From researcher to founder: the identity and skills shift
Core idea#
Research trains you to be right. A startup needs you to be useful, quickly, with incomplete information. The skills overlap more than people think (forming hypotheses, designing tests, persisting through failure), but the goal, the pace and the audience are different, and some research habits that served you well actively get in the way.
The shift has two parts:
- Identity. Your sense of worth stops coming from papers, citations and peer respect, and starts coming from customers, revenue and the team. Many researcher-founders find this harder than learning any new skill, and few talk about it.
- Skills. You need a working level, not expert level, in selling, hiring, managing money, negotiating and making product decisions. Most of these can be learned in months through deliberate practice, not years.
This playbook names the habits to keep and the ones to drop, and gives a 90-day plan to close the biggest gaps before you need them.
Who it's for: researchers in the 6 to 24 months before or after founding, and PIs considering a CEO or CTO role in their own spinout.
What the evidence says#
Few PhDs want to found; many want to join
A survey of over 4,000 US science and engineering PhD candidates found 46% interested in joining a startup as an employee, but only 11% who expected to found one (Science Connected, on Roach and Sauermann). Founder interest was most strongly tied to personal preferences for autonomy, risk and managerial work; an entrepreneurial environment made people more likely to want to join a startup, but did little to create founder interest in people without those preferences (Roach and Sauermann, Management Science 2015).
The practical point: if you find the commercial side draining rather than interesting, being an early technical hire or CSO (playbook 2.12) may suit you better than being CEO. That is a legitimate and valuable path, not a lesser one.
Customer discovery training works#
NSF's I-Corps programme, based on Steve Blank's Lean LaunchPad method, trains researchers to interview customers before building. Its 2025 biennial report says 52% of National I-Corps teams since inception have been linked to startups, and that 57% of entrepreneurial leads in FY 2023 to 2024 were graduate students (NSF). The method is the single most useful tool for the identity shift, because it forces you out of the lab and into conversations where your expertise matters less than your listening (playbook 1.19).
Industry experience is the strongest predictor#
Prior work in the industry a startup enters more than doubles the chance of a high-growth outcome (Jones and Kim). For an academic, the closest substitute is deep, early exposure to customers, industry partners and the people who would buy, plus a co-founder or early hire who has worked in the industry.
Habits to keep, adapt and drop#
| Research habit | In a startup | What to do instead |
|---|---|---|
| Hypothesis-driven work | Keep it. This is your biggest advantage. |
Apply it to customers and the business model, not only the technology: "Utilities will pay for X because Y; we'll know in 20 interviews." |
| Getting it right before showing anyone |
Drop it. Months of polishing before feedback burns runway. |
Show rough work early: a one- page spec, a slide, a prototype that half works. Ask "what's wrong with this?" |
| Precision in language | Adapt it. Precision about claims stays essential (investors and customers will check). Precision as caveats on every sentence loses the room. |
Lead with the conclusion and the number. Keep the caveats for the follow-up and the data room. |
| Consensus and peer review |
Adapt it. Advice is useful; waiting for agreement isn't. |
Decide with 70% of the information, write down why, and set a date to revisit. |
| Research habit | In a startup | What to do instead |
|---|---|---|
| Depth over breadth | Adapt it. You still need depth in the core technology; you now also need breadth. |
Get to "good enough" in sales, finance and hiring. Hire or partner for depth there later. |
| Credit through authorship |
Drop it. In a company, credit goes to the team and the result. |
Give credit away generously; it builds the team. Publishing becomes a strategic choice (playbook 2.04), not the goal. |
| Grant-writing | Keep it. It's a real skill and non-dilutive funding matters in deeptech. |
Write grants for milestones that investors also value, so one piece of work serves both. |
| Working alone on hard problems |
Adapt it. Hard thinking still matters; isolation doesn't. |
Block deep-work time, but spend at least a third of the week with customers, team and investors. |
| Time measured in years | Drop it as the default. | Plan in quarters, with monthly checks. Long-horizon R&D still gets years, broken into milestones. |
The identity part#
Three shifts that founders who came from research often describe:
- From "expert" to "learner". In the lab, you were often the person who knew most. In a sales meeting or a board meeting, you may know least. Treat it as fieldwork.
- From "my work" to "our company". Your best contribution may soon be hiring someone better than you at the thing you used to do.
- From peer approval to customer evidence. A colleague saying "elegant" counts for less than a customer saying "when can we have it?". This can feel like a loss. It helps to keep a foot in the research community (a seminar, an adjunct role, an advisory board) so that part of your identity isn't simply cut off.
Closing the skills gap: a 90-day plan#
Step 1: Find your gaps#
Rate yourself 1 to 5 on each of the founder skills below. Then ask one person from industry to rate you. Pick the two lowest to work on first.
| Skill | What "good enough" looks like at pre-seed |
|---|---|
| Customer discovery | You've run 30+ interviews and can say what customers would pay and why. |
| Selling | You've asked someone for a pilot, a letter of intent or a purchase order, and heard yes or no. |
| Storytelling and pitching | You can explain the company in one sentence and a 10-minute pitch to a non-specialist. |
| Money | You can build an 18-month budget and runway model and explain it. |
| Hiring | You've written a job description, interviewed and made an offer. |
| Negotiating | You've negotiated a contract or licence term and know your walk-away point. |
| Managing people | You've given someone difficult feedback and set clear goals. |
| Product decisions | You can say what you'll build first, what you won't, and why. |
Step 2: Learn by doing, in small, real steps#
- Weeks 1 to 4. 15 customer interviews (playbook 1.19). One pitch to a friendly investor or incubator manager, for practice. Build a first budget.
- Weeks 5 to 8. 15 more interviews. Ask for one pilot or letter of intent. Shadow a founder for a day, or sit in on a sales call.
- Weeks 9 to 12. Write the one-page thesis (playbook 1.20). Pitch three times. Re-rate your skills and choose the next two gaps.
Step 3: Fill the rest with people#
You don't need to be good at everything. A business co-founder, an experienced adviser or a fractional CFO can cover what you can't learn in time (playbooks 1.12, 1.18, 3.04).
Worked example#
Arjun, a materials science postdoc, scored himself lowest on selling (1) and money (2) . In 90 days he ran 32 interviews with packaging companies, built a runway model with help from his incubator's finance mentor, and asked two companies for paid pilots. One said
yes, at a small fee. Selling went from 1 to 3, and he learned that his first customer cared more about cost per tonne than performance, which changed the product plan.
Differences in India, the US and the UK#
- India. Status attached to academic and government posts can make the identity shift feel like a step down in the eyes of family and peers (playbook 1.17). Institute incubators (1.11) often pair founders with mentors from industry, which is the fastest route to commercial exposure. Formal customer discovery training is less common; running your own I-Corps-style interviews is a real differentiator with investors.
- US. NSF I-Corps, NIH I-Corps and DOE programmes are widely available through university hubs, and many universities run founder bootcamps. The culture celebrates founders, which can push people into the CEO role who would be happier as CSO or early technical hires.
- UK. Programmes such as ICURe (Innovate UK's customer discovery programme for university teams) and Royal Academy of Engineering Enterprise Fellowships fund time for this shift (1.10). Spinout culture often assumes an external CEO will be brought in (3.04); decide early whether you want the role.
Programme names and eligibility change; check current details on each programme's site.
Scripts#
Opening a customer interview without pitching#
"I'm not here to sell you anything; there's nothing to sell yet. I'm a researcher trying to understand how you handle [problem] today. Can you walk me through the last time it came up?"
Asking an industry contact to rate your skills#
"I'm working on moving from research into running a company. Here are eight skills. Could you rate me 1 to 5 on each, based on what you've seen? The low scores are what I most need to hear."
When an investor asks "Are you the CEO?"#
"For the next 18 months, yes. I'm closest to the technology and the first customers. I've been building commercial skills deliberately [give one example]. If we reach a stage where the company needs a different kind of leader, I'd rather see that early and help bring them in."
Common mistakes#
- Waiting to feel ready. The commercial skills come from doing them badly at first, not from courses.
- Treating sales as beneath the science. Early customers are a source of data. Selling is the experiment.
- Over-caveating in pitches. Precision is a strength; hedging every sentence reads as low confidence.
- Hiring only researchers. Teams of people like you share your blind spots.
- Cutting off the research identity completely. Keeping a connection to your field helps morale and recruiting.
- Assuming you must be CEO. Many successful scientist-founders are CSO or CTO.
Related playbooks#
- 1.01 Should I start a company?
- 1.12 Finding mentors and early advisers
- 1.16 A fast business curriculum for PhDs
- 1.18 Solo founder or team?
- 1.19 Customer discovery from inside the lab
- 2.12 Professor as co-founder, adviser or CSO
- 3.04 Bringing in a business CEO
Sources#
- Roach and Sauermann, "Founder or Joiner?", Management Science 61(9), 2015 (abstract; Science Connected summary; Cornell News).
- NSF, 2025 I-Corps Biennial Report.
- Jones and Kim, "Most successful entrepreneurs are older than you think".
Last checked November 2026.