At this stage, the government is not reliant on any contractor and usually does not need data rights. It can influence open standards and documented interfaces along the way by maintaining competition. Only when government considers making a major purchase of the capability and down-selects to a single company is the government threatened by vendor lock. A successful development strategy should delay data rights negotiations to this stage when both parties in the relationship have something to lose. The vendor faces the loss of a major production contract while government faces the loss of competitive pressures.

 Non-Disclosure Agreements. Government often needs delivery of the software code and software documentation in the early phases of development in order to conduct proper testing. Rather than acquire data rights to accomplish this kind of testing, a non-disclosure agreement (DFARS 227.7203-7) offers a model in line with commercial practices that provide a legal basis for enforcing confidentiality. This practice will help defer the acquisition of data rights, particularly for nontraditionals.

Conclusion

For software-intensive systems, government has relatively low needs for data rights if it can keep competition open and enforce industry best practices in terms of documenting interfaces. Owning the operational data is also imperative to interoperability and enterprise-wise capabilities, but these rights can be deferred until production contracts and fielding. These considerations provide government a way to engage nontraditionals as a “fast follower” rather than existing guidance better suited to its role as a “first mover.”