Carnegie Mellon has one of the most successful technology transfer records in the country and invests heavily in helping its students and faculty bring research to the marketplace. Through vehicles like the Center for Technology Transfer and Enterprise Creation and Project Olympus, an on-campus incubator, CMU is helping students blaze the notoriously treacherous path from thrilling research and bright idea to marketable product.
I recently had the chance to pick the brains of the folks leading these programs: Kit Needham, Entrepreneur-in-Residence and Director of Project Olympus, and Reed McManigle, Senior Manager, Business Development & Licensing and Mentor in Residence, Center for Technology Transfer and Enterprise Creation (CTTEC).
Together, Project Olympus and CTTEC have helped students and faculty launch dozens of businesses, including Safaba, which was acquired by Amazon, Qeexo, whose touchscreen upgrade is widely deployed by Huawei, Oppo, and others, BossaNova Robotics, which is deploying shelf-scanning robots for Walmart and others, and DuoLingo, which is valued above $700M.
Also: How Duolingo uses AI to disrupt the language learning market
Their insights and advice to entrepreneurs are especially valuable if you're sitting on a can't-miss technology you plan to build a business around.
You both hear from lots of CMU students and faculty with big ideas. What are the biggest mistakes or misapprehensions you see from technologists who want to transform research into the foundations of a business?
Reed McManigle: At CTTEC we deal primarily with inventor teams of faculty and PhD students, and inventions that have resulted from (perhaps years) of research funding. With this starting point, it's easy for the inventors to be convinced that the science/technology is the key driver of success, instead of the actual drivers of finding a real, sizable problem to address; building and managing a team; developing and executing a go-to-market strategy, etc.
Is that more or less true for every technology, or are there some categories that tend to fall prey to those problems more than others? I guess another way to ask the question: Is there a category of tech that you just know will present a very tough path for an entrepreneur?
Kit Needham: The hardest tech in my experience is apps. In the ten years that I've been here, only one app has been successful. They are relatively easy to build but getting market awareness and adoption is very, very difficult. Further, they aren't really solving big problems and if they are, they aren't sufficiently distinct or better from the other solutions that customers are using.
Reed, I know you've got some different thoughts on that question. Can you elaborate?
Reed McManigle: Almost anything that is non-software based, such as medical devices/biotech, energy, chemistry, materials science, is harder to commercialize. The challenges are that these technologies are generally far from being market-ready, or even market-testable, when they are no longer eligible for the basic research-oriented federal funding on which they were based. While these types of technologies have the potential for huge impacts and upsides, they require significant amounts of time and money to reduce technical and market risk.
Prior to having some meaningful feedback from potential customers (or in the case of medical technologies -- some efficacy data), it's next to impossible to recruit outside entrepreneurs or to solicit investment. SBIR/STTR [Small Business Innovation Research and Small Business Technology Transfer programs, also known as America's Seed Fund] funding can be a bridge over this gap, but it's a slow process and requires someone (often a PhD student or post-doc) to commit to being in a startup that has scant funding, instead of pursuing a more stable job in a larger company or an academic career. SBIR/STTR funding only pays for research, so it may not be possible to involve an experienced business partner/entrepreneur with the scientific founders in the early stages of the company, which can stunt their business development efforts.
Given all that, how can researchers and inventors build a case that there's a customer for their products given their limited resources and backgrounds primarily in tech, not business?
Kit Needham: We start almost every one of our startups doing customer discovery interviews (the right way; they are not allowed to talk about their idea). Our recommended minimum is 100 stakeholders. If done correctly, pivots take place before the product/service is even developed, there is more confidence of market-market fit, knowledge of how well current solutions are working, priority of what is most important to the customers, etc. In short, they are solving the right problem, and then they go back and ensure they are solving it the right way.
Reed McManigle: We push our inventors to do 'customer discovery' to develop deep understanding of the market needs/pains, competitive approaches, evaluation/purchasing processes and metrics for new solutions, etc. and with that information in hand, develop a minimal viable product and line up pilot testing with potential customers.
Why the historical indifference to entrepreneurship at research institutions? Given that many of the best ideas and research come out of universities it seems like that bridge would be more logical. Yet many universities seem to be falling down on the job.
Reed McManigle: Historically tech transfer and entrepreneurship was seen as a distraction for a successful academic career. Strong publications and a history of successfully pursuing grant funding for research are the key metrics for advancement in an academic career. CMU has had a much more entrepreneurial history and culture than many other universities, but even here it has taken decades, multiple success stories, and long-term investment in a tech transfer and entrepreneurial support ecosystem to get to the point we are today (supported in large part by our outstanding foundation community).
From an institutional perspective, it's worth noting that most universities don't break even on their tech transfer operations. Patent prosecution and skilled staff are expensive and difficult to justify solely on the basis of economic return. It can be very difficult to determine which technologies are worth pursuing at the early stage at which they are disclosed in a university environment (WAY before VCs make their decisions, with typically a 1 in 10 hit rate for big returns), but even so, many investments need to be made with low hit rates for a financially successful outcomes. Less than 1 percent of university inventions generate more than $100k in licensing revenues (per analysis done at Stanford).
Also: Launching and building a startup: A founder's guide (free PDF) TechRepublic
One of the things that sets CMU apart from many other universities is its 'standardized' approach to licensing to startups. This reduces the time, complexity and hostility of negotiations. Importantly, CMU's approach recognizes that many startups are bootstrapping, and don't have the ability to pay substantial sums for an IP license in their early days. Therefore our startup licensing approach, which combines equity and royalty compensation to the university, requires no upfront licensing fee, no annual minimum royalties, and no royalties on revenues for three years. We also give the startup an option to defer reimbursement of past patent expenses for three years (while the university continues to cover new expenses incurred) in return for additional equity, and the option to 'incubate' the company in their university lab, in return for additional equity.
A lot of technologies have gone from "out there" to mature in a short amount of time. Drones, AR/VR, and robotics, for example, are more and more becoming the domain of huge companies. In which technology categories do you see the most potential for researchers-turned-scrappy entrepreneurs?
Kit Needham: One of CMU's blessings is our research is leading edge. However, market readiness takes longer. For example, all of the above, while growing, are still at the very, very early stage of the growth curve.
In a practical sense, the biggest current opportunity is around machine learning and artificial intelligence since that can be applied to improving existing, already adopted products and services.
Reed McManigle: Each of the examples you mention have been under development in universities for decades before becoming 'overnight successes'. The pace and breadth at which they are now being deployed masks the long history that led to this. I would put AI/ML and robotics at the top of the list of technologies 'whose time has come' in terms of corporate and investor interest.
CMU has great and well recognized strengths in these areas, which sometimes overshadows the great but less recognized strengths we have in energy, materials, chemistry, medical devices, healthcare IT, etc. As mentioned above, many of these sectors need more time and money to get to a first customer, which makes it more difficult for a researcher to be a 'scrappy entrepreneur' on his/her own. We try to help them find outside, experienced entrepreneurs to partner with to form a scrappy entrepreneurial *team*.