Developing successful data products in a structured and strategic way

KI Service - Industrie 4.0

By Nils Klute, Specialist IT Editor and IoT Project Manager at eco – Association of the Internet Industry

Service-Meister relies on open platforms and an AI-based ecosystem. Now the development of the first data products has started in the workshop. In a two-part series, we accompany the experts. Read in part 2: What distinguishes successful data-driven services.

Search for relevant information online and get matching hits within seconds. Or have books recommended to you online in line with your own preferences, instead of having to rummage through the entire range: “Examples such as Google and Amazon show what capital is invested in data products,” said Dr. Florian Wilhelm of inovex. Together with Dr. Alexander Löser from Beuth University of Applied Sciences and Prof. Steffen Staab from the University of Stuttgart, the data scientist hosted an online workshop in which the speedboat projects from Service-Meister took part. The goal: To not only find ideas and determine directions to develop own data-based services for the service in Industry 4.0, but also learn the methodology. Because: “Successful data-driven services require a structured and strategic approach,” said Wilhelm.

First examine the data landscape, then develop a data strategy

Which data is available in the company and which is missing? Where does data originate, where is it stored and in which formats is it available? And which data sources can perhaps be aggregated to fill information gaps? In order to develop data products at all, companies must examine their own data landscape. Which makes the task challenging: “Data is not evaluated or is only processed in silos, which means that a cross-company context is missing,” said Jörg Bienert, chairman of the KI Bundesverband and partner in the data science consultancy Alexander Thamm, in an interview on servicemeister.org. “In order to use data in the company efficiently and holistically, it is necessary to create a comprehensive, company-wide data strategy.”

The Tesla example shows why this is necessary: The manufacturer first laid the foundation for its business model with its digital services. Where smart vehicles initially lacked the data needed to drive autonomously, intelligent cars simply collected this information themselves. With Fleet Learning, Tesla processes data from individual cars to gain insights for all cars. This creates a self-learning, self-optimizing system. Even if users have to brake or take countermeasures in order to correct an artificially intelligent (AI) decision, the control software learns from that. “Feedback loops like this one continuously optimize our own data product,” said Wilhelm.

Applying data products to the user context

To continuously optimize digital services and capitalize on smart capitalization, data-driven services must be at the right point in the value chain. Take the example of buying sneakers: If you want to jog, you choose a shoe, draw up a training plan, determine routes and daily. “The decisive factor is the context in which users use the product,” said Wilhelm. “A shoe manufacturer could develop an app that suggests routes, monitors your own run and provides training plans.” The digital service supports beginners and experts alike. And all users improve the overall system with their data.

“These are also interesting concepts for Service-Meister, but often the necessary database is lacking, especially in the B2B area,” said Wilhelm. Compared to B2C markets, the B2B industry not only has fewer customers for a specific solution, but also many and varied requirements to consider. And where B2C providers can pass on development costs for data products to all of their many customers, the situation with just a few B2B users makes this financially unattractive. In order to solve this dilemma, Service-Meister plans on offering general, preconfigured AI building blocks that can be used to implement data-driven services in industrial services in a particularly cost-effective manner.

Data products in SMEs: Money flows, data does not yet

What problem does my data product solve? Who is the target group? And how does the development pay for itself? In Service-Meister workshop, Wilhelm relied on a Predictive Analytics Canvas. “Start-ups also plan and work with these templates in order to develop themselves and their business model in an agile manner,” said Wilhelm. “The canvas provides orientation, covers tasks such as a checklist and ensures that all steps on the way to your own service are followed.” Structured, the template guides users from the idea to the Minimum Viable Product.

“Although money does flow from the customer to the provider in small and medium-sized businesses, data is rarely involved,” said Wilhelm. “For digital, data-based products and services to flourish in this country in the future, it is not only industrial services that need to rethink.” According to a study by the Business Application Research Center (BARC), only 17 percent of companies are exploiting the financial potential of their data. A quarter of the study participants even completely rule out that their company will earn money with data in the future. Still: Twenty percent think it is conceivable. And four out of ten companies are running corresponding pilot projects, according to a 2019 study by BARC.

In the first blog post about the workshop, you can read why platforms need success stories and where the potential for Europe and Germany lies. “Whether mechanical engineering, logistics or chemicals – industries like these are still accessible as platform markets for local companies. Companies must now make the right decisions and make long-term investments,” said Dr. Alexander Löser from Beuth University.

You liked this article? Then subscribe to our newsletter and receive regular updates on similar topics and the Project Service-Meister.