This is advice for people who have done scientific research at a university and are considering starting a company to commercialize it.
At YC, we've funded more than 75 companies in this situation. We also recently went on a Bio Tour where we went around to research universities and talked with hundreds of students and professors in the life sciences about commercializing their research. These are the most common topics founders have asked us about.
Deciding who should spin out
In a typical spin-out situation, there are several people who worked on the research, including a mix of students, post docs and faculty. The first thing you need to decide is who is going to work on the company and who is going to stay at the university.
A lot of scientific founders have misconceptions about how to structure their founding team. Here are the most common ones.
Misconception 1. You can start a company while continuing your academic career
Here is a blunt fact that often makes founders uncomfortable: your company has little chance of success unless someone who worked on the original research is willing to leave their university role to start this company.
If you are a student graduating soon, you can just wait until you graduate; that’s a perfect time to start a company. But otherwise, at some point you will have to make an intentional decision to leave so you can run the company.
Misconception 2. You should find a CEO to run the company
Too many scientists believe that they should remain in academia and find a CEO to start a company around their invention. This is almost always a bad idea. For one, it’s hard to find a great CEO to run a company at this early stage. Truly great CEOs are scarce, and they usually have much better opportunities available to them than running an idea-stage startup with no funding. As a result, most scientists that try this approach either never find a CEO or end up settling for a mediocre one. It’s even worse if your university tries to find a CEO for you.
But even if you could draft any person in the world to run your company, it still probably wouldn’t be a good idea. The best CEO for this stage is one of the people who did the original research. The people who did the original research will be far more invested in the success of the venture than any outsider. They are also far more qualified to build a company around it because their domain knowledge of the field is much more valuable than whatever general business skills an outside CEO would bring.
A related misconception is believing that the research is done and that all that’s left is to commercialize it. If this were true, perhaps an outside CEO would make sense. However, it rarely works out that way. Usually you find that the thing the market wants is not quite the thing that you’ve invented, and that more research needs to be done. The original inventors can take this feedback and make adjustments; an outside CEO will just be stuck.
Misconception 3. You need someone with business experience on the founding team
Many scientists think that to start a company you need someone with prior business and financial experience. This is just not the case. In the first couple of years, there is typically very little "business" to be done, and whatever business skills you need you will pick up along the way. Most of the scientists we fund at Y Combinator have no prior experience in business.
People who work in business like to make it sound hard, as if business were like quantum physics, a field that needed to be studied for years to master. The fact is, it's not even close.
Misconception 4: You should raise money first, then leave the university
Often, people are unsure of whether they want to risk leaving a stable academic role to pursue a startup. So they take their idea and pitch it to some local VC firms. They figure if it's a good idea, the VC firms will fund them, validating the idea and giving them a smooth transition out of their university job and into a well-funded company.
While VC firms will occasionally fund spin-outs this way, usually they don't. Unfortunately, too many founders get turned down by VCs and conclude that their idea must be bad and give up. Actually, the issue is that it's just too early at this point to raise money from VCs.
Typically, founders will need to work full-time on their company for 1+ years before it is ready to raise a multi-million dollar round from VCs. In the meantime, they sustain themselves by self-funding from their savings, getting government research grants, raising a small amount from friends and family, or raising a small "pre-seed" round from angel investors, accelerators, or seed funds.
Founders who won't quit their job before they raise money often get stuck in a catch-22. They are waiting for an investor to take a bet on them before they quit their job. But the investors are waiting for the founders to believe enough in their own company to quit their job!
What I recommend
The ideal situation is that two or more people from the lab who did the work leave together to start the company as cofounders. One full-time founder is also ok. One of the people who leave to start the company should be the CEO.
In many cases, other people who were involved in the research want to remain behind at the university but still contribute in some way. That’s fine. Those are often called “academic cofounders” or “scientific cofounders” and they can still be very helpful. But the founders who are going to be full-time are the most important.
Deciding when to spin out
In the early stages of developing a new technology, you’ll make faster progress still at the university, taking advantage of university resources. It’s the ideal place to do the initial experiments to prove that your idea could work. You can even do some testing of market demand for a new product, through programs like NSF I-Corps or just by calling up potential customers/stakeholders. At some point, though, that will flip, and being at a university will start to slow you down, because universities are not set up to commercialize technologies.
It’s possible to leave too early and possible to wait too long before leaving. Unquestionably, though, the far more common mistake is to wait too long.
Most founders wait too long because leaving is scary. Academia is a comforting environment. No one is pressuring you to leave and leaving seems risky so the natural thing to do is to keep delaying it. There's a temptation to make the technology perfect before spinning out, and there's always "one more experiment" you could do. If you don't stop this cycle, you'll never leave.
Often after people do leave, they realize that a lot of the work they did in the last year was wasted, because some of their assumptions about what the market wanted were wrong. They also realize that they are now moving so much faster as a company that they could have saved months of time by spinning out a year earlier.
Splitting up Equity
After you've decided who is going to be full-time on the startup and what everyone's role will be, you'll want to split up equity.
As important as this decision is, founders often don't have a good framework for making it. Here is the framework I recommend. It has just two rules1.
Founders who will be working on the company full-time should get equal or nearly equal amounts of equity.
Founders who will be leaving their job to work on the company full-time should get much more equity than founders who are going to remain in academia. Academic cofounders should typically own no more than 10% unless they are going to continue to be hands-on2.
My colleague Michael Seibel previously wrote a great essay about why rule #1 is so important. Rule #2 is important because it is the full-time founders who will invest years of their life exclusively in making the company successful, and they need to have enough ownership that it makes sense for them to do that.
The biggest conceptual mistake I see scientific founding teams make here is that they think the purpose of allocating equity is to reward past contributions, when actually it’s mainly to anticipate future ones.
Here's a blunt fact about starting a company. If you are going to make a new company successful, you will probably have to work on it for 7-10 years post spin-out. That's a long time!
If you are just spinning out of a university now, you might feel like you're halfway done, but actually you are on mile 2 of a 26 mile marathon. The academic founders may have been instrumental in the first mile, but it is the full-time founders who will be primarily taking you the other 25. The equity split between founders has to reflect the expected contributions over the whole marathon.
One consequence of this is that your equity split in the new company will not necessarily have any relation to your seniority within the original academic team. It’s often the case that the people leaving are more junior, while the senior people / faculty remain. In that case, the founders who leave will end up with much more equity than their former boss. This can be an awkward conversation, but it’s entirely sensible.
Negotiating with tech transfer offices
If you are going to commercialize research started at a university, you will probably need to negotiate the rights to the intellectual property. The group at a university that does that is the technology transfer office.3
In the past, tech transfer offices had a well-deserved bad reputation. They were known for being slow and bureaucratic, and for forcing onerous terms onto fragile young startups. Many times the terms they insisted on strangled the very companies they were trying to create. There was so little transparency in the industry, it was hard for founders to know what terms were fair.
Fortunately, things have gotten better. There is now much more information available for founders. Tech transfer groups at the universities in major startup hubs like Harvard, MIT and Stanford now give startups reasonable terms (though they still take too long to do it). At universities that have not seen many successful spin-outs, it's hit-or-miss. A few universities are now using “express license agreements”, preset agreements that require little to no negotiation; hopefully this will become more common.
There are typically four key terms4 in these agreements.
Royalty. This means that you pay a percentage of revenue or profits to the university. If this is too high, it can affect the viability of the company to raise money and operate. Ideally you would make this zero. If you can't do that, try to keep it < 5%, and to have it terminate after a certain number of years and/or a certain level of payments.
Milestone payments. I.e., "You owe us $250K when the company raises its first $10M", or “You owe us $500K when you reach Phase II clinical trials”. Because cash is scarce in the early days of a startup, you want to keep these as low as possible. You should never need to spend more than a few percent of the money you raise.
Exclusivity. If a license is not exclusive, the university could theoretically turn around and license the same IP to a big company to go compete with you. This sounds like a real problem, but often it’s not. For many inventions, in practice other companies won’t know how to use the IP and won’t value it until you've done years of work further developing it (at which point the university-owned IP isn’t sufficient).7 It may be optimal to have a non-exclusive license initially with an option to make it exclusive later, or a right of first refusal clause.
Here is some advice for negotiating these agreements.
You should get in touch with founders of other companies that have recently negotiated agreements with the same office. Find out what terms they got and ask for advice on negotiating strategy. You can also ask investors, lawyers, and advisors. You should get as many data points as possible.
If you’re a student or post doc, it’s valuable to have the buy-in of the professor running your lab. Professors have sway at universities and will give you leverage over a stubborn tech transfer office. You also want to make sure that they don’t have any competing plans to do their own spin-out with the technology. Often by getting them onboard as an advisor early, they will be helpful in securing a good deal, and they’ll also give you credibility with investors.
If the agreement feels too onerous, ask yourself if you need it at all. It might be cheaper to recreate something similar on the company’s time.8
More dramatically, you might end up not even using what you’re planning to license, because startups pivot all the time. I’ve worked with many companies that fought tooth and nail over their tech transfer agreement, only to find that a year later they'd totally changed their approach and abandoned the IP they fought so hard to license! One way to protect against this is to ensure that any royalties are directly tied to the use of the technology.
Consider taking an option to license the IP in the future, rather than negotiating a full license agreement now. An option is often much cheaper and simpler to get, and allows you to defer the final negotiation for six to twelve months. That also gives you time to see how much you are using the original IP before committing to licensing it.
Beware of well-meaning but bad advice from university entrepreneurship offices. Some entrepreneurship offices at universities are great, but unfortunately some are not. Worse, some have their own agendas, like helping local investors. Consider whether the people you’re talking to have a track record of many truly successful companies when listening to their advice.
Start the conversation with this office as early as possible. This will give you more time to work out an agreement and also let you find out the lay of the land.
Don't wait for the agreement to start the company. Getting an agreement can take 6 months or longer. Many investors (including YC) will fund companies before they have an agreement in place. The more progress you make on the company, the more leverage you have in the negotiation.
After the spinout
You’ll need to incorporate your company. If you are based in the US (and possibly even if you aren’t), you’ll want to incorporate as a Delaware C Corporation, no matter which state you are physically in.9
It may well make sense to continue collaborating with your lab. They may produce new work which you want to license.
In some cases, you may want to continue doing experimental work using university labs. University core facilities are commonly available to companies, albeit for higher fees. It’s possible to save a lot of money using university resources instead of buying equivalents commercially. That’s fine, as long as it isn’t slowing you down significantly and doesn’t create IP issues. Unfortunately there is often a tradeoff between speed and cost.
A big adjustment for founders from academia is internalizing a different incentive structure. In academia, you’re rewarded for new discoveries and for publications. In startups, there is zero reward for new discoveries and hardly any for publications. In startups, the only thing that you are rewarded for is making tangible progress towards a commercially valuable product.
A side effect is that in startups, there is no intrinsic reward for doing something new or difficult. Any time there is a shortcut where you can copy or buy something that already exists, you should take it. In startups, you want the “new part” of what you are doing to be as small as possible, and everything else to be as boring and low risk as possible.
Another big adjustment is the pace. In startups, you are racing against the clock. If you don’t hit milestones before your current funding runs out, your company will run out of money and die. That kind of hard constraint forces focus like nothing else. As a result, founders who leave academia to do YC often tell us they got more done in the three month YC batch than the prior year.
If that’s made startups sound stressful, they are, if only because so much is at stake. When you’re starting your own company, the highs are higher and the lows are lower than any regular job. If you want to learn more about what it feels like, there’s a great essay called “What Startups Are Really Like”.
There has never been a better time to start a biotech or hard-tech company. There is far more funding available now than ever before, and a well developed ecosystem to support founders at every stage. Today most grad students and postdocs choose between staying in academia and getting a job in industry. We think there will increasingly be a third option: to start their own company.
I couldn’t have written this without the help of many YC founders who did spin-outs themselves. Special thanks to Uri Lopatin, Birgitt Boschitsch, Aaron Lazarus, Alexis Rovner, Juan Medina, Glenn Markov, John Ramunas, Wesley Wiersen, Lindsay Amos, Ravi Pamnani, and Andrew Jajack for reading drafts of this.
1. Separately, it’s also very important for all founders to have a vesting schedule for their shares. It should vest over at least four years.↩
2. Being “hands-on” means more than giving advice. It means spending at least 1 day / week doing active work for the company - writing grants, being in the lab, going to sales meetings, etc. Their equity vesting should be tied to them continuing to take a hands-on role.↩
3. The names vary a little - sometimes it’s called the “office of technology licensing”, or the “innovation office”, but I’ll use this term.↩
4. Here are some less critical but still common terms and quick advice for them: (5) patent prosecution - you should have control of the patent filing and patent strategy, (6) sublicensing - likely important for you to be able to sublicense, (7) funding requirements - try to avoid, (8) upfront and annual fees - ok if relatively small, (9) rights to future IP coming out of the same lab - may or may not be worth it, (10) pro rata rights so the university can invest to maintain their ownership - acceptable if the university can decide quickly and if it only applies to priced rounds, not convertible securities.↩
5. Sometimes the terms call for a “percentage of proceeds from any liquidity event” instead. This is worse than equity because it can’t be diluted. So if your university insists on this because they can’t accept equity, it should be a very small percentage, about ⅓ of the equivalent equity percentage.↩
6. Another wrinkle is anti-dilution clauses where these equity percentages don’t dilute until you’ve raised $xM.↩
7. Related to this, some universities require a “marketing period” where they shop the invention around to other potential buyers but usually this is a formality because they rarely get interest.↩
8. Also, if you have a reasonable path to recreating it, that can give you leverage over a tech transfer office. If they know you are busily working on recreating it, the value of the license deal goes down every day.↩
9. Two easy options for doing this are http://clerky.com/ and https://stripe.com/atlas.[↩]