University Startups and Spin-Offs: Guide for Entrepreneurs in Academia (2015)
Part I. Strategies for University Startup Entrepreneurs
Chapter 2. The Lean Startup Changed Everything
When Eric Ries published his book The Lean Startup in 2011, he brought about a paradigm shift in business modeling.1 Only a few of the ideas in the book were groundbreaking or new, but many people were seeing them outlined neatly and logically for the first time. According to the Lean Startup method, “build, measure, learn” is at the heart of entrepreneurship. This means testing, measuring, changing, and retesting several hypotheses before a product launch. Real potential clients give feedback on minimum viable products (MVPs), which helps tailor the product or service to the market in small steps. Learning and experimenting therefore trump planning and politics. Businesses focus on product development close to their customers to improve their impact from day one. Hallelujah!
When the book came out, I was recovering from a venture that had started out promisingly. Flush with seed investment in the mid-six figures, the company I cofounded in San Francisco was about to revolutionize the trading of copyrights, trademarks, and patents on a digital platform. Offices in the Mission District were rented. A board 12 members strong, including executives from eBay, was summoned. A fancy code shop full of developers was hired to program the platform. Hours were spent in conference calls, with phone-ins from different continents. Things were looking up, until the financial crisis in 2007/8 came around. The second round of financing fell through, and the startup collapsed under its heavy overhead without having launched a single line of code. This was the opposite of a lean startup, and it is painful to remember all the silliness. We wasted much time, not to mention the investment of the venture capital firm. Lucky for us, they understood that bad things happen to good people, which softened unwinding the mess a great deal. Having such lenient investors was by no means a foregone conclusion.
According to the Lean Startup method, we should have approached the business differently. Instead of trying to build a new universe to revolutionize the licensing market with a big bang, we should have thought about our MVP first. If we had coded it ourselves, we could have tested it daily with actual potential users and clients. We probably would have learned that the original idea was much too large and doomed to fail and that we should focus on a smaller niche from the beginning. We could have found the most profitable niche through further testing. With clear signals for or against the need for such a platform and a much stronger idea of what the code should look like we could have engaged professional developers to finish the beta for launch. Going about the project this way would have required only 10% of the investment, with much more learning and insight along the way. Instead we were flying blind based solely on our own rigid assumptions, which we even vehemently defended to investors and advisors.
Before the Lean Startup method became widely adopted, most business founders focused primarily on launching the perfect product and much less on feedback from the market. Long brainstorming sessions about the perfect strategy replaced simply trying something and seeing what happened. Entrepreneurs founding a company today should figure out how to run it lean to make the best use of their resources.
The Lean Startup method is not only for tech startups. Multinationals like General Motors use it for their in-house R&D as well. Steve Blank goes as far as calling it a new strategy for the 21st century that goes hand in hand with other changes like democratization of knowledge and capital, low startup costs, mobility, globalization, and convergence.2
There are various schools of thought about launching a company. For university startups, I find the Lean Startup method ideal. This chapter explains how to apply it with two examples. But let’s first understand the main underlying principles.
Incremental Product Development
Under the old paradigm, most university startups think they have to brainstorm extensively to arrive at the perfect product. After meticulous planning, they build it behind locked doors and finally launch it in the market with a big splash. This is expensive and risky and also impractical. There is a more intuitive approach to develop products, which the Lean Startup method refers to as incremental product development. This is how it works: you first create an MVP based on your assumptions about what the market could be willing to accept. This is obviously not the full-featured version of your ideal product, but the opposite. The simplest, least-expensive product you can produce quickly with the sole purpose of getting feedback from real potential clients is your MVP 1. Now you continually improve your product, called pivoting in Lean Startup jargon. You test the next assumption with MVP 2—for example, when you have fixed the flaws in MVP 1 and made it more complex with added features. Along the way, the product becomes better and better, and it is always approved by the market. At some point, you feel it has developed into a robust commercial version that enough clients are willing to pay money for. Only then do you launch your first actual product (the final MVP) in the market. Even after launch, product development still continues. You keep collecting information about how you could improve the product in an upgrade or the next version. Figure 2-1 illustrates this process.
Figure 2-1. Minimum viable product (MVP) testing with “Build, measure, learn”
For each MVP, you need to test two hypotheses: the value hypothesis (whether anybody in the market wants it) and the growth hypothesis (whether there is sufficient demand in the market to have a business). This boils down to, “Do people want our product?” and “Will enough of them pay for it?” Often, both hypothesis tests can occur at the same time. MVP testing is about experimenting and recording your findings—not in the lab, but in the real market with real potential clients. By leaving the building, you shift the question from “Can we build it?” to “Should we build it?” If your market gives you the thumbs up, then you have a business, and the startup is ready for takeoff. Otherwise, keep experimenting with other MVPs.
Most university startups, even those with repeated grant funding, have never conducted such tests and are not going to do so anytime soon. This is unfortunate, because in a global marketplace you cannot afford to fly blind. You are up against millions of highly motivated incremental product testers who are validating hypotheses on a daily basis to get a head start. Spending time writing grant proposals will put you at a disadvantage.
Simplify, Then Simplify Again
The idea of testing MVPs sounds intuitive—that is, until you try to apply it in your own startup. It can be surprisingly hard to come up with a quick product that can be tested easily. Not only that, but you need to be able to implement that feedback rapidly to improve your product. If your MVP is too complex or expensive to build, you may collect some feedback, but incremental development will be tedious. Finding the right MVP is therefore the first order of business for a startup.
As a software company, especially when your product involves a mobile app, you can quickly mock up a user experience with design software or even on paper. It is less easy when your startup involves hardware. The MVP is never what you would like to make, but what you can make today with little funding and no additional materials. Toward that end, Table 2-1 outlines a simple process to find possible MVPs.
Table 2-1. Table to Arrive at Minimum Viable Products (MVPs)
The goal of this exercise is to find an MVP that you can create quickly with the materials you already have. Go down the list until you cannot simplify any more. As a rule, the most-reduced product is always just a picture of what you want to make. This may sound trite, but many startups think their technology is too complex to create an MVP. You may have an animation, a screenshot, or any other visual representation of your product. Get feedback on this first, using a checklist like that in Figure 2-2. Then you can step up and test a physical MVP. Ideally it should cost you nothing, because you can easily put it together yourself. You should be able to produce this MVP in a few days.
Figure 2-2. Checklist for MVP testing
Entrepreneurs generally like to start with the maximum viable product in mind. To scale that down to the minimum is easier said than done. Researchers in particular feel like frauds unless their product is flawless. In startup entrepreneurship, this is impractical. Pivoting a buggy product wins the day. Have you thought hard, but you cannot come up with any ideas to simplify the MVP? Engage your team in a brainstorming session. Perhaps also invite the accountant and the guy who does IT support at your university. Those without domain knowledge sometimes have unexpected suggestions that help you strip down your MVP. You will have difficulty to gather timely information in the market if your product is too complex. The worst outcome is to perfectly carry out a plan that results in a product nobody wants. Better to make the simplest prototype you can think of and show it to potential consumers today.
The Lean Startup in Action, 1: Battery Startup
Imagine the following fictitious example. Three PhDs named Adam, Ben, and Christine are working on their postdoc, researching a new smartphone battery technology. Not only is their battery biodegradable, it also fully recharges in ten minutes. Two years of diligent research have yielded promising results, and the postdoc is nearing its end. One evening, over beers, the team talks about their future, and someone makes a joke about starting their own battery firm. Hours later the trio is still brainstorming about how to do that, and it becomes obvious the entrepreneurial bug has bitten them. The idea sounds even better the next morning. But where to begin? They have published several papers about the technology and given presentations at congresses. Industry leaders have expressed vague interest in “doing something together,” but none of the team knows much about business or raising funds. Their scientific advisor supports the startup idea. She proposes that they write a business plan, make a budget, and apply for grant XYZ. The grant-award date is 12 months away.
Instead of getting sidetracked in writing a lengthy business plan and a grant proposal, the team should first gain clarity about their motivation and product. They should start confirming their idea while they still have the infrastructure of the university available. Is the battery (or a variation of it) something the market wants? If so, what price will buyers be prepared to pay for it? Does the team really want to run a company? They should start thinking about their MVP. Instead of the finished battery for the smartphone, this will be perhaps a larger rechargeable battery that the team could create more quickly. In the best case, the MVP is a good-enough product they can produce in a small series for market testing without much effort. Before they produce the MVP, the team should leave the lab and speak with manufacturers and wholesale and retail customers about the product. Would they carry it in their stores or make their distribution channels available? Will they pay price X? What about price Y? The first MVP could be an advertisement in which the battery is available only on paper for preorder. In love and war, all is fair. Another simple MVP could be a campaign on a crowdfunding platform such as Kickstarter. The goal is to get real-life data from potential clients fast, without spending much time and money.
Once you start thinking about possible test cases, you will come up with many. The key is to engage your thinking along this path and get busy—not in the bubble of the lab, but in the real market.
Back to the battery startup. Within a month of the first hypothesis test, one of the team, Adam, has thrown the towel and left for his native land where a position on a tenure track beckons. Good for him. Now it is up to Ben and Christine. They have found that the biodegradable, rechargeable battery they imagined is too expensive to make. However, if they forget about making it rechargeable, they can offer it competitively. But then it will no longer work in a smartphone. Interviews and responses to the mock advertisement have shown that a deep market exists for biodegradable batteries. The simplest battery they can produce is a car battery. Tests with potential buyers have validated both the value hypothesis and the growth hypothesis: people want a biodegradable car battery and are willing to pay for it.
Let’s recap. The team wanted to produce a smartphone battery, but it became obvious that they needed to simplify their MVP. The original idea involved complex manufacturing, device dependence, and other roadblocks that can only be overcome with scale. As a result, they had to think about other applications of batteries, such as AAA, 4.5 volt, CR, and car batteries. Whichever of these was the simplest for them to make, they chose as their MVP. Tests of the car battery in the market showed that there were willing buyers for such a product. Thus, with a car battery, they have a product that satisfied the value hypothesis and the growth hypothesis. With this, they have a valid business idea and can move forward.
The startup duo can now decide whether they still want to embark on the venture. If they do, they may approach a leading battery manufacturer and offer to enter a joint venture to further improve, test, and finally mass-produce the battery. If this happens, they will not need to raise funding because their joint venture will be a subsidiary of the battery manufacturer, in which both Ben and Christine will own equity. The university will often also own a small stake in spin-off companies, which is only fair.
To engage the battery manufacturer, Ben and Christine had to first confirm their idea. They had to find out the original product they planned was impractical and too expensive to produce. Only through testing in the real market could they match their technology with existing demand. Because they were running lean, they developed a useful product inexpensively. Had they not done that, someone else would have had to figure out the right product based on their technology. This is often unattractive for joint venture partners, who have enough on their minds with their own problems. Yet if you can prove to them that a profitable market for your validated product exists, their risk is much lower. They only need to finalize the product and can then integrate it in their sales and distribution channels.
This is an idealized story, of course. But note that no business plan was at the heart of this team’s success. Many assumptions about the project changed, which would have resulted in pointless updates to a business plan that nobody will ever read. And no grant proposal was submitted. Had the team done that, they would have robbed themselves of momentum, valuable data, and deeper insights about the nature of the market. This eventually laid the foundation for the future of the business. As a bonus, they learned that one of the cofounders was uncomfortable with startup life and preferred a 9-to-5 job. Sorting this out later down the road would have been a messy affair.
Because lean product development is inexpensive, startups have strategies at their disposal to test ideas and assumptions right away. From the beginning, they have one foot firmly planted in the market, collecting feedback and incrementally developing products. The founders immediately see what startup entrepreneurship entails and can decide if they are cut out for it or not. Gaining momentum as early as possible is extremely valuable. The lean startup will always have a leg up against those who sit at their desks, formulating perfect business plans even before they test their product ideas in the market.
The Lean Startup in Action, 2: Electric Scooter Startup
Let’s look at another fictitious example of MVP testing. In this case, a prototype exists that incorporates several individual technologies. This is frequent with university startups.
Assume that in his PhD, Simon has built an electric scooter that he wants to bring to the market with a startup. The scooter has a slick design with several innovations in the technology of the wheels and the closed body. Other features include a newly developed accelerator, regenerative braking, and a quick-charging battery. Simon believes the retail price of the scooter should be around $2,000. A prototype of the scooter exists in the lab, but he is unsure what to do as the first step to commercialize the idea. How should he go about all this with the Lean Startup method?
First, Simon should test the value hypothesis with potential buyers of the scooter. Do they want what he built in the lab? If the scooter drives safely, he could test the prototype right out of the gate. However, this is not his MVP. What will Simon do if he finds that test drivers do not like the prototype? Is there something wrong with the wheels, the body, the accelerator, or the brakes? There are too many moving parts, which will prevent him from gathering meaningful feedback, and he will be no wiser than when he started. Instead he should simplify the MVP and break down the tests into several steps.
Simon can start by testing any of the scooter’s components, but he should focus on the one that is easiest to test. Is it the battery, the wheel design, the electronics, certain features of the build, or something else? He should find the MVP, with emphasis on minimum. If he is testing the wheel design, then he can take an existing scooter and just add his wheels. He has already tested them in the lab to see how much traction they have and how many hours they last with normal use. Now he has to leave the building and invite some potential customers for a ride to see what happens. Do they prefer his wheels, or are they indifferent? Simon should measure how his MVP scooter fares compared to a normal model. He may learn that customers love the wheel design, because it provides a new kind of driving experience. If this is the case, he could commercialize the wheels alone and approach a manufacturer. Right from the beginning, Simon’s startup has a product with the potential to generate revenue. If he was still brainstorming about how to build the perfect untested prototype, he might never have known that a marketable product was lying around in the lab, unused.
Simon has tested the wheels. Test riders felt that they are “more fun” than traditional wheels, because they are easier to navigate with. He records this and moves on to the next MVP. To find out how close to the market his original prototype is, he adds parts of his scooter to the MVP one by one and tests how long the market still accepts it as a product. The next component of the scooter is the accelerator. Simon adds it to MVP 1 and arrives at MVP 2 (wheels and accelerator). He tests it in the center of town, where most potential buyers of the scooter are. One day, it begins to rain unexpectedly, and the accelerator stops working. He never knew of this flaw, because up to this point, he only tried it in the university’s air-conditioned hangar. He can now find out what the problem is, improve the accelerator, and test it again.
Then Simon moves on to the brakes. Afterward he tests the closed body. The original prototype of the body turns out to be impractical for most potential users. So Simon mocks up changes with CAD software and shows the renderings to potential customers, just to see what they say. He adapts and improves the design until the market gives him the thumbs up.
He moves along like this until he has tested all the different features of his prototype. The wheels and the accelerator deliver a new driving experience, but the fast-charging battery adds too much weight, which slows down the scooter. He learns which individual parts—the wheels and the accelerator—satisfy the value hypothesis of the market. So, Simon could start with them alone as his first products. Had he tested his prototype in one go, he would have gotten either a positive or a negative response and would not have known where to start with improvements. Based on his findings in MVP testing (illustrated in Figure 2-3), he can now commercialize the individual components and build a scooter the market wants.
Figure 2-3. MVP testing of the electric scooter
Simon confirmed several value hypotheses. How about the growth hypothesis? Will people pay $2,000 for the scooter? If some testers already want to buy the scooter, then yes. Otherwise, Simon can speak with potential customers about pricing and find out this way. He may learn that adding the closed body leads to most of the cost, so he has a bigger market if he produces a conventional, open scooter. In either case, Simon has first-hand intelligence about his market and his customers’ preferences.
To have a viable startup, both the value hypothesis and the growth hypothesis need to check out. When they do, a startup is infinitely more advanced than one that is still brainstorming. If Simon is not exhausted yet and wants to pursue entrepreneurship with his startup, then he is in the game, up and running.
A Startup Coach Can Be Valuable
In the real world, the process of creating and testing MVPs is easier said than done. Doing so often requires experience and unconventional thinking that may not be native to the founding team. It is therefore helpful to engage a startup coach for the research team in the early stages, when the idea seems daunting and many insecurities about the path ahead exist. Universities are often open to engaging external advisors for startups. Of course, this should not be theoretical advice or yet another weekend workshop in startup entrepreneurship. An MBA or economics student from another faculty who will work on the case as part of their master’s thesis also is not enough. The team needs a dedicated, experienced entrepreneur who will enjoy working with them for several weeks in the lab to address their challenges.
Research shows that it is often very effective to learn a new skill by seeing an expert doing it and then repeating it.3 This approach trumps theoretical advice any time, as well as self-directed learning by doing. If startups have access to a seasoned entrepreneur who works with them one-on-one, this can make all the difference. It can accelerate the learning curve by two or even three times while avoiding many early common pitfalls and distractions. If the team has built their own network and reached out to other entrepreneurs, they may suggest someone they see fit as their advisor. A startup coach should obviously have firsthand entrepreneurial experience, but domain knowledge is unnecessary. The whole point of working with a coach is to get an outsider’s view and a fresh perspective on the team’s challenges. Domain knowledge exists already on the team in the form of the founders. The added value of the startup coach lies more in management: getting the startup out of the building and into the market as soon as possible.
Do You Need a Business Plan?
A business plan in the classic sense is a thing of the past. It is largely a thought experiment with little merit once a business has gotten underway. Some entrepreneurs admit that but insist the business plan can be a guideline to keep them on track and their thinking focused. To me, it is the other way around: once you start, you need to be as flexible as possible to pivot and incrementally change your products and services. A rigid 50-page document will only hinder you in that regard. I have never written a business plan for any of my businesses in almost 20 years. Launching a startup is an exercise not in thinking, but in doing. The practice of requiring a written document before taking any action is against the nature of entrepreneurship. Flexibility and time to market are much more important than perfect planning and consideration of eventualities. A business plan stands in the way of getting started.
Of course, you still need to think about your startup in a systematic manner before taking the plunge. This is important, but you need to do it with an open mind. A relatively new approach is the business model canvas, which is basically a business plan condensed onto one page.4 It does away with the need for speculation about potential revenue down the road and focuses on a few specific key questions about the business and its market. An updated version of the business model canvas has been proposed by author Ash Maurya.5 His lean canvas is what I use for the startups I work with (see Figure 2-4). Although the original canvas is good to analyze a business, the lean canvas helps entrepreneurs get started quickly. I also find it more useful when I need to discuss the business with third parties.
Figure 2-4. Lean canvas, adapted from the Business Model Canvas (www.businessmodelgeneration.com), licensed under the Creative Commons Attribution-Share Alike 3.0 Un-ported License
When filling out the canvas, make a different one for different assumptions. For example, if your startup is in Singapore, you should show how it applies to the local market, to the wider market in Southeast Asia, and to all of Asia including China. Analyze different scenarios for your startup on different canvases. If you speak with different third parties, make a canvas for each of them. The problems, solutions, unique value propositions, distribution channels, customers, costs, and pricing models will vary from market to market. This is a good exercise to understand the platform you are playing on.
When you begin, you will quickly find that all this looks simple but is far from intuitive. Prepare yourself for the fact that your first canvasses will rarely survive much scrutiny and will need an update soon. This is fine, because the canvas is the opposite of a business plan set in stone; it is a fluid document you use to organize your startup and put the different puzzle pieces together over time. Treat the canvas as a work in progress, and carry it around with you. Update it as needed, when new findings emerge from market feedback or you feel a pivot is necessary. The lean canvas and the financial model, which we will visit later are both the cornerstones of your toolkit to present your startup to the world.
1Eric Ries, The Lean Startup (New York: Crown Publishing Group, 2011).
2Steve Blank, “Why the Lean Start-Up Changes Everything,” Harvard Business Review (May 2013).
3Robert Greene, Mastery (New York: Viking, 2012).
4Alexander Osterwalder and Yves Pigneur, Business Model Generation: A Handbook for Visionaries, Game Changers, and Challengers (Hoboken, NJ: Wiley, 2010).
5Ash Maurya, Running Lean: Iterate from Plan A to a Plan that Works (Sebastopol, CA: O’Reilly, 2013).