What is a deep tech company, anyway?
Tech, hard tech, deep tech, frontier tech? And does it even matter?
Now that I’ve got this corner of the internet for writing about deep tech startups, it seems useful to describe just what a deep tech startup IS, and why distinguishing it from other types of ventures is important.
My husband and I debated this topic the other day. He said SpaceX is a deep tech company, while I said it isn’t. He argued that SpaceX’s launch vehicles are such complex systems that the only way to test their “products” to full functionality and performance is to integrate them together and launch them. I countered that while that may be true, there weren’t any scientific discoveries that needed to be performed to build their vehicles. Indeed many engineering and business feats were necessary: modifications to proven rocket designs, surely; new manufacturing processes and suppliers, of course; aggressive project management compared to the status quo, absolutely; a paradigm shift in how the government and really all humans think about space travel, check. But science breakthroughs? Not that I’m aware of.
SpaceX provides a good example of a tough tech or hard tech company: a company with inherent complexity that means it can’t mature the product and/or demonstrate meaningful market traction incrementally. In other words, hard tech startups can’t apply Eric Ries’s Lean Startup Methodology. This complexity that stymies incremental progress can come from the technology itself, from the regulatory environment, or from the way customers will use the product as part of a larger system. There was little scientific uncertainty in the early days of SpaceX, though there was uncertainty in demand that Elon Musk charged through towards his vision of how the launch industry should be, instead of how it had been for 50 years.
Founders, investors, and the like, used to traditional startups, may try to impose incremental milestones on hard tech companies, but often these contrived markers don’t get at the heart of the true risk, or they take the startup off its path to product-market (P-M) fit, in both cases wasting precious resources in the early days.
Most people know what are considered tech companies: a company with perhaps substantial development costs but that can produce and sell its products or services with low variable costs and high margins, like software businesses. Positioning a company in the tech sector did, and to an extent still does, give the perception that it is hot and will usher us into the future; it helps with hiring, fundraising, and PR. Despite having ‘technology’ in the name, these companies are lumped in with what I think of as traditional startups: businesses commercializing software or primarily off-the-shelf (OTS) hardware solutions with tractable development paths. While these companies have little to no technological uncertainty, they tend to have market appeal that is unknowable or uncertain ahead of time. They are a good fit for the Lean Startup Approach, because the tech can be developed incrementally using test environments and existing supporting components, and customers can be probed well enough for real indicators of adoption along the way. Techniques like A/B testing with wireframes and focus groups provide useful information for traditional startups because of the lack of complexity; customers can be transported well enough into the mindset of pain from the way they currently have to do things, to the relief of a world where the startup’s solution exists.
My company, Accion Systems (now called Revolution Space), and most of those I work with today, are deep tech companies. A deep tech startup is one whose first product requires the successful maturation of a novel scientific idea through research and development (R&D). In addition to this first product, a deep tech company will likely engage in future R&D cycles as well, to introduce new features or versions of their product, or to bring new product lines to market as their business evolves. A deep tech company’s uncertainty generally lies in the technological development and end solution to the problem they are tackling. More often than not, the market appeal is obvious, at least for the first few customer segments, but the problem has remained unsolved because the technological solution that will enable the winning product is not yet identified nor proven. Most deep tech companies will find it hard to follow an incremental strategy.
Frontier tech companies can really be any of the above, but their businesses focus on emerging sectors like VR/AR, cryptocurrency, and quantum computing.
Great! So now what? For deep tech companies, making these distinctions is important because applying frameworks and strategies that have worked for traditional startups either doesn’t make sense, or worse, can be harmful to their outcome. This is because there are a handful of implications that arise from the differences.
First, the trial and error nature of R&D means that the phase from idea to market has unknowable development timelines and resource needs. Perhaps a bit uninspiringly, the time to first product in deep tech can be multiple decades. Before you abandon your idea, keep in mind that in addition to including some of the most important problems of our lifetime (which should be very motivating), these ideas have often already been matured to the proof of concept or even prototype stage in an academic or research lab setting, usually with government or industry funding. But the bottom line is the timeline, and therefore the resources required, are unknown, and building a business in this type of uncertainty requires different techniques and tools.
The second implication is that regardless of how much work preceded the deep tech venture, the reality is that the founding team is building a company whose main purpose is to perform the R&D required to develop their first market-scale product, for the foreseeable future. So the team, organization structure, activities, systems, and processes should reflect this reality. Traditional startups spend their first formative years ruthlessly focused on burning down market and business model risk, by contrast.
The last major implication is that deep tech startups often have a harder time engaging with and getting meaningful feedback from the market without a ready product. There exists good advice to do the cheaper, faster validation of your venture and product like traditional startups can, but highly technical products usually have long sales cycles and multiple payers/approvers, are often intricately linked to and operated in conjunction with other parts of a customer’s solution as part of a bigger system, and are sometimes a replacement for an existing solution with complicated switching costs and inconvenience, making it hard to get a true reaction from a CAD drawing or spec sheet of aspirational attributes. (Not to say that deep tech founders should forsake customers in this early stage! More soon.)
These implications propagate throughout the systems, processes, strategies, and cultures of deep tech companies, and where existing resources fall short of accommodating the unique requirements of these businesses, I aim to share tools from my experience through this newsletter.
Stay tuned and carry on solving these messy, important problems!