Welcome back to our course on commercial open source software startups and how to spin one

off from university. We're in the third and final part of the course looking at how software

startups work and how to spin them off from a university context. Today's topic is how research

at a university relates to the startups we are talking about. It's probably a bit shorter because

this part of the course is still developing rapidly and future versions of this talk will

be expanding on the topics. Still we need to understand how research from a university or

research lab context makes it into a potential startup and more importantly first of how to

manage such research that it can possibly be turned into a startup. Obviously not all research becomes

a startup. So we will look at how software development and university context takes place,

how to build a successful open source project from that and then in the next lecture we will

look at specifically how to create a how to spin off the possible startup based on the open source

project from the university. The focus of the lecture today arguably is more on my PhD students

than master's students but I believe that you can transfer the situation from research with my PhD

students to what master's students do during their master's program and of course master's

students are welcome in our research projects to spin off together with my PhD students into a startup.

So the first thing to notice is that it's actually not necessarily easy and can fail

easily because there's a goal conflict between what you do as a researcher and what you do for a

startup. In research your desired output is research results and these results are made real to other

people and give credit to your resume by way of research papers and in the case of PhD students,

doctoral students in the form of a dissertation. So you focus as a researcher very hard on getting

those additional insights to such a level of quality that you can write a research paper about

it. Such level of quality is not necessarily a successful or even usable piece of software

which however is what the startup is about. The desired output of a startup is that

viable business model and it certainly involves for a software vendor a software product

and so a software product as the basis or as the expression of the value proposition of

the business model of a startup and the software product really requires you to go

the extra mile that you don't go for code that you write to just get research results. So there's

clearly a goal conflict in how you spend your time. As a consequence I believe it's different

people possibly or at least a shifting focus over time that you need to have if you want to turn

research into a startup. Here as you start a research project large or small initially depending

how basic the research is you will focus 100% on that research to get your research papers out the

door. And then over time and over phases we will see more about this next time. You shift your

focus from research results to the actual startup work meaning the software work the open source

of the work and at the end it's solely product development and no research results any longer

and you should at that point of time also not be at the university any longer but working in a

startup or for a startup. But how do you get there? You need to actually actively manage towards a

goal, towards a set of phases or stages of going from basic research with innovations that had not

happened before yet to a software product that probably incorporates these innovations in a

market ready form. I see or I have seen at university at colleagues who are entrepreneurial and

myself that there are two variants similar to the BDFL model in open source and the peer group

model. The BDFL model maps onto the solo founder model so there is actually usually not the

professor but a research fellow, wissenschaftlicher Mitarbeiter, wissenschaftlicher Mitarbeiterin,

who is entrepreneurial and who in a given research project in the beginning

plays the engineering management role and also often the engineering role as well.

But that research fellow, wissenschaftlicher Mitarbeiterin will not be able to develop

everything themselves. Rather they work with master students. That's at least at German

universities the norm. The numbers play out nicely for one PhD student in a given year there are

easily five master students and for each master student there are maybe three to four bachelor

students. So Germany has a lot of master students that all need to write a final thesis to achieve

their degree. So a research fellow, a PhD student basically, in some research group has a stream of