Going from research to a startup: 3 things I got wrong in my first 4 months

Doing a PhD and building a startup have a lot in common. You decide on the topic you work on, the input/output relationship is more direct than in a big company, and the sense of ownership of your work is greater. But transitioning from research to entrepreneurship still requires some mindset shift, and if I had to start again, I would definitely do things differently.

For context, the BRIDGE Proof of Concept is a Swiss federal grant that is given to a federal research institution like ETH Zurich to employ you for a duration of 12 months. You’re getting up to 130’000CHF, which largely covers salary and research costs. The mandate is to “support researchers on their way to entrepreneurship”, as outlined in the first article of their regulation. I researched shape generation for embroidery during my PhD and wanted to create a company around this idea, so this was very well fitting. I applied and got the grant.

Talk to customers first

My PhD student contract with ETH ended in January 2026. It was not mandated at all to build a company, but to complete my PhD, which primarily involves doing research! I defended in November and deposited my thesis in January, but also spent my time until the last minute doing research. We submitted a paper to SIGGRAPH, which got in and will be presented next month in Los Angeles. While the paper is related to embroidery and its technology might end up powering advanced embroidery auto-digitization, we released the code under MIT license. So the transition from my PhD position to my BRIDGE-funded position has actually been somehow abrupt: I was actively pushing for an academic deadline at the end of January, and at the start of February, I had to become an entrepreneur.

During my studies, I followed entrepreneurship programs like Talent Kick or Startup Campus that are aimed at training students. So I knew theoretically that you don’t start a company by solving a technical challenge. You start by trying to understand what’s worthy to solve. But still, I just finished a research push at that time, so I started February with what I was trained to do: technology development. I was building cool stuff without checking whether I was the only guy in the world who cared!

After one month, I took the VentureLab’s Business Creation ICT course. The first day emphasized having “customer interviews”. This expression still sounds funny to me because I associate the word “interview” more with hiring people than with customers. Also because they are not yet customers, and the principle is not to sell an idea, but to understand what people do so that you can uncover what they might want.

If you don’t know where to start, I would strongly recommend you read The Mom Test. It’s packed with useful information and actionable rules, and on top of that, it’s pretty cheap. One of my favourite “rules of thumb” from the book is that there is no such thing as a “meeting that went well”. It either ends with a commitment or not. My very first meeting with a potential customer ended with a “we will contact you soon”. I wish I had read The Mom Test before to understand that I had to push for a date for the next meeting or for meeting someone else in their company. Instead, I thought that this non-committing sentence meant that they were interested. I ended up the meeting thinking I should work on what they told me they wanted!

Spoiler: they never contacted me back.

Seek mentorship as soon as possible

The BRIDGE grant comes with a CHF 2,000 voucher for coaching. I didn’t realise I could simply use it, so two months in, I applied through the Innosuisse platform for the CHF 5,000 ‘initial coaching’ voucher instead, declaring my BRIDGE fellowship in the application, since both come from Innosuisse. In hindsight, I may not have needed the second one, but extra coaching funding doesn’t hurt, and everything was above board with both BRIDGE and Innosuisse.

Having an Innosuisse coach is wonderful, and you should do that as early as possible. The main reason is that you get someone who is paid by Innosuisse and who does not have a conflict of interest, so they can tell you what they think is in your interest to do. They don’t tell you what to do (you’re the one deciding), but they can answer some questions very precisely, because they can see things you can’t see due to their experience.

For example, I was advised by someone else to incorporate an AG (LTD) and not a GmbH (LLC), mainly because an AG allows for a way more flexible way to handle share distribution. But an AG requires CHF 100,000 of share capital, of which at least CHF 50,000 must be paid into a blocked account at incorporation (the rest is subscribed but can be paid later). A GmbH only requires 20,000CHF. I am a solo founder anyway, and I wanted to incorporate now so that I could separate the legal entity from me (e.g. for liability reasons) and so that I could process payments easily. Why? Because I want to check whether there is a willingness to pay for my product, which in turn is required to understand whether I reached Product-Market-Fit. An AG, at my stage and as a solo founder, doesn’t make much sense. If I realise that nobody wants what I’m building, then it’s easier to liquidate a GmbH anyway. My coach advised me that way, and it unlocked my decision to incorporate. Had I gone through the AG path, I would probably have asked an angel investor to de-risk my financial risk. Or I might not have made the jump.

The VentureLab course is another example where you get to learn about how to create a company from people who have done it before. They’ve been in the game, and they answer your questions precisely on many topics. Of course, some topics might come a bit too early to land, and I will need to revisit the resources when the time comes, but it doesn’t take much of your time to attend, so the downside is practically 0. The free opportunities to get knowledge in the startup ecosystem in Switzerland are amazing. Use them.

Clarify IP now

ETH has a special office for transferring IP and licensing. If you want to incorporate a company based on your research results, you can follow these steps to become an ETH spin-off. You will notice right away that you need to prepare for an introductory meeting 6 months before incorporation. After talking to one of the responsible people, they told me they can be very busy (they’re in demand and we should respect that). It’s really no joke: it does take time, and you should handle that way before incorporation.

After developing the first version of my software prototype, I went on Reddit to ask whether some people would be interested. And some people were! So as said previously, I now need to collect real data about willingness-to-pay to confirm that I can reach PMF with my product. For this, I need a company. For this, I need to have clear IP separation. And for this, I need to talk to the ETH IP transfer office. I started to send emails in mid-April and could get a meeting just before the beginning of May.

That was way too late. What was done during my BRIDGE fellowship, and hence technically under my ETH employment time, was the software prototype. Hence, this software belongs to ETH. So I need to have the rights from ETH to use it commercially. So to use it for a company, I would need to enter long negotiations over a prototype that only took two months to build. If it really takes 6 months to negotiate it, it means that I would have to wait October-November just to be allowed to incorporate and check PMF. If I take 3 months to check PMF, I’m already at the end of the BRIDGE fellowship. In the meantime, I could do research on topics that nobody cares about, because I can’t get reliable data about what people actually want. Not good (and it’d be a waste of taxpayer money).

Two things made this workable. First, the research itself (the part done during my PhD) was open-sourced under an MIT license, so the underlying ideas are free for anyone, including me, to build on commercially. Second, and more importantly, I talked it through with the transfer office directly. We agreed on a clean separation: the prototype code developed during my ETH-funded BRIDGE time stays at ETH as part of the fellowship, and the GmbH operates on code I re-engineer during GmbH time. I confirmed this scope with them in writing and in a meeting, and incorporation went ahead on that basis. As a bonus, the re-implementation is structurally better than my first attempt, so the few weeks aren’t entirely wasted.

Eventually, all of that extra work could have been avoided if I had clearly taken the time to understand the IP constraints before I started working on any prototype. My mistake; hopefully you won’t do the same as I did. My goal is to now do research during my ETH time that could later be used for the company, and incorporate an AG at this moment with ETH licensing. Provided that I can make it work, of course. But because I will be able to check PMF, I know that if it works, it will more likely prove useful to people.

Summary

Three things I’d do differently starting again: (1) talk to potential customers from day one, before building anything (and read The Mom Test first). (2) Get an Innosuisse coach immediately; the guidance is worth far more than the time you spend applying for the funding. (3) Clarify your IP situation with your institution as soon as possible: it takes months, and getting it wrong can cost you a lot of time or money.