This is part 2 of 3 on software for preclinical R&D. Read part 1 here: “Software Isn't Eating Preclinical R&D, and That's a Problem.”

As recently as 2014, a market that’s now massive received little attention from investors. That market was for developer tools, which help software engineers be more productive. At the time, TechCrunch noted persistent reasons venture capitalists avoided investing in them: “The market isn’t very large, preventing huge scale; there are only a handful of potential buyers, leading to small exits; and most venture capitalists don’t understand the needs of developers in the first place to even make an informed investment.”

Fast forward six years and all this has changed. Investment increased about 10% per year in that timeframe. Last year alone, investors put $2 billion into open-source and developer tool startups. And there have been big exits. In 2018, for example, Microsoft acquired GitHub, a tool for managing source code, for $7.5 billion in stock. Last year, cloud application monitoring service Datadog went public and quickly achieved a $10 billion market cap.

Swap “developers” for “bench scientists” and you’ll see parallels between the developer tools market and the preclinical R&D software market. In the first article in this series, I described how immature the latter market is. This results in poor software support for bench scientists developing lifesaving drugs. In this article, I’ll describe why investors have mistakenly avoided the market. This is the key reason it continues to languish. And it's also why the opportunity remains so untapped.

Six reasons venture capitalists underinvest in preclinical R&D software, and why they’re wrong

In the past four years, I have met over 300 investors while raising money for BenchSci, an AI-powered platform for designing preclinical experiments. We were fortunate to be successful in our fundraising (over $44 million raised to date). But it wasn’t easy. And other than Benchling, there are few similar success stories I’m aware of.

Why? Because investors don’t believe you can build a company in this space that provides a large enough financial return. This belief is based on six understandable yet incorrect assumptions about a market that, like the developer tools market, is ripe for disruption:

  1. There aren’t that many scientists, so the user base isn’t big enough. Compared to other software markets, even vertical ones, the number of users in preclinical R&D is admittedly small. While it’s hard to get exact numbers, most people in the industry believe there are about 500,000 research scientists worldwide. But the number of potential users becomes even smaller when you realize that 70% of these scientists are in academia, and investors hate academic customers as they assume they won’t pay much for software. So that leaves about 150,000 potential users. Since the most common pricing model is seat-based, multiplying 150,000 by any reasonable value doesn’t yield a large enough addressable market. This is all correct. But the incorrect assumption here is that seat-based pricing is the model. Since enterprises gain the most from more productive scientists, they will pay for value rather than seats. So the assumption is false.
  2. There aren’t a lot of logos. If we move away from a seat-based pricing model to an enterprise value-based pricing model, the market size increases. But investors will argue that there aren’t enough enterprises in life science to achieve sufficient scale. After all, there are only about 25 large pharma companies, and the industry continues to consolidate. This is true, but it misses some key facts. First, these enterprises have significant R&D budgets, hence will spend $5-10 million per year to achieve significant R&D savings. Second, the total market includes large pharma, mid-size pharma, and thousands of biotechs under similar pressure to achieve faster, more cost-effective R&D research.
  3. There aren’t any success stories (exits). There is only one Silicon Valley success story within the life science industry: Veeva Systems. In fact, Veeva is the largest vertical SaaS company in the world. It currently develops software for clinical research and commercial activity, such as sales. It doesn’t develop software for preclinical R&D. But Veeva’s story actually reinforces how much investors underestimate life science software’s market potential. When Veeva raised its first and only round of VC funding, they estimated the total addressable market at around $500 million. Today, the company’s annual revenue is over $1 billion. A similar opportunity exists in preclinical R&D software.
  4. They don’t understand the user base. To truly understand the complexity of preclinical work, you have to experience it. Since the majority of investors don’t have a PhD in biology or a related field, they don’t really understand the space. I have learned it is hard for investors to invest in a company that solves a problem they haven’t experienced and don’t appreciate. But it’s not impossible, as the developer tools market shows. Remember that six years ago, TechCrunch explained lack of investment in developer tools as in part due to investors not understanding the needs of developers. But that didn't stop the market from growing, or VCs from eventually investing.
  5. There is limited IT spend today. Pharma has limited IT budgets for preclinical R&D, as I have already described. Investors know this, and it’s one reason they avoid the market. But the thing is, it’s a catch-22. Investors see limited budgets and therefore don’t fund new solutions that companies will pay for. But then the budgets never increase, because pharma IT departments have no new preclinical R&D software to buy.
  6. They believe little money goes into preclinical research. For some reason, there’s a common belief that “the real money is in clinical trials.” I have heard this so many times. But it is false. About 30-40% of the drug discovery cost is in preclinical R&D.

You can’t solve meaningful problems in preclinical R&D software without venture capital

Of course, VCs might simply state that the life science industry doesn’t need them. After all, there are many successful bootstrapped software companies. Even GitHub was able to build a user base of 100,000 engineers and a successful company before raising their first venture capital investment. Why can’t scientists do the same thing?

There are two reasons for this: the complexity of the software, and the time to revenue. Engineers can build software for other engineers. Scientists can’t build software for other scientists. To truly build game-changing software tools for preclinical scientists you need a diverse team of PhD scientists, software engineers, and machine learning specialists. That costs a lot of money. And that money takes a long time to earn back. It can take 18-24 months to sell your software to a pharmaceutical company. Add 90-day payment terms and you’re talking about over two years before you start generating cash flow. This combo makes it almost impossible to build a software company in the space without VC funding.

In short: the market is ripe for VC funding, providing both outsized returns, and outsized benefit to the world. If this funding doesn’t materialize, it will continue to slow the pace and quality of drug discovery. In the third and final part of this series, I’ll explore the potential impact.