Data Science Marcel Hellman Header

Digital services

Our Thermomix is smart: How Data Science makes it even smarter

Data Science isn’t something new. But with digitalization, IoT appliances like Thermomix ® and the huge amount of data being required, the field has gained new dimension and relevance. I can experience this change at Vorwerk Digital, where we recently working on making our appliances and services even smarter. What my role as a Data Scientist looks like, which challenges we are faced with and how to solve them – here are some Insights and three practical tips to really understand data.

Big Data, Smart Data, Data Mining, Data Engineering, Predictive Analytics – when it comes to data, the list of buzzwords can be long and confusing. So, it makes sense to find someone who brings light into the darkness of the data jungle. Therefore, the field of data science is a much discussed one. And still: a lot of people imagine “Brainiacs” working in the dark depths of a corporate lab brooding over heaps of statistics to make processes more efficient. From my everyday life I can contradict: Data Science is much more than that, because new knowledge about the customers’ needs can motivate a whole team to experiment with new features and enable new looks outside the box.

Introducing smartness must start during product design

In doing so, we not only gain insights for marketing purposes, but also improve products and make them even smarter. When it comes to Data Science at Thermomix ®, we e.g. try to identify patterns in the user behaviour while cooking, that may later have a direct impact on the product design. Think about the thinning of onions: in our data we see, that customers use various specific functions of our kitchen machine – like cutting and steaming – to make them succeed. So, one learning we could draw for the future: We combine certain steps into one and give our users another selectable function that saves them some interaction with the device. This is just a simple example of how Data Science directly impacts the customer experience. Of course, some conditions must be met and obviously understanding the data is key: “What data sources are involved?” – “What quality does the data have?” – “How can we provide a wider picture by using data from other sources?” are just some questions to be answered. At this point I would like to share three learnings with you that have helped me in my own work as Data Scientist.

Data Science Marcel Hellman Portrait
Marcel Hellmann

Marcel Hellmann studied physics in Aachen and Heidelberg and completed his doctorate at the Cancer Research Center in Heidelberg. After six years as a business analyst in corporate IT at HeidelbergCement AG, he joined Vorwerk Digital in 2017 as Senior Data Scientist.

My personal tips for understanding data and using it

Tip 1

Look for the right tools

Find tools to help you cross the boundaries of different data sources and make all your important data available at a glance to become searchable. If databases A and B are separated, it can be a lavish process bringing them together, that nobody wants to go through. At Vorwerk we therefore have established a helpful but clear technology stack – from MapR for processing and preparing data to ClickSense for creating individual dashboards, that serve as showcases.

More on this topic

How Big Data helps us optimize the cooking experience

Tip 2

Look for flexibility

If you want to make devices smarter using data, you need an environment where you can try new things and develop prototypes quickly to iteratively approach a problem. At Vorwerk Digital I’ve found that agile setup and work for a department that acts as a kind of start-up within the company.

Tip 3

Look for the right questioners

Data science is about asking the right questions. But one data scientist alone can never know all the relevant ones. Rather, he must look for the right questioners from different perspectives and departments – whether from product owners, engineers or the direct sales representatives. In my colleague Maria Coronado, for example, I’ve found a competent questioner in the field of marketing.

My most important personal learning is: Only in a functionable team setup, that brings DataScience and Product Intelligence teams together, is it possible to complement the physical product with virtual "data-fueled" products to guarantee a seamless customer experience.