Solutions, Garbage Cans, and Platforms: How to Drive Commercial Solutions into the Military


 “Laws and institutions must go hand in hand with the progress of the human mind. As that becomes more developed, more enlightened, as new discoveries are made, new truths discovered and manners and opinions change, with the change of circumstances, institutions must advance also to keep pace with the times.”
– Thomas Jefferson, 1816


Over the past half-decade, the Defense Department has instituted a series of defense acquisition changes to break down the barriers that separate warfighters from “nontraditional” commercially-focused companies. While privately owned Chinese technology companies are eager to sell to their country’s military forces, the situation in the United States is more complicated. Barriers, such as stale requirements, talent and cultural problems, and lengthy, opaque contracting processes, stand in the way, disincentivizing nontraditional domestic companies from doing business with the U.S. Defense Department. In response, newer defense-innovation organizations such as SOFWERX, the Defense Innovation Unit, and the National Security Innovation Network have focused on expanding the talent pool, revolutionizing contracting approaches and timelines, and developing new problem-curation processes to re-posture the Department of Defense to be more “commercial” friendly. Employed effectively, one might assume that these changes alone will break down the barriers separating domestic businesses and government.

Yet, more is needed.

The issue is that these operational changes do not directly address the larger strategic problem — the Defense Department’s dependence on a linear, requirements-first acquisition approach. This approach begins when users generate requirements, planners and programmers build a budget to pay for those requirements, and acquirers execute acquisition programs that apply that budget to meet those requirements. While the recent changes start with problems as opposed to requirements, they still follow the same linear trajectory. Organizations carefully curate problems by listening closely to user demands, scout companies to find available solutions, and acquire resource sponsors to pay for it. Either way, the underlying theory behind the approach remains constant. Start with a problem, then search for a solution.



Figure 1

The limitation inherent in a requirements-first approach is that catering solely to a user’s known problems limits an organization’s ability to recognize the extent to which externally derived (e.g., nontraditional) technical innovations solve “unknown” problems. As Clay Christensen pointed out in his seminal work on disruptive innovation, large incumbent companies often become successful by listening closely to their customers’ demands. Yet, they often perish by the same hand. Over time, by listening too closely to their customers’ perceived needs companies tie themselves to antiquated business strategies, ultimately losing to more adaptable startups who are better able to harness innovations to cater to their customers’ realized needs. Whether it is the mini-mill disrupting the steel industry or in-home content delivery disrupting home movie rentals, large incumbent companies often fail to see how their customers’ needs change given the introduction of a novel technical innovation. In turn, the U.S. Steels and Blockbusters of the world go bankrupt. As Henry Ford is said to have said, “If I had asked people what they wanted, I would have made faster horses.”

When a large, successful company is disrupted, there tend to be two critical failure points. First, as companies grow, they generate organizational inertia focused on exploiting existing technologies to meet customer needs. This inertia intensifies with corporate growth, making it increasingly difficult (and cognitively dangerous) to disregard existing ways of doing business in order to adapt to emerging technical innovations. Second, over time customers generate strong product dependencies, anchoring themselves to the company’s existing technologies as reference points to describe their problems. Thus, customers base their problems in “known knowns,” i.e., known needs and known capabilities. In combination, organizational inertia and product dependency generate both informational and cognitive asymmetries that limit the company’s ability either to source external innovations or to recognize an innovation’s potential when they do.

Relying solely on a linear requirement-to-solution approach fosters similar organizational inertia and product dependencies within defense acquisitions. Acquirers and warfighters alike minimize searching for innovative, nontraditional solutions while exploiting their existing connections, resulting in stovepipes that blind the Defense Department to external innovation. Thus, even when a new problem-driven approach discovers new technologies, the entrenched acquisition system prioritizes the purchase of “new old” solutions — buying new technologies that fit neatly into old ways of doing business — when it may have been better to purchase a wholly new capability.

And this is a big problem, given the pace of external technological change. New technology applications are, more often than not, funded and developed external to the Defense Department. Experienced entrepreneur Steve Blank places commercial advancements in technologies at two to ten times faster than those developed or acquired by the department. Therefore, even under the assumption that acquirers are using well-curated problems and searching outside their stovepipes, placing the burden of searching for, and finding, externally derived yet useful technologies solely on an acquirer’s shoulders will always be outpaced by the velocity at which potential commercially-derived solutions are introduced. As a result, the Defense Department’s current approach results in a bottleneck that restricts the ability to absorb external innovation and, therefore, work with nontraditional companies.

This is not to say that a requirement-first approach is unimportant. It certainly is important. As with large companies, defense acquisitions can be far more successful by listening closely to their users. The organizations who are currently fighting this fight are making incredible headway in delivering capabilities to the warfighter. The fundamental issue is that the requirement-first approach itself is only part of the battle.

In a commercial setting, disruptive threats to industry incumbents often come from new companies who are not burdened by existing customer demand. Instead, they start with an innovation which reframes the customer’s problem and alters their demand. The innovation may even make existing problems obsolete. Where Redbox fixed the known DVD rental problem, Netflix provided a novel streaming service that made DVD rentals obsolete.

The current requirement-first approach limits the adoption of commercial innovations because it requires the defense acquirer to find the company and alter their product, instead of allowing the company to find the user and reframe the problem. As a result, the innovative, commercially focused company sees the military as an opaque marketplace that is not available to negotiation. Rather, what the Defense Department needs is a complementary approach to requirement-first acquisitions that incentivizes the nontraditional company to engage with the military and reframe problems.

As a complement, this article offers a platform approach to acquisitions. Theoretically, a platform incentivizes solutions to find problems while problems are also looking for solutions. In the requirement-first approach, acquirers exist to find solutions for problems. In the platform approach, acquisitions professionals build and manage platforms for solutions and problems to find each other. This is done by incentivizing commercial technology companies and military users (e.g., warfighters, acquirers) to interact on the same platform, while also acting as a guiding hand to ensure that only promising matches are advanced. The platform approach can be best understood in three parts: (1) “garbage cans”; (2) network effects; and (3) incentives.

In 1972, Michael Cohen, Jim March, and Johan Olsen, world-famous organizational scholars, used “garbage cans” to illustrate how solutions seek problems just as problems seek solutions. They argued that problems and solutions exist independently, and that the ability for an organization to produce innovations (i.e., solution-problem matching) depended on the ability for both solutions and problems to find each other. They explained their position with a simple model — the garbage can. Essentially, they argued that you can view the choice opportunity of an organization as a “garbage can” filled with “various solutions and problems.” The choices available (i.e., potential solution-problem matches) depend on a number of things — the size of the can, what garbage it currently contains, the speed at which garbage is thrown in and emptied out, and so on. The insight was that the garbage can modeled an organization in which adding solutions is just as important as adding problems — that solutions can find problems just as problems can find solutions. Varying conditions, such as the pace of technical change (i.e., the speed of garbage introduction), shape the heterogenous effectiveness of that search (i.e., if many solutions are continually being introduced, they have a higher chance of finding problems).

The critical point is that as the restrictions of search (e.g., who can search, how search occurs) decrease, the probability of an optimum solution-problem match increases. For example, Access Structure A0 represents a “garbage can” in which every solution (column) can access, or find, every problem (row). A2, however, represents something closer to current defense acquisitions. In this garbage can, every solution has access only to the problems with which it is directly associated. Thus, the set of available solution-problem matches is significantly reduced.

Figures 2 & 3

Source: From a “Garbage Can Model of Organizational Choice” by Michael Cohen, Jim March, and Johan Olsen

The “garbage can” model shows that solution-product matches become more effective as the structure restricting search decreases. This requires, however, an efficient way to reduce restrictions. Reducing restrictions through simply adding more acquirers, for example, requires significant resources and would incur a significant cost to the taxpayer.

Enter network effects. The key is to create a network that allows nontraditional solution providers to search for problems while military problem owners are seeking solutions. By allowing solutions and problems to search for each other, the number of available “network nodes” (i.e., problems and solutions) doubles. The powerful thing about networks, however, is that as the number of nodes doubles, the connections between nodes increase quadratically, increasing the likelihood of matches exponentially. Such network effects make solution-problem matches both more likely and more efficient.

Yet, as powerful as network effects are, they require participants. The garbage can model assumes solutions and problems are engaging in simultaneous discovery. This leads to the last important insight: incentives. Incentives are critical because they encourage network participation. For example, some credit card companies give cardholders 5 percent cash back. Why? Because it adds cardholders to the network and therefore drives value to the retailer, which, in turn, drives value for the credit card company. As much as we’d like to believe that credit card companies are good companies that just want us to be happy, “free money” is offered to incentivize our network participation.

In total, we can consider the combined mechanisms as a platform approach to acquisitions. In our context, a platform is defined as an incentive-backed, multi-sided network that allows solutions to find problems as problems look for solutions. In turn, platforms are better able to absorb commercially derived solutions via increasing the probability of finding solution-problem matches.

Think of Amazon. Amazon is great not because it’s a marketplace, but because it’s a platform. Yes, consumers may think of it only as a market because they log on to Amazon to consume. Amazon itself, however, is a third party that matches consumers with problems, producers with solutions. Airbnb, Uber, and Tinder are all different platform companies. Their successes were found in using the three principles above: (1) that solutions need easy ways to find problems; (2) that network effects quadratically increase the number of potential problem-solution matches; and (3) that incentives are required to increase the numbers of problems and solutions on the platform. The modern platform company is changing the way we consume, but it also changing the way we supply, and in doing so it is removing the information and cognitive asymmetries present in daily life.

Today, the Defense Department’s linear acquisition process is not a platform. It is built to support problems trying to find solutions, which makes it exceedingly difficult for solutions to find and reframe problems. However, acquisitions can be a platform. At least a portion of defense spending should focus on a platform approach to acquisition. We need a fresh perspective that requires the government to take the view of the solution-providing, nontraditional technology company who is uncertain about their potential role in national security. By taking this perspective, with the intent of driving interactions to the warfighter, the acquisition community can remake itself in the mold of a modern platform company. Phrased another way, instead of acting as the intermediary, we can remold the acquirer’s role to act as the catalyst, connector, and curator of the solution-problem matching process.

Figure 4Case in point: the Defense Department’s Small Business Innovative Research (SBIR) program. This program has been the focus of much of my recent research on the impact of public funding on startup growth, and so it is a program I am both well familiar with and a big supporter of. However, the traditional program approach limits itself to a small set of companies — specifically, those companies who are both well versed in the requirements-first acquisition process and uniquely specialized to meet highly specified requirements. Currently, the program takes a requirement-first approach by publishing problems (i.e., topics) against which they award contracts for innovative small companies to solve. These topics are written by military organizations who use the opportunity to publish their requirements with the hope that companies will submit proposals to solve them. Last year, the Defense Department published nearly 1,000 unique topics — 1,000 problems to solve. How is a company, unfamiliar with the defense acquisition system, supposed to sort through 1,000 problems to find the one that they can solve, and do so in an efficient manner? How much flexibility is there to reframe the problem to better fit the solution? And what is a company’s incentive to undertake all of the work to do so?

There are simple steps for the program to start to act as a platform. For example, instead of posting 1,000 “problem” topics, post one “open topic.” The open topic would allow any company with a commercial solution to apply for funding to find problems to solve. Given some preliminary award metrics (e.g., commercial viability of the company, applicability to national security), an acquisition office could fund companies to perform customer discovery. It would give the innovative small companies a chance to see how their novel solution might reframe, and be reframed by, military users’ problems, and thus be more acceptable to those companies seeking to introduce external innovation to the Defense Department.

In June of 2018, the Air Force allowed our small team to experiment with an open topic. Building from the strategic intent of then-Secretary of the Air Force Heather Wilson and Dr. Will Roper — the assistant secretary of the Air Force for acquisition, technology and logistics — to become a faster adopter of externally developed technology, Air Force SBIR partnered with our small team at AFWERX (the Air Force’s innovation and experimentation organization) to develop the SBIR Open-Topic Platform. Companies engage on the platform through a single application process. If companies make it past an initial screening, they are paid up to $50,000 for three months of customer discovery. If, during customer discovery, a solution-problem match is found, the Air Force Open Topic pays for a second contract to develop, adapt, and transition their commercially derived solutions into military operations. Furthermore, if the company finds a military problem owner willing to fund a pilot effort, the Open Topic will match the funding that the problem owner provides. Thus, warfighters who seek solutions on the platform are essentially paid to do so. By making it easier for companies to interact with the Air Force, these incentives drive participation from both companies and warfighters, leading to higher and faster transition rates.

That being said, the SBIR Open Topic Platform is not a magic pill and will certainly have downsides. Monopsony companies who are developing bespoke, military-only technology would not be a good match for this process, and are better suited for the traditional approach. Companies who have not worked with the military before will have to learn the technological requirements to function in a high-reliability setting. Acquirers who have not worked with nontraditional companies before will have to become comfortable with receiving calls from companies pitching solutions. A platform approach must therefore be complementary to the traditional approach and requires a number of additional changes: to acquisition training, management style, and cost and accounting practices, to name a few.

The best thing about platforms, however, is that the more people that are on the platform, the better it is for both companies and warfighters. It is what makes platforms the only economic tool where monopoly benefits both consumers and producers. Depending on a critical variable — the cross-elasticity — the more individuals that are on a platform, the greater the consumer surplus for both. For warfighters, this would mean more companies to work with, and therefore a greater chance of finding innovative commercial solutions. For companies, this would provide more potential buyers, and thus greater access to military markets. It would also provide a relatively quick understanding of whether the company really could, or even wanted to, do business with the Defense Department, thus limiting the agonizing back-and-forth of contractual negotiation that limits the desire of so many companies to sell to the military.

Certainly, the program will work with earlier-stage products and services, but I argue that this makes the approach even more valuable. Those companies at the beginning of product development will gain access to a large set of prospective buyers through this platform, resulting in higher rates of co-creation and leading to quicker product development as compared to the existing, linear process. Platforms should drive both speed and quality, instead of just quality. Just as Amazon uses its platform to generate new consumer products, emergent technology companies can utilize the platform to develop novel defense products.

While the SBIR Open Topic Platform is still new, early signs show strong results. This platform approach has resulted in a radical teaming of a number of highly motivated, yet traditionally “siloed” Air Force organizations to execute the approach at scale. After only one year, the Open Topic initial “Phase I” investment of about $12 million led to over $80 million in non-SBIR (public, in the form of government customers, and private, in the form of equity-based investment) follow-on revenue and funding, with the majority of contracts being awarded to companies who had not previously done business with the government. As a result, the new approach has matched externally derived innovations from nontraditional companies with paying Air Force customers, demonstrating the potential for the rapid transition of innovative, externally developed technologies into the military (as compared to conventional SBIR commercialization metrics).

As a co-creator of the Open Topic, I am admittedly biased. And, certainly, mission-focused, usage-backed metrics are needed. It will take some time to understand the effectiveness of this new approach. However, these early signs indicate how a platform approach to acquisitions can enable solutions to find problems quickly, representing a changing paradigm for defense acquisitions.



Jason Rathje is a captain in the U.S. Air Force and a partner in AF Ventures, a newly established group experimenting with better ways to partner with commercial technology ventures. As a recent alumnus of the Air Force Chief of Staff’s Strategic PhD fellowship, he holds a doctorate in Strategy and Policy from Stanford University. His current research investigates the impact of government funding on corporate growth in both large, incumbent and small, entrant firms. The views expressed here are his alone and do not necessarily reflect those of the U.S. government or any part thereof. 

Image: U.S. Air Force (Master Sgt. Barry Loo)