What is it about the “valley of death” that both government and commercial software executives strive to cross?

The term means one thing to government and something different to industry because they are each seeking different goals. In the government, it refers to transitioning from a research and development project to a program of record. Among venture-backed tech companies, it refers to transitioning from pilot programs to larger customer engagements and the valuation-enhancing financing rounds that ensue. Yet this is extremely difficult to do.

When thinking about how to cross the valley of death, instead of a bridge, a better metaphor may be a stepladder; taking measured steps that include sizing the market, honing the product and ensuring that existing revenue is going toward funding product development that enables solutions to chart a path out of the “valley.”

The reason transition is so difficult is that the expectations that accompany new concepts and technologies undermine the commitment to long-term efforts to adopt the capability. Consider the Pentagon’s approach to artificial intelligence, which for decades has been stuck in a cycle of investment, disappointment, forgetting and investment again.

The “hype cycle” starts with initial R&D and early adopters driving the technology into the marketplace. This is typically accompanied by initial venture or government funding, higher unit prices, and a lot of product customization to meet customer requirements. Broader market acceptance generates hype around the promise of the technology’s ability to address legacy problems.

At this point, this “hype” creates an impossible standard that can lead to a loss of support and officials looking for the next technology that will surely solve all their needs. Exacerbating the issue, the companies driving the development and adoption of these maturing technologies also face heightened competition for follow-on rounds of venture capital funding.

These are the headwinds that trap and kill technologies in the valley of death.

Despite apparent customer success, many of these companies end up going out of business. The vast majority of companies do not survive the valley of death and do not make it to the promised land of a scalable and profitable future.

This excess supplier problem, faced by both venture-backed technology companies and government R&D projects, was documented by Carlota Perez, who saw this boom-and-bust cycle of new technology as laying the foundation of future technology adoption.

So what strategy should a true commercial company that wants to support the U.S. government adopt to cross the valley of death?

The answer is to pursue a stepladder approach of sizing the market, honing the product and ensuring that existing revenue is going toward funding solution development and not recreating a purebred defense contractor.

What is essential to determining product-market fit is getting to know the customer base you’d like to sell into first. The government doesn’t buy technology and doesn’t buy capabilities; it buys solutions. A commercial startup should devote initial resources getting to know prospective customers, what their trust pain points are and how your solution is better than the competition.

Many commercial technology companies fail to consider how their business model aligns with government contracting and the bureaucratic stipulations that encumber supposedly lucrative contracts. Working with government requires understanding the ins and outs of the literally thousands of pages of regulations: intellectual property rules that are unique to government, the Truth in Negotiations Act, commercial items, colors of money, Section 889, Buy America rules and streamlined acquisitions. These regulations include price controls, contracting audit requirements and artificial timelines that await new entrants.

Then the real headaches begins with obtaining “authority to operate” software solutions on a government site, complying with the new Cybersecurity Maturity Model Certification standards, and ensuring your software-as-a-service solution does not become a service-as-a-software business model that capsizes a lofty venture-backed valuation.

In other words, it is suspiciously easy to slip from being a product company to a services company chasing revenue in the government sector. To prevent this, companies must assess the specific contract encumbrances that come with the federal acquisition regulations, many of which government officials may not even be aware of themselves.

While you explore product market fit, look for defense contractors that can be engaged in co-opetition. Ideally, your commitment to your product allows them to favorably integrate your capability into their solutions being delivered to government customers — a recipe for a long-term business relationship. Conversely, you can take advantage of their infrastructure and contract vehicle options, and perhaps even squat at their office when visiting your growing customer base — COVID-19 be damned!

As a startup builds its business, it must also develop a strategy for programs of record. As noted defense analyst Eric Lofgren has shared, this does not necessarily involve “winning” that big contract but rather developing solutions that can then be sold through Part 8 of the federal acquisition regulations using operations and maintenance funds. Many startup offerings may fall into this category, supporting larger companies that raised the venture funding necessary to compete for the program of record outright. Other options include subcontracting for the prime.

The valley of death is real, but it persists because startups fail to address the government market in a methodical way. Rather, they look for limited traction such as an R&D grant to validate their interest in federal contracting and then commit too early, in a haphazard manner. As a founder myself, I advise others: Never assume your future success will rest on government investment. Always remember that your technology is dual use, starting with the commercial sector that it originated in. Most of all, don’t worry about building a bridge across the valley when a longer ladder will do nicely.

Stephen Rodriguez is a managing partner at One Defense, a senior adviser at the Atlantic Council and a senior innovation adviser at the Naval Postgraduate School.

More In Commentary