On starting a business at Fraunhofer - Interview with IVV alumnus Andre Schult

Your career as an entrepreneur started at Fraunhofer? What was your experience like?

As project manager for assistance systems at Fraunhofer IVV in Dresden, by the end I was leading a team of 13 employees. It was a great time, and I was able to implement ideas quickly. I can’t say whether this is specific to our institute, but the huge freedom of thought we had is certainly unique in that, when I got the idea for the assistance system for production environments, neither psychology nor computer science were part of the institute’s expertise profile. Even so, I still received support and encouragement for my idea from my supervisors, Prof. Jens-Peter Majschak and Dr. Marc Mauermann. We started by launching a few research projects in collaboration with psychologists and then, in 2019, began building Maddox, a software based on a licensed Fraunhofer technology. Since then, Maddox has not only been successful with pilot customers but has also won a Transfer Award. Our company, Peerox, continues to maintain an intensive exchange with Fraunhofer IVV and other institutes, and that collaboration is very important.

My former group manager had previously also started his own business, which is when I first began to consider that option. The AHEAD program prepared me mentally for the process of founding a company. It also gave me the chance to meet many fascinating entrepreneurs from other institutes, who I’m still in contact with to this day.

What would you say the opportunities are like for spin-offs at Fraunhofer?

Fraunhofer offers a lot of opportunities for employees. It provides the chance to experiment with state-of-the-art technological developments, in close proximity to the demands of industry. Employees have to pitch, present and sell to industry partners, withstand adverse conditions, deal with minor contracts and learn how to handle finances. In my experience, this makes Fraunhofer a better career choice for entrepreneurs than starting a spin-off from a university, for example. Jobs like that aren’t usually meant until retirement, though; they’re more like a “fast lane.” So my advice is, if you want to develop something cool and you’re looking for funding, you won’t get it from a big industrial company or a medium-sized enterprise — you’ll get it from Fraunhofer!

What could be improved?

People who are pursuing their own ideas have above-average motivation. So Fraunhofer should actively support employees who want to start their own businesses and give managers and employees even more encouragement to become entrepreneurs. With exciting assignments combined with the highest security, Fraunhofer offers the ideal conditions for a job until retirement. If you’re interested in founding a company, you have to go from that security to suddenly thinking very entrepreneurially. Successful spin-offs can be a good return on investment for Fraunhofer, even if not every start-up is a success. At the same time, I’ve seen projects where the research has been completed, but they’ve never gotten any further. Transferring something to market is a big step, and there aren’t always suitable candidates. With Peerox, for example, it took us about three years to get a marketable product.

What can rise the appetite for risk?

You obviously have to earn money at the end of the day. The high potential of a spin-off comes with high risks. A service business model allows an institute to generate turnover in a constant and predictable way. There are limitations, however, such as the advertising ban. Looking back at my own project: As a publicly funded, non-profit organization, Fraunhofer is subject to conditions that would never have allowed us to achieve this growth. Now I can sign contracts that, understandably enough, I’d never have been allowed to sign at Fraunhofer. In our case, these opportunities and freedoms have led to growth that has benefited Fraunhofer, too.

To return to my earlier point, if Fraunhofer could establish itself more strongly as an organization where people learn not only how to work with technology, but also the ins and outs of business, I’m certain that a lot of smart and committed people would see it as very attractive. “Want to start a business? Come to us and we’ll show you how!” Fraunhofer would definitely do well with a pitch like that. I should also say that starting a business isn’t the only way forward. A place at Fraunhofer could also make it easier to get into a big corporation or medium-sized enterprise.

How much personal risk did you take when you started your company?

Security is an issue, of course. To protect my family from taking the hit if I miscalculated with the business, I limited our investments in the business to current savings. We were also fortunate enough to receive funding from the EXIST program run by the Federal Ministry for Economic Affairs and Climate Action. That gave us the opportunity to develop and test a business model alongside research and development at the institute and to prepare the foundation of the business.

Were there any other funders?

At first, we deliberately didn’t take on any investors. It was important for us to work with customers and learn in customer projects. Now, we’re working with two governmental institutions, but we’re not using conventional risk capital. We’ve decided to take a bootstrapping approach, growing from our own funds and revenue. We don’t want to be told what direction to develop the product in. VCs often have the primary goal of growing a company quickly and then selling it at a profit. That’s not how we think about businesses. With Peerox, we want to still have a great team and an excellent product on the market in 20 years’ time. That’s why we’re perhaps taking a slightly slower path, but then we’re lucky enough to be in a field where that’s an option.

We’ve talked a lot about the process of starting a business, but what about the idea you and your IVV colleague Markus Windisch built the business around? How did that happen?

After studying mechanical engineering and working as a research scientist at TU Dresden, I started at Fraunhofer developing process technologies for industrial customers, most recently as manager of a group. In 2015, we were fixing problems with filling yogurt pots in a dairy factory. Our team spent the whole night looking for a solution until an employee called Susi started her shift and taciturnly wiped a particular nozzle with her fingertip. I immediately got to thinking about technical solutions. My existential crisis set in when I realized that this employee’s knowledge couldn’t be automated in a cost-effective way. I spoke to the employee, and her response was that she’d explained the procedure several times. It became clear to me that the problem was about communication, not technology. We can find videos on digital platforms to show us things like how to fix problems with washing machines. There are numerous multi-layered reasons why this type of classic knowledge management hasn’t yet been implemented in production in a sensible way.

However, even with all the data available, you can’t fix a problem without knowing the terminology. Knowledge management systems are generally keyword-based. If you don’t know the name of a particular component, you’ll never manage to find the solution. That’s actually a big problem in production. This is where we come in: The software uses machine data to do the searching on behalf of the human. With the help of machine learning algorithms, the software identifies particular situations and suggests solutions. Based on the responses, the system gradually learns the right suggestions.

How is the idea being received?

If I put all my knowledge into a database, then I won’t be needed any longer — that was one of the most common points that came up in the feedback from the experts at the company. There were others who saw it as a monitoring tool and still others who didn’t want any digitalization projects. It was clear to us that even the best solution would fail if we didn’t bring the employees along with us. Together with psychologists from TU Dresden, we explored how people tick, what drives them, what their anxieties are and when they feel manipulated, and integrated those findings into the product.

What are the measures being taken?

There’s no one trick or one measure; people are too complicated and different. I like to cite the example of a tradesperson in a video explaining how to apply a silicone joint. I can use that to work out a solution to a problem, but I’m still far from being a tradesperson. Here, passing on knowledge works like advertising. Our platform also allows people to prove their own abilities. Skilled workers are often bored by minor tasks and prefer to take on bigger projects. Using the software therefore increases the “value” of an employee.

What’s the next step for your product? What’s your assessment of the continuing automation in industry?

We’ve invested in the product and piloted it successfully with some customers. Now we need to show that we can standardize the rollout process to serve a larger number of customers and allow the product to work in a wide range of settings. Thanks to my IVV background, we have a very strong presence in the packaging industry, particularly in the food and pharmaceuticals sector. In principle, though, we say that our product can be used wherever machine data is available and where it’s important who the employee standing in front of the machine is. There are highly automated processes where the human factor no longer affects the quality. Generally speaking, it won’t be possible to reach those in a cost-effective way. Development is moving forward, and fewer people are required for operations. At the same time, the level of complexity is increasing when it comes to operating the machines. I see very few factory floors that no longer have lights because there are no people working there at all. In most industries, humans will always be needed.