Between process mining and data science
Wil van der Aalst is professor for process and data science at RWTH Aachen University and part in many successful start-up foundations.
Collective Incubator: Prof. van der Aalst, you hold the chair of Process & Data Science at RWTH Aachen University and you are considered the founding father of process mining. First of all, what is process mining all about?
Prof. Dr. Wil van der Aalst: Process mining is a technology that is linking data science to process science. I think most people are familiar with data science. They immediately think about things like AI, machine learning and doing things with data. Process mining bridges this working with data with more traditional process-centric approaches. The easiest way to explain what process mining does is to look at the typical process mining project. You start with extracting event data from any system. Ranging from SAP to a hospital information system, any information that is sort of time-based can be extracted and used for process mining. Thus, you can automatically learn the processes as they are actually executed. This often leads to very surprising insights such as that the real processes differ greatly from the expectations of the management. Therefore, the gained knowledge presents an incredible opportunity for improvement. Using process mining, you can automatically detect bottlenecks and their causes. Process mining can be used to immediately see compliance problems ranging from fraud to inefficiencies or for example customer-unfriendly behavior as well as their root causes. You can even predict such problems using process mining. Moreover, the current trend in the market leads to process mining tools starting to automatically take actions in case something does not run as expected.
Collective Incubator: Before we come to the benefits for companies that you already described here, you with your chair and your various other scientific engagements, stand for the research side of this topic. You are one of the most cited computer scientists regarding process mining. What are the current topics in the foreground of your research?
Prof. Dr. Wil van der Aalst: What is very funny is that based on the research that I started to do in the late 90s, many companies were founded. There is always a very direct connection – much more direct than in most other fields – between the research that we do and what ends up in commercial products. At the same time, many of the research questions that we work on, are research questions that we identified 20 years ago and that are still not completely solved. For example, the problem of process discovery – learning good process models from raw event data – is still a scientifically incredibly challenging topic. I think it will take years to get it right. That doesn't mean that the things currently are not usable, but there are still these core process mining challenges, which are open problems where we book improvements each year. Nevertheless, new things are popping up. For example, we are now trying to learn process models that are much closer to how organizations and people actually work compared to the notations that we had before.
Collective Incubator: Compared to when you started your research in the 90s, did some other universities join the research competition and get on track? How fast can the scientific world adapt to such new topics?
Prof. Dr. Wil van der Aalst: For a very long time my group and I were basically the only group in the world working on this topic. Before, I had been very successful in the area of process modeling and process analysis. I also wrote the first textbook on workflow management systems. The work that I did then was very famous because of all of the things that we did in the area of workflow management. Hence, many people did not understand why I was making this transition. At that point in time, there was not so much data and people were very happy with the more model-driven approach. However, I felt many of these models that we were creating had very little to do with reality. So I wanted to address this issue and that's why I jumped onto the topic of process mining. It did not exist and for a very long time, it was just us. What you could see is that after six or seven years of research, the first of my students started to create companies. Some of them were a bit too early and not super successful. Others were incredibly successful. Celonis is a very nice example of a company that has been building on the research to create probably the most exciting IT company in Germany at this point in time.
Collective Incubator: Celonis, a startup and now the leading company in the field, is currently valued at several billion dollars. Its customers are corporates like BMW, Lufthansa and most of the German hidden champions. What do the companies gain from process mining and especially from companies like Celonis?
Prof. Dr. Wil van der Aalst: Celonis is clearly a market leader and is currently valued at two and a half billion euros. Important to see is that this is not a single tool, this is really a market. In total, there are over 35 pure process mining companies and additionally, process mining capabilities are increasingly integrated into other tools. The gain for companies using these tools is, that by using process mining, they can see where the frictions in an organization are. They can see where the bottlenecks or compliance problems are. Often, an organization has a lot of low-hanging fruits. Just by looking at these inefficiencies and making simple actions, it is easily possible to save literally millions and millions of euros per year. For example, Siemens is one of the bigger users of Celonis. Within Siemens, six thousand people are using the Celonis software. Just an interesting statistic: If you look at the order to cash process within Siemens, which is a standard process that any organization has, then within one year the different ways of executing the order to cash process were found to have nine hundred thousand different variants. If you think about this number, nine hundred thousand different ways of getting money from their customers after product delivery, then you immediately realize that there is a lot to be gained. The field is moving from visualizing this and showing the diagnostics to also automatically take action. For example, if you have a supplier that is continuously changing the price or is doing something that is very bad for your organization, then you can automatically take actions to address these things.
Collective Incubator: In your opinion, Celonis and its competitors, do they capture the whole market already, or would you advise computer science students to still pursue ideas in the field of process mining?
Prof. Dr. Wil van der Aalst: I think that this combination of processes and data is completely logical if you think about it. That is a huge base. Moreover, if you look at the current market penetration at a global scale, the market is just starting. In Germany, of course, the number of organizations that are using process mining is already quite high, but the potential of this market is a hundred or thousandfold of what it currently is. It is also clear that that will happen, but what you will see is that there will be diversification in the market. If people want to start in this market, they need to think of an edge, which is not exactly the same as what the current systems are doing. For example, traditionally there are all kinds of suppliers of business modeling tools for documentation and these types of things. What you see is that these organizations are rapidly including process mining capabilities. I think you will see that in all kinds of other fields, this will be an integral part of the tool.
Collective Incubator: Now, the founders of Celonis moved to Munich at a very early stage. The reason being, according to their own statement, the access to talents and customers. Do you think such a Start-up could also have been established in the vicinity of RWTH Aachen or TU Eindhoven? Or does it have to be Munich?
Prof. Dr. Wil van der Aalst: Of course. Many of these Start-ups also exist at this stage. For example, if you look at Eindhoven, several companies were founded there. Perhaps one of the more interesting companies is a company called ProcessGold, which is now part of UiPath. UiPath is still a couple of factors bigger than Celonis. I think these Start-ups can happen in any region. I think the access to talent, so the access to students that know about these things is very important. Of course, there are different regions where this is easier than in other regions. When I moved to Aachen, I realized even more what I knew before: What is a pity here in Aachen, is that we have a super good university, but there is not that much industry. If you compare it to Eindhoven, I would say that around Eindhoven, there is much more industry.
Collective Incubator: There is no Philips in Aachen.
Prof. Dr. Wil van der Aalst: Many people know Eindhoven from Philips, but the company that is much more influential is, for example, a company like ASML. You have things in your pocket that contain chips made using their systems. They have about 80 percent of the global market of machines to produce chips.
Johannes Schäfer: So what would you suggest for Aachen as a city region, for RWTH Aachen as a university. When we look at ATEC, we will have almost 2500 founders, investors, scientists and students coming from all over Europe together. What do these players - like the university, the city of Aachen - what do they have to do in order to establish a network and an environment that produces more Start-ups and has stronger bonds to the companies?
Prof. Dr. Wil van der Aalst: Yes, I think in general, because there is not that much (physical) industry here around Aachen, it makes sense to try to stimulate more IT-related research. Because for that, you don't need to have the factories, etc. It makes perfect sense and I think you can also see that by the more recent Start-ups. Many of them are in this IT-related space. So that's a good strategy. The question is how to improve it. As I said, there are many students here in Aachen, that have a very strong background. So that's incredible potential. Perhaps, the city of Aachen could become a bit more international, to not just attract German students, but also to create a place where people would like to be. I've seen that in Eindhoven. Around 30 years ago, Eindhoven was a city where people would not like to be found dead. What you see today is that it kind of transforms to a very international city. Not a city for tourists, but it's a city for experts etc. The close vicinity of Aachen to Belgium, the Netherlands could make it a very international city within Germany. I think that there is some potential that is not being used. Also, given the fact that the city of Aachen is a beautiful city. Nevertheless, it could be more vibrant, more international than it is today, in regard to transport, etc. I think also the rector of the RWTH has made this comment coming here, that there is more potential when you look at these types of things.
Collective Incubator: Do you think this is a general problem of German cities being part of Germany, which has a very common internal market? Whereas the new markets are not internal, but thanks to digitization, international, at least European markets. Is this an advantage of cities from the Netherlands that have always been more international and related to the world?
Prof. Dr. Wil van der Aalst: Yes, I think all things are relative. If you compare Germany to France or Italy, you could argue that Germany is very international, people speak English, etc. At the same time, it is relative and I think if you look at the things related to it, data science, artificial intelligence, etc. this is a completely global market where you should try to attract talents from all over the world to work in these areas. I think Aachen has a potential also being close to the border that it does not realize at this stage. In Germany, for example, the car industry is very important. Many of the students that studied here in Aachen, would traditionally work in the car industry or suppliers to the car industry. And I think one has to realize that this will change, that other industries will become more important and that there is a need for thinking about new products and services.
Collective Incubator: Do you think Germany - German cities and industries, like the car manufacturers and others - are ready for digitization and for digital business models like working with Celonis?
Prof. Dr. Wil van der Aalst: I think that they want to and to some degree, they are also ready. Interesting is, if you think about process management in general, then there is a very strong European tradition in doing this. And this tradition is particularly strong in Germany. The fact that SAP is a German company seems to be no coincidence. It has to do with the fact that here in Germany we like to study and manage these processes in a scientific way. I think many companies are ready and they are also prepared to do this. That probably explains why the penetration of process mining software in the German market is among the highest in the world.
Collective Incubator: Ok, last question on a very personal note. In your environment and surrounding, many Start-ups have been founded. Did you ever ask yourself whether it's a good opportunity to be part of one of the Start-ups and pause your scientific career for a while and become a founder?
Prof. Dr. Wil van der Aalst: I started to work on open-source software at a very early stage. So what we did in Eindhoven with the ProM tool in the field of process mining, but also with systems like YAWL in the workflow management domain. I've always been an advocate of open-source and I also have this idealistic vision that if the government is paying you to do research, you should, in principle be neutral and want to share it with many others. My model has always been to do things in the open-source domain. If I have ideas, I try to share them with as many as possible. I'm involved with many organizations from whom I don't get money and I consider it to be a success if these companies take over these things. So that has been a foundational choice. The problem would be if you get involved in one of them, there is no step back anymore. Following, you have given up in a way, this independence, and you can no longer claim to be neutral. In my view, there are many things that you can no longer do. Other professors will think about this differently, but that's the way I feel.
Collective Incubator Dear Professor Van der Aalst, thank you very much for the insights and the interview.
Johannes Schäfer, Kim Nebe
Photo: Bart van Overbeeke Photography