While RapidMiner initially followed the relatively typical service-based business model of an IT service provider, today the company relies on a licensing model with more than 100 partners worldwide. “In addition to being innovative, IT organizations need to build large ecosystems and communities of developers to survive. Companies that believe they can do it all alone will perish.”
Klinkenberg and his business partner recognized this early on. They measure their success mainly based on downloads and subscription rates, so they can easily understand what works and what doesn’t. “In the beginning, RapidMiner was a tool for experts. The first version had no design interface and was difficult for new users to utilize. When we added the graphical environment, user data rose dramatically. Similar to when we added the installer.”
Today, the tool offers online tutorials, application templates, and transparent example processes with sample data records and an auto modeler function. The most important thing for Klinkenberg: “Facilitate operation and increase automation while maintaining transparency and flexibility. Add to that our large community and marketplace, so other companies and universities can build their own projects based on our software and share RapidMiner expansions with the community.”
RapidMiner now works across all industries: Automotive, aviation, chemical, metalworking, food, insurance, banking, internet. Above all, the topic of ‚predictive‘ is now established and growing rapidly. Nevertheless: Many organizations are still in the testing phase. Even industries and companies that are already engaged in predictive analysis still have much untapped potential, says Klinkenberg. “Although many machines are already connected in industry, there are still many unused interfaces along the production chain. And the chain itself is often forgotten.”
According to Klinkenberg’s assessment, the greatest effects can be achieved in production and industrial manufacturing through the comprehensive use of data: Production becomes more plannable and efficient, products are better and more individual, and environmental resources are conserved. Klinkenberg also sees strong effects in the healthcare and medical industries. Proper use of data will not only allow for more personalized treatments, it will also help prevent medical problems such as heart attacks and strokes – and with the appropriate treatment recommendations.
To work efficiently with data, it is particularly important to break up data silos and to connect the data across all processes. Klinkenberg advises: “Just start somewhere. Often you can achieve a lot with minimal effort. But most of all, you acquire a sense of your own data quality along the way. External consulting can help, but it is becoming increasingly important for companies to build their own data science competence. This creates a real competitive advantage.”
Since 2013, the company has also maintained a location in Boston and thus has a direct comparison of developments between the US and Germany. “The US is a bit ahead of Germany and Western Europe. Americans are often quick to opt for an innovative solution but throw it overboard more quickly if it does not work right away. In Germany, the start-up time is a bit longer, but development is more systematic and sustainable.”
Politicians must also become more aware of data science, demands Klinkenberg. “It is important to me to establish the topic in Germany more strongly in politics and to reduce fears in society. Germany is a very strong industrial nation and on a good trajectory. But the wheel is turning faster and faster, and local companies are now competing with global corporations. There are huge potentials and they have to be recognized. Certainly not everyone has to become a data scientist, but everyone has to be aware of it. This should already be taught in school.”