Setting up a pool of data scientists to serve the different business departments within Innogy. That was the initial plan, says Robert Egging, International Head of Retail Data Strategy at Innogy (E.ON group). At least until they examined which topics the different departments were requiring data support for. The needs revealed to be in a different area.
Egging’s team discovered, on the basis of a so-called ‘data stock taking’, that analytical capacity was not the issue. “Our people were able to develop the most beautiful models from a multitude of sources. However, these models only provided limited answers to the actual needs of the business owners. The efforts in the field of artificial intelligence and data science consequently yielded less business value than potentially attainable.”
It turned out that there was a particular need for an ‘execution manager’, i.e. someone who ensures that the developed models actually do seamlessly connect to the user stories of the business owners. This person should additionally involve the controller, in order to make sure that she also gives her approval to the proposed model and correspondingly signs for the expected return on the requested investments. “This validation process cannot be put on the shoulders of most data scientists. Most of them feel uncomfortable playing this game”, Egging argues.
Therefore, he started looking for someone he, jokingly, calls a ‘consultant and nerd in one’. He could not find anyone fitting this description internally. Luckily, data scientists with a great empathetic ability appeared to be working at Pipple. “They employ business savvy mathematicians. These people like to develop a model, but also to sit around a table with a management team. They are able to explain in simple terms to those teams what they are doing and how this is valuable.”
The value resulting from the collaboration with Pipple is clearly evident. They are able to predict which of the large business customers can easily pay their bills and which yield the greatest conversion. Furthermore, models are being developed to help determining which propositions are most tempting for specific customers.
“By sharing our knowledge and expertise, we are able to jointly make giant leaps forward.”
Egging expects that the number of opportunities will only increase. “We have 22 million customers at Innogy. The impact of an operating model is already substantial very quickly, because it is easily scalable: the same trick can be applied to more segments in even more countries.” The effect is moreover amplified as a consequence of the acquisition by E.ON. “It means a doubling of the number of customers. Besides, our new colleagues also have all kinds of models on the shelf. By sharing our knowledge and expertise, we are able to jointly make giant leaps forward.”