Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.


How founding a company compares to graduate school

Grey budgie standing apart from green budgies in a branch.

Credit: Michael Blann/Getty

They say there’s no such thing as a free lunch, and yet that was exactly what was on the menu once a week during graduate school. Every Thursday, PhD students in our department took a guest speaker from a different university out to a local Indian buffet for lunch — paid for by our university. We’d all give two- or three-minute summaries of what we were working on, get some feedback and guidance, and then spend the rest of the lunch getting advice from someone outside our bubble on our life, research and career struggles.

The free lunches were all fun and games until, like clockwork, the question came: are you all applying for jobs in academia? Furtive glances. Fallen shoulders. Knowing looks all around the table. Closest friends shared smirks that betrayed a thousand private conversations. But we all knew the drill: the two bravest and most senior PhD students would take centre stage to recount their latest high-impact contributions and ask for any insider information on which professors might be interviewing and where.

Even though we all know that the number of graduating PhD students far outweighs the number of academic-job openings, free lunches are about as close as most graduate programmes come to enabling candid conversations about life outside academia. Most of us learn almost nothing about working in industry and the ways in which your career as a scientist might change.

I received my PhD from Yale University in New Haven, Connecticut, focusing on treatments for mental illness. The team I was in published research showing that machine-learning algorithms can predict treatments that are likely to be successful in treating mental illness1,2,3, potentially reducing the amount of trial and error in treatment and the total cost of care.

I co-founded a health-care company in the third year of my PhD, along with two other students at Yale, because it felt like the fastest and most efficient way to bring this technology to clinical practice. Start-up companies can raise funding, hire people and partner with clinics much faster than would be possible in academia.

Some of my closest friends from graduate school have gone on to work in data science and analytical roles at health-care and technology start-ups. Here’s what we learnt during our transitions.

You won’t have to do everything any more

Graduate students have a host of responsibilities during research projects: reviewing literature; designing experiments; drafting ethical protocols; doing statistical analyses; writing custom software and hardware; the recruitment, scheduling and coordination of participants; presenting findings; and writing reports.

In industry, even in small companies, there might be a team of five or six specialized people — probably being paid more than they would be in academia — dividing up these responsibilities, with job titles such as statistician, data analyst, software engineer, operations coordinator and medical science writer.

Good science takes place in industry

For many, academic science is the purest kind of science — especially for those with tenure who don’t have commercial conflicts of interest and who are funded by government bodies. But this doesn’t mean that the science happening in industry is any less rigorous. In fact, companies might have an even greater need for quality to ensure that their products function as intended and outperform their competitors’.

People holding drinks stand and sit in rows in a casual office.

Staff members at a Spring Health party.Credit: Adam Chekroud

The guiding principles for our company have always been scientific thinking and the scientific method. We form hypotheses about what customers want or what users will prefer, define metrics for success, design the smallest experiments that could test our hypotheses and then use the results to inform our thinking. In practice, we measure everything, test many things and iterate constantly to understand what is working.

Science in industry happens faster

Unlike academic institutions, most companies measure their progress in quarterly performance reviews. This means that they often need to measure results and finish projects quickly. The science that does happen is often more applied, with clearer objectives and definitions of success; it’s also better funded and resourced compared with projects at universities. It is normal to have to show weekly progress. This is quite different from academia, where there is less time pressure to demonstrate progress, but also less opportunity for structured feedback.

You might miss some aspects of academia

In graduate school, everyone around you is probably interested in, and good at, the same things as you. In industry, people have a wide range of skills: sales, marketing, human resources, customer satisfaction, operations and management.

Although your colleagues will all be working towards a common company vision and mission, you might not ‘nerd out’ with them about that new software library, biological pathway or preprint article. I miss the heated debates we used to have over lunch about random theories, designs or projects that weren’t constrained by timelines or real-world use.

Setting up a company or working at a start-up is like getting your PhD

Both start-ups and research projects attempt to do something that has probably never been done before. In both cases, you have to persevere despite ambiguity, uncertainty and risk. You take a big, amorphous goal, and you break it up into smaller, testable hypotheses to create clarity as you go. And in both you have to learn new skills quickly. Throughout graduate school, this happens in lab meetings, seminars, courses and research projects, and in industry you have to learn how to build products, do user testing, market effectively and sell compellingly. (And it might not work out even if you do everything right!)

Some elements of building a company were new to me, and I had no parallels in graduate school. Hiring people, developing a user experience, finding and defining your customer, and building a sustainable business model are all challenges that founders face. But many of the logistical barriers to starting a company are easier to conquer than ever. Online services such as Stripe Atlas can give you the paperwork you need to set up your company, bank accounts and web servers. Many other companies offer software that make human resources, payroll, and financial administration easier. These types of tool allow you as a founder to focus on the things that you’re good at, and to spend less time on administration.


This is an article from the Nature Careers Community, a place for Nature readers to share their professional experiences and advice. Guest posts are encouraged.


  1. 1.

    Chekroud, A. M. et al. Lancet Psychiatry 3, 243–250 (2016).

    PubMed  Article  Google Scholar 

  2. 2.

    Chekroud, A. M. et al. JAMA Psychiatry 74, 370–378 (2017).

    PubMed  Article  Google Scholar 

  3. 3.

    Chekroud, A. M. JAMA Psychiatry 74, 1183–1184 (2017).

    PubMed  Article  Google Scholar 

Download references


Nature Careers


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing


Quick links