How Technology Research Compares to Doing a Tech Startup

iStock_000007647676LargeIt isn’t too often that people mention technology research and tech startups in the same breath. After all, technology research happens out of view in university labs or corporate research centers and is published to a select few Ph.D. types. Tech startups, on the other hand, usually get the more splashy press with their billion dollar IPOs and their disruptive ambitions. How could any two things be so different?

But it turns out they aren’t so different. As both a technology professor and a tech entrepreneur, I’ve had the opportunity to reflect on how both endeavors compare. There are many similarities, but also many differences, and often the same talented students who consider joining a tech startup also consider graduate school. I hope this blog post may help them make their decision about which path to take. And it’s certainly possible to do both, as professors like my friend Shwetak Patel so successfully demonstrate.

More Similarities Than People Realize

Let’s start with the similarities between technology research and tech startups. It should be no surprise that while many startups begin in the proverbial garages of Silicon Valley, many also spin out of university research labs. In fact, government-sponsored university research is essential to many of the discoveries behind companies like Google, Intel, Qualcomm, Apple, Microsoft, and others, accounting for over $500B/yr. in tech-giant revenues according to the National Research Council. Google, after all, was an outcome of the graduate work of Larry Page and Sergei Brin funded directly by the National Science Foundation. Many of my current UW colleagues have tech startups of their own: GraphLab from Carlos Guestrin, SNUPI from Shwetak Patel and Matt Reynolds, Trifacta from Jeffrey Heer, and AnswerDash from me, Andy Ko, and Parmit Chilana. From the start, many startups have their roots in the research activities of university professors and their students.

The single biggest similarity between technology research and tech startups is that both require the ability to work productively under conditions of extreme uncertainty. In research, uncertainty comes from solving open problems whose solutions aren’t known. (If they were, it wouldn’t be research!) In startups, uncertainty comes from not knowing whether the product or service you are creating is going to solve a genuine pain, get noticed, and gain traction fast enough to avoid running out of money.

Most people seem to grow pretty uncomfortable when having to make progress under uncertainty. They are more comfortable with concrete plans than uncertain possibilities. Big companies have plans. Researchers and startups explore possibilities, and need people who can work through them with imperfect information. Possibilities excite some but terrify others because it is always unclear which possibility should become the next priority, and choosing one possibility means excluding another.

Another similarity between technology research and tech startups is that, to be successful, you have to work a ton. There’s just no substitute for 12-hour days. Why is so much time and effort needed? Again, it comes from the extreme uncertainty. Since you don’t know which of your efforts are going to pay off, you have to do as much as you can, hoping that one of the arrows you shoot will hit the bullseye. The more arrows you can shoot, the better your chances. And while many arrows don’t guarantee success, too few arrows do guarantee failure when the target is uncertain.

Different Purposes Mean Different Work


Differences also run deep. Most of these differences stem from two fundamentally different purposes. Other differences arise in terms of focus, what constitutes impact, and how teams function.

Purpose. Technology research is fundamentally a scientific endeavor. Like all scientific endeavors, its purpose is to answer open questions at the forefront of a field of knowledge. Its goal is therefore knowledge discovery. In technology research, that usually means knowledge of how to do something not known to be previously possible. It may also be the discovery of a property of technology, computation, or information that was not previously understood.

Startups are fundamentally focused on commercializing something new. The goal is to create value and ultimately make money. The goal is not answering open questions. In tech startups, it doesn’t matter so much if someone has “done it before,” if you can do it better, faster, more reliably, or for a lower price. If there’s a buyer willing to pay more for your innovation than it costs you to make and sell it, you’re in business. In tech startups, it’s fundamentally about winning, not just discovering.

Now, discovering can certainly lead to winning, because new discoveries can lead to monetizable advantages over existing solutions. But this is not a requirement of technology research, and to place such a requirement on research is to corrupt its purpose and drastically cripple its scope. This is a point too often missed by those calling for universities to be run more like businesses. With different purposes must come different ways of judging success.

Focus. Because the purposes of technology research and tech startups are different, their foci are different. Research focuses on demonstrating possibilities, while tech startups focus on realizing possibilities for a market. In research, it only takes one successful instance to show that something’s possible. But tech startups must create more than one successful instance. They must implement their vision across many instances to scale. This means supporting multiple versions, platforms, devices, etc., and worrying extensively about edge cases. Technology research largely gets to avoid such worries because they aren’t central to demonstrating new possibilities.

Impact. From different purposes come different measures of impact. Technology research aims to have impact through significant fundamental breakthroughs. Such breakthroughs can be thought of as step functions propelling whole fields forward. This kind of impact can take time to emerge and isn’t accurately measured by market response. On the other hand, tech startups aim to affect as many people as possible and generate commercial value, so market responses—number of customers, rate of adoption, amount of revenues—are appropriate measures. In a sense, research impact can be thought of as depth—deep discoveries that penetrate the heart of a field and reside there for a time. By contrast, startup impact can be thought of as breadth—widely used products that gain a devoted user base with enough buyers willing to pay.

Teaming. Technology research usually takes place in small teams of 2-5 people. Success usually depends on the efforts of a few. This arrangement is similar to startup founding teams, but to succeed, startups need to grow beyond just 2-5 people. The reasons for different team sizes should be apparent from what I’ve said thus far: achieving depth through singular breakthroughs that demonstrate possibilities is very different in scope than achieving breadth through widespread adoption where all edge cases are covered. That takes a small army.

A Metaphor to Capture the Differences

I like metaphors as models for thought. My metaphor here relates technology research and tech startups and even large tech companies as being part of one overall “tech ecology.”

The tech ecology is a stream that begins in the mountains, becomes a river further down, and eventually finds its way to the vast ocean. Technology research is at the headwaters; a mountain stream. It’s narrow, precise, not easily observed, and far away in time and space from the ocean. Tech startups are further downstream, and as they grow, they swell into rivers looking for the ocean. The ocean is where the waters disperse, become widespread, big, broad, powerful, and easily accessed by anyone.

Mountain headwaters can certainly influence which rivers grow large. But by the time they do, they have undergone significant transformations in their breadth, audience, power, and capacity for impact, the very things companies strive for.

So Which to Pursue?

If you are a student in a technology field and wondering whether to pursue technology research or join a tech startup, the following questions can help you decide:

  • If you get excited about making new discoveries, go into research. If you get excited about making products that people use, join a company.

  • If you like being the first to demonstrate a new possibility, go into research. If you like refining and perfecting things, join a company.

  • If you get excited about your own individual impact, go into research. If you get excited about the impact of an organization you’re a part of, join a company.

  • If you like working alone or in small teams that will never grow, go into research. If you like the possibility of growing an organization, join a company.

  • If you like asking fundamental questions, go into research. If you like removing barriers to the adoption of a specific product, join a company.

Whatever your path, working in technology today is as exciting as ever. It is a privilege to shape the technology landscape, whether as a mountain stream or a surging river. We should all count ourselves lucky to find our place in the tech ecology.

Topics: University Startups