Wolfgang Huber offers lessons learned from bootstrapping Matrix Requirements into leader in medical device documentation software.
Wolfgang Huber on Lessons Learned Bootstrapping Matrix Requirements
On Wed-Oct-9-2024 Wolfgang took part in a Q&A at Lean Culture on “Lessons Learned: Startup Success Without External Funding”
In this Q&A Wolfgang Huber will share the invaluable lessons learned from bootstrapping and scaling a technology company without relying on external funding. He will discuss the importance of understanding industry-specific challenges, the power of a determined team with a shared vision, and the strategies that enabled Matrix Requirements to become a global leader in its niche.
Attendees will be inspired by Wolfgang’s story of perseverance, innovation, and the entrepreneurial spirit that turned a gap in the market into a thriving, self-sustained business. This talk is not just about the success of Matrix Requirements but also about the lessons learned, which will help other entrepreneurs pursue their visions with tenacity and creativity.
Edited Transcript of Q&A
Transcript Edited for Clarity, Hyperlinks added.
Sean Murphy: Today we’re talking Wolfgang Huber of Matrix Requirement on what it took to successfully bootstrap a software startup in the requirements management space for medical devices, a highly regulated space. Wolfgang and his partner, Yves Berquin, started Matrix more than a dozen years ago. They grew from customer revenue from their efforts and the efforts of the people they hired for the first ten years. It’s now recognized as a leader in the field of software requirements management for medical devices.
Wolfgang, welcome. Please share a little bit about your background and what you were doing before you started Matrix.
Wolfgang Huber: First of all, I need to say I’m German. You can hear my accent. Before I started Matrix, I studied computer science in Germany, got a Masters in that, and then spent 20 years working in more or less successful startups. A few were in the entertainment industry, but most were in the medical device field. Then, 12 years ago ,I started Matrix.
Sean: So what led you to found Matrix? What problem or need motivated you?
Wolfgang: I was VP of R&D in a medical device company, managing a team of 10 to 15 people. In addition to managing the team and releasing the product, I was responsible for the technical documentation of the product. With a medical device, there is quite a lot of documentation. You need to create thousands of pages of documents describing the product, how it works, how it has been tested, and what the risks are–all with lots of links in between them. That work must be done by someone who knows the whole product inside and out: it’s really hard to delegate, so it finally ended up on my desk. I spent many days and nights with Excel and Word, writing fairly stupid documents of a product which was already finished. It was a lot of boring work, but you needed to describe everything very exactly.
Sean: Just for clarification, for the audience members that may not be familiar with it. When we talk about documentation for a medical device, these are essentially specification documents that are reviewed by regulatory bodies. These are long, serious, and detailed documents.
Wolfgang: You need to document all of the user requirements, how they match your marketing claims, and how you implemented them. You need to detail all of the risks, how it was tested, and who did the testing. All of these details must be traceable without error. Once you have the documents, you must hand them to the FDA or European notified body. They review it, and if they don’t like it or find a gap or error, they send it back, and you need to fix it. You may need to wait for six months before you can have a second try. You submit the documentation when your product is almost finished, which means if you fail, the whole company needs to wait for six months until you fix what may be a few minor errors.
Moving Beyond a Document-Centric Paradigm For Medical Device Development
Sean: The state of the art when you started Matrix was a document paradigm: you submitted a set of these Word and Excel documents. It was a document-centric approach, right?
Wolfgang: Yes. I mean, there were a couple of products out. IBM offered Doors, but it was basically a replacement for Word and Excel. So, one person would sit down, write documents, and get them signed off. Of course, the problem with this approach is that it usually means that person becomes a bottleneck.
Sean: So you and Yves looked at each other after your last company shut down and said, “That was fun; let’s do it again.”
Wolfgang: Actually, the end of the last company was a bit painful. The investors started to pull the plug, and Yves and I tried to protect the employees. We worked with the investors to ensure everyone was paid–that went on for a year or so, and when it finally shut down, we both said, “We need a break.” We did not want to join another startup and go through the same thing all over again. We wanted to do a bit of programming and realized we had some ideas about addressing the pain of developing this documentation.
Sean: So Matrix started, not exactly as a side project, but as a way to decompress from the painful denouement of your prior startup?
Wolfgang: We agreed it was a good problem to work on, so we each decided to take a month and work on it. I made a UI prototype, and he made a prototype of a back end and a reporting engine to complement it. After one month, we had two prototypes. We looked at them and said, “If we could connect them, we would have a product.” So that’s what we did, making changes as needed, and we had a prototype product.
Sean: You two had collaborated successfully before and knew each other’s strengths and working styles.
Wolfgang: Exactly. We had different interests. He was more like a back-end server database guy. He likes everything which has to do with money, which I hate. In the beginning, my idea was to give away the product for free for everyone, and then some point, maybe make money or not. He was more interested in making money to pay for his family. We had a similar background but different strengths.
Sean: So you decided to incorporate and formalize the business relationship?
Wolfgang: That took a little bit longer. We said after one month, “Let’s make a first prototype together and see if there’s a market.” In three months, we had a prototype we could show to external people to explain what we were doing. We did a bit of research. We figured out that the market was actually very big. We spent the next three months trying to find customers–trying to figure out what to do to find customers who would pay for for our minimal viable product. In our first six months, we had a very ugly “homemade website.” Our first paying customer paid $25 a month, which was not zero! It was not someone we knew before; they found us on Google on the ninth page of some search results. So we knew, because of how hard it was to find us, they were really in pain. They really needed to search to find our solution. So we said, “Okay, we have a customer, it’s working, let’s continue! Let’s see if we can find some more customers.” Three months later, nine months after we had started, we had so many customers that we thought, “Now we need to start a company because we are actually earning a little bit of money.” That’s when we really decided to form a company and continue.
Strangers Become Customers
Sean: So you had taken money from strangers, which is a very strong existence proof of need or demand. You waited until that happened several times and pulled the trigger. Where did your early customers come from? Were they very small medical device companies? Where did you find them?
Wolfgang: Actually, they found us. Our typical customer profile matched our position before we started: a VP of R&D or Head of Development who was in so much pain creating this documentation that they looked for something more modern and would spend half an hour on the web trying to find a solution. But indeed, mostly smaller companies.
Sean: You incorporated in Germany but picked heavyweight form of company for a two-person firm.
Wolfgang: Yes, we chose to incorporate as a GmbH, which is quite painful and usually takes a lot of time. But we did it because we wanted this image of a serious German company, committed to quality and perhaps a bit slow moving but not a company which lives for a few months and then dies. So, it was a bit of marketing to get that German company started.
Sean: Then, due to the nature of your product, you chose to be a regulated company in the EU?
Wolfgang: Exactly. We decided, mainly for marketing, that our company should be a regulated company, just like a medical device company. We would follow the same standards as our clients. We did the same documentation for our products that the customers need to do for their products, and that was, in one sense, marketing. But it also helped our customers because we needed to prove that our product worked. If we were certified, it would be easier for them to believe the product was good.
Matrix Today
Sean: Where is Matrix today?
Wolfgang: We grew a company in the first ten years to about 15 people, and still without any external funding, marketing, or sales, just basically field experts, programmers, and specialists in the medical device area. After ten years, about three years ago, we decided to seek external investors. The main reason was that it was getting more and more difficult for us to hire the people we wanted to grow. You need to hire more than just your friends and people you know from your network. You need to hire strangers, which means you need HR people. So we were, I mean, it sounds a bit stupid, but we were earning too much money, and then suddenly you need some HR people, or the hiring process might take six months or more, and you lose momentum.
Sean: With just fifteen people, you can fit the entire company around one table–or one Zoom screen. You can remain flat and relatively informal, but at some point, as you add people, there’s a phase change where that doesn’t work anymore.
Wolfgang: We didn’t have a secretary, so we each organized our own travel. We didn’t have marketing or sales. If we wanted a new website, we looked at it as a programming problem. We decided to change gears and make it a real company with a real headquarters, where some people might work together in one room. But neither Yves nor I wanted to be CEO of that company. We did not want to manage sales, marketing, or finance. We didn’t want to hire people; we wanted to continue to focus on our product and our customers.
So when an investor called, we picked up the phone for the first time. They had been calling every other week since we hit a million Euro ARR. We were very lucky that one of the first people to call us had the same vision as we had: to make a product that helps medical device companies and to be in exactly that market and nothing else. So they understood what we were doing, and we really liked them. They liked us, so we actually got them on board. With their help, we hired a CEO, and afterwards, the company changed quite dramatically. So we went from 15 to like 45 people in one and a half years. Wow.
Engineering Still Fully Remote
Sean: Are you still fully remote, what with your headquarters building?
Wolfgang: Yes, Matrix is still fully remote, even though we have a lot of people sitting in Paris because the investor and the CEO came from Paris, so that kind of became a natural headquarters where new people like marketing and sales are located. But all the engineering and all the technical people are still everywhere in Europe and the US.
Sean: Cool! When you look back, what key skills or expertise you brought to Matrix had an impact on your success?
Wolfgang: One critical thing was that I had worked in many successful startups, although some failed very quickly and some were sold off. I had worked as a software developer in these companies but was also interested in everything else. I learned about finance, marketing, and sales by listening to what those people had to say, watching what they did, and paying attention to what they predicted. I went to see customers with them and was interested in everything good and bad that can happen in those conversations. So, I saw good salespeople, bad salespeople, and good marketing managers. I saw big business plans for some VC funded companies and more realistic plans that were achievable in others. With Matrix, I tried to build on everything which I thought had gone well in the prior 20 years
Sean: With an engineering-heavy team, they must be generalists who provide customer support, help with pre-sales, do QA, and play many other roles. They are doing more than just writing code.
Wolfgang: We knew it was important to hire people that we knew who were talented and would be happy to talk to customers. We hired people like Yves and myself, who were fascinated by medical devices and excited when they could talk to customers. We didn’t want to have people who just wanted to sit at a desk, take orders, and deliver so-so results.
Sean: By keeping it small and flat, anyone who needs to be in a conversation can take part.
Wolfgang: We kept the company super open and transparent, and that’s still true today. Everything was communicated through Slack, and everybody could see and listen to everything. Everyone heard the good news when we won a customer. When we lost one, they heard about that too, and our explanations. Everyone heard about new software releases; it was an open, ongoing conversation.
Key Accomplishments: Team and Culture
Sean: Interesting! Looking back, what are two or three things you accomplished that you take the most pride in or are a source of significant satisfaction?
Wolfgang: The first is the team we recruited and retained. I am super proud of how well we worked together. We never forced anyone to work specific hours. Everybody had their own tasks and objectives. If someone was, for example, a parent with a small kid, the small kid was sick, they may not have worked the whole day. They may have checked email to see if something important had happened, but they didn’t work. But perhaps, at 9 pm, they might have a conference call with someone in the US after their child was asleep. It was up to them to decide how to do the best possible job rather than just working eight hours a day. Our whole climate of working together as a team made me very happy.
Sean: It sounds like everyone wanted to be part of the team and jointly set their work standards. It’s great that everyone wanted to contribute and understood that things happen. I think it’s a great culture that lets people choose how best to contribute and allows them to put their family first.
Wolfgang: Everyone knew that if they did not produce results the company would go belly up. We were a young startup without outside financing so we had to deliver.
Sean: Matrix turned out to be the first of a new generation of web-based tools for medical device development. You must see that you have positively affected medical device development and certification: as you look back over the last thirteen years, how do you assess the impact on your customer’s design and time-to-market challenges?
Wolfgang: When we started, we were the first tool in our space to be completely web-based. So, in the past, everything was super secret, and firms relied on either in-house solutions on their own servers or Word and Excel.
We envisioned making the documentation a shared responsibility of the whole engineering team instead of one person writing all of it. Nowadays, we enable everyone to contribute important information about their work.
For example, a software developer will update the database, explaining how he implemented a particular feature. Then, someone writing tests will relate to that implementation and explain how it can be tested. But it’s not one person writing the document. It’s now a whole team that participates.
Our experience before Matrix was working in medical device startups and small companies. A project manager or the head of R&D wrote these documents because they were the only people who really knew everything. It was a complete waste of time.
Web Access Enabled Wider Collaboration in Customers’ Engineering Methdology
Sean: I assume that, in addition to allowing the entire team to get involved, another impact of your web-based approach was to allow firms to bring in remote people to collaborate in a way that would not have been possible in a document-centric paradigm.
Wolfgang: Exactly. In the past, someone wrote, for example, a requirement management document for the product, and then a project manager started to call people and tell them what to do. Each engineer only got a little piece of the job; they never saw the whole document–and, therefore, the bigger picture. But with Matrix, they can now see the whole picture. You can see why you need to implement a particular feature. While it might not be useful for the product, it might be a risk control. So, you can now see the risks and where they fit into the product pipeline and development.
Sean: So in the same way that the Matrix culture was very open internally, you allowed your customers to gain a holistic perspective. Some questions have come in, let me stir those into our conversation. When you say you hired people you knew, you were not just hiring in Belgium where you were based or in France where Yves was.I’ve been neglecting my role as a as a moderator here, we’ve got a number of questions I’m going to stir in. So the first was, when you say people were known to you, you weren’t just hiring in Belgium where you were based or France where Yves was.
Wolfgang: We initially hired people who were mainly in Europe, but later, from all over the world if we had worked with them before.
Sean: That would not scale to an arbitrary size, but you did scale it to a team of 15.
Hiring Followed Revenue
Wolfgang: Whenever we had enough money coming in to hire someone, we hired someone. We never projected hiring, saying, “Okay, and we need to hire five people beginning next year. Let’s start searching for them.” Instead, it was, “Now we have enough money to pay someone. Let’s get someone as quickly as possible.” We didn’t have a buffer to hire people before we could pay them out of current revenue. We waited until we had the cash flow and hired someone as quickly as possible.
Sean Murphy: It sounds like you were strict about hiring following revenue and had an implicit “no layoff” policy.
Wolfgang: We didn’t have a choice, we did not have investors so we did not have a big bank account. We planned month to month and always worked hard to stay above zero.
Sean Murphy: Another question: where did you publish to find your first customers?
Wolfgang: We wrote a very simple website without external help or designers, explaining how we solved the documentation pain point we thought our customers were feeling. It was less about the benefits of our product: we explained to an engineer how we solved this problem, and then we waited for Google to do its Magic.
We didn’t follow the marketing rules you learn in school where you describe the benefits and the ROI. We felt we had been in the engineer’s situation many times and asked ourselves what they would want to see on a website. When I was an engineer, I did not like to talk to salespeople and get a lot of “blah blah blah” before I even knew what the product could do and actually tried it. So we let people try it as quickly as possible by putting it online and letting engineers play with it. We knew they would quickly understand how Matrix worked and determine if it would solve their problem.
We knew we would lose potential customers because they would not understand our solution or would have a different idea about what the product should do. However, we felt it would not matter as long as the market was big enough; we preferred to support people who understood our concept and thought the same way we did because it would create fewer problems after the sale.
Started with Customer Interviews In Parallel to MVP Development
Sean Murphy: Question from the audience: how did you know that customers are really interested if you didn’t have a product to sell them?
Wolfgang: We had a minimum viable product, which did not do much, but we started selling it after six months. In the first six months, we visited many companies that were potential clients. We didn’t try to sell; we listened to the pain they had doing documentation. For example, we called Siemens and asked for an appointment with an engineer doing requirements management. We told them, “Don’t worry, we don’t have a product to sell you. We just want to learn what the problems are so that we can build a product that might help you.”
And many were open to those conversations because they were unhappy with their current approach and saw room for improvement. In the case of Siemens, they had an in-house solution and were still unhappy. Other companies had licensed Doors from IBM for several million and were unhappy with how long it took to implement it. Some used Word or Excel; others didn’t understand how to write the documentation. Many were happy to talk to us, and we listened and asked questions to see if there was really pain and a need.
Sean Murphy: How did you and Yves craft the MVP? Did you have some rules of thumb?
Wolfgang: In the beginning, I wrote a user manual. I sat down and asked myself, “What does the customer need to do with our product? How do they find the data? How do they enter it? What else might they look for in the user manual? I wrote the first description of the product in the style of a user manual.
Sean Murphy: Another question from the audience. How did you get the regulatory bodies to take you seriously when you’re not associated with a company that’s doing a medical device? How did you get approved?
Regulatory Certification Increased Early Customer Confidence in Product
Wolfgang: That was quite difficult. We needed to get approved in Europe by what is called a notified body. We had two problems.
First, we didn’t do a medical device ourselves, so we had to find a reason why we should certify us. Most notified bodies we approached thought it was not possible. But we knew the regulations and knew that was not true.
Second, we did not have an office: any regulatory body, whether it’s the FDA in the US or a notified body in Europe, they want to visit you. They want to come on-site–they even make surprise visits. But when you don’t have an office, it gets a bit difficult. So we asked pretty much every respected notified body in Europe if we could get certified. Most told us they were not interested, and some said to come back in two years because of the long pipeline for regulatory changes; they were not taking new companies.
Then we got lucky. We found a new notified body in Eastern Europe–I think it was in Slovakia–that was just starting up. We were probably one of the first people to call them and ask if they could certify us and if we could do it remotely.
They said yes, and we were probably one of the first startups to pass their certification, but it’s an EU certification. They had to follow the same rules as any other notifying body, so we had the same certification as if we had gone to the TUV.
Sean Murphy: Another question from the audience: did you calculate a runway when you started? Did you set a budget for how much time you would give it?
Wolfgang: We never made a business plan. We did calculate that it was a big market, and there was no other solution like ours, so if we don’t screw up, it should work out.
We were always very opportunistic; we did not plan too far ahead. We did not work on a plan to convert 30% of the existing market in two years. We looked for our next customer: how can we have success with them in two or three months?
How do we find them? Do we need to call 100 startups? Do we need to go to a conference? Do we need to send emails? What do we need to do next to move forward? We worked in a very short time frame of two to three months.
Sean Murphy: You knew this was a new paradigm because of your prior experience developing software and working in medical device companies. After your problem discussions with customers in the first six months, you had high confidence that this new approach would work. And it was a large market. How did you decide what to do next?
Always Start with “How Can We Measure Success?”
Wolfgang: We always started with, “How can we measure success?” If we would not be able to measure success, we would not do it.
For example, in the beginning, we participated in some accelerators and things like that. Their people would tell us, “You need to do a website with AB testing. So you have two different websites, and you figure out which one works better, and then you focus your effort on the one that works better.” But we knew that would not work when we only had ten visitors a month; we would not be able to decide based on statistics from so little data.
We did not follow the standard book of doing startups.
Another example is that if you talk to any salesperson in any serious company, they talk about sales funnels. So we estimated that out of 100 people in an email outreach campaign, 30% will respond, and 20 will pick up the phone if you call them. And in three months, we will get ten customers.
After three months, you figure out maybe, if you’re lucky, you have the number of predicted customers, but many did not come out of the funnel; they came out of nowhere. But in those same three months, you figure out ways to improve your approach, and at the same time, you are dealing with such small numbers that you never have a statistical basis for calculating your funnel.
Then you realize that what is really valuable is knowing all of your customers and understanding your prospects:
- People who call you.
- People who find you on the web and fill out a software download form.
- People you meet at a medical device conference.
You need to talk to them and understand how they think about their problems. You need to stay close to them: everyone is now a person, not a statistic or a number in a model. It’s not about automating something with a lot of salespeople.
While all this outreach is going on, you have to keep your customers happy as well because if you have one happy customer and he talks to a prospect, now you have two customers. So it’s much easier to have a few happy customers who spread the word than trying to convince people who don’t even feel the pain or see the need for your product.
Competition: Initially Word, Later Very Well Funded
Sean Murphy: All this time, you’ve been essentially replacing Word and Excel. Microsoft isn’t messaging against you, saying, “Excel is just fine for medical device documents.” You were competing against inertia, the way it was always done with Word and Excel. Was there a point when you started attracting competition following your paradigm?
Wolfgang: There’s always been a little bit of competition. Still, at one point, an American startup raised tens of millions of dollars to do pretty much the same thing, and they started doing a lot of marketing. Luckily for us, they didn’t really focus on their product. Instead, they focused a lot on marketing and explaining to everybody why they needed a web-based tool to solve the medical device documentation problem.
Sean Murphy: So they helped to educate the market about the need for your product?
Wolfgang: We were quite happy after we realized that when a company was considering our competitor, we needed just enough market presence for them to find us and compare the two solutions. Their product wasn’t very good, so they chose to put it behind several layers of salespeople: it would take months to get a quote.
While on our website, you can see the price and start using our product within two minutes without a credit card. We took completely opposite approaches.
At first, it was a little scary to see someone with millions of dollars in marketing pushing into the market. But as we got to know them better, we became less afraid and learned how to use their messaging to our advantage. That was a surprise. Another surprise for me personally was how easy that whole thing was in the end. I mean, if you really know the pain of your customer, and you solve that pain, the rest kind of automatically follows.
First Decade: Everyone Has Regular Contact with Customers and Prospects
Sean Murphy: So in the first ten years, everyone on your team is still talking to customers or prospects at least once a week?
Wolfgang: Yes, everybody on the team would talk to one or two customers at a time. Sometimes, you would talk to a customer a lot for a month and then not hear from them for eleven months, but we always stayed very close to them, so we always had excellent support. For example, if someone called us to ask a question, we would return within minutes. An engineer would stop development and talk to them or find a solution as quickly as possible because building a list of tickets wouldn’t help our customers.
Sean Murphy: That approach is the opposite of the advice that some people give. Paul Graham has this concept of the maker schedule where you minimize interruptions and don’t treat customer inquiries as urgent phone calls. Instead, you get back to them when you’re at an appropriate stopping point. But you and your team treated customer inquiries as urgent interrupts.
Wolfgang: My point of view is that customers are the most important thing because, in the end, they pay your bills, and if they stop paying you, you don’t need the next generation of your product. If your software has a problem and you solve it quickly, then they will be happy. So they might recommend your product to someone else. I don’t see any downside in helping your customers as quickly and efficiently as you can.
Knowing What You Know Now…
Sean: A question from the audience: knowing what you now now what advice would you give yourself in your first startup?
Wolfgang: Good question! We were a bit too introverted in the first startup I founded with another colleague, so we worked on our product. We had our ideas and built a marketing campaign, website, and whatever, but we didn’t spend enough time talking to customers. If someone had told me to forget about the website or a finalized product and first talk to all your customers, that would have been very useful advice. Another piece of advice is to train your engineers about what the company wants to achieve. It is not your product but the customer pain you want to solve so they can talk to customers more effectively.
Sean Murphy: This has been delightful. Do you have any final remarks or suggestions you want to make for folks that are considering bootstrapping?
Wolfgang: Listen to as many potential customers as possible to get as much advice for your company as you can. You can get good advice from more experienced entrepreneurs, but follow any of it blindly. Most of the advice you will get on how to run your company will not fit your company. So you need to make your own decisions and be creative: it’s up to you if it works or not. If you just follow the standard way of doing things, you won’t succeed. If you do exactly the same thing as your competitors, you won’t succeed. I hope that’s helpful.
Sean Murphy: If people listen, I think you have given them a very good roadmap here.Thank you!
SKMurphy Take
I was broadly familiar with the Matrix Requirements story before the presentation but was impressed by several principles they followed in bootstrapping a successful company:
- Even thought they had worked in medical device firms they invested six months in customer interviews to make sure they were solving a real pain point for customers.
- Stay fully remote to attract talent throughout Europe and then worldwide. Talented people can self-organize on a team when leadership is transparent about strategy, decisions, and challenges.
- Build a high-performance team supported by a humane culture.
- Initially hired people they knew or who were recommended. Simplified interview process and encouraged potential employees to take a risk on a startup. This is also a testament to how Wolfgang and Yves had treated the people they had worked with previously: old teammates were willing to follow them into a risky situation.
- Had deep knowledge of customer needs because they each had worked for more than a decade in the types of firms they wanted to serve.
- They understood their customers were risk averse when it came mission critical infrastructure and chose to submit to regulatory oversight and selected a heavyweight corporate form to signal their intent to provide a high quality product.
- The engineering-heavy team prioritized customer support, minimizing churn and maximizing referrals.
- They relied on cash flow budgeting: hiring followed revenue.
- They leveraged product functionality to allow customers to self-select: invested in support instead of sales and marketing.
- Made some key architectural decisions early that enabled them to pioneer a new paradigm for medical device specification and quality management.