Data Science, Day to Day: What’s Typical?

Using data analytics to make business decisions is obviously a major part of business plans across industries. But data science itself is still so new that even within the same industry, knowing what kinds of problems companies are running into on a day-to-day basis is elusive.

In a recent webinar was organized by Mindtree, “Analytics Best Practice Webinar from the Experts at Allianz Group and Unilever,” Mindtree noted, “Analytics problem statements are very cryptic. It could just be a one-liner.”

So we asked our guests—Dr. Andreas Braun, Head of Global Data & Analytics at Allianz Group, and Aneesh Chaudhry, Global Analytics Director at Unilever—about their typical problems and challenges are, and how they approach them.

Allianz: Cataloguing the Best, Testing the Rest

Braun said that Allianz, a financial services company with core business in insurance and asset management, has come up with a catalog of 120 or so “umbrella use cases.” These are templates that they can demonstrate and show to businesses, with evidence that they can be implemented very quickly and bring meaningful benefits.

One of the biggest areas for these use cases falls under customer and consumer analytics—ways to cross sell, up sell and improve customer retention, for instance. In this area, Braun says there are some very simple and proven analytical approaches which that companies can rely on.

“It was quite impressive to see that even small analytical advancements, in a group as big as Allianz, made a significant business impact,” said Braun. “For example in a small country we had a first use case for retention. We improved customer retention by 20 percent or so, and the top line improvement of the whole business of just 2.1 percent or so. But add it up and we are talking half a billion here, so that was a really quick win.”

On the other hand, Allianz is also using analytics to experiment and test new customer products and business models. But he says they do this testing abroad, for instance in Asia. An example: using health record data to predict diseases like diabetes.

“It creates a win-win situation, he said. “We use technology and analytics to improve the diabetic situation of such a person. And of course it can also save a lot of money too if people don’t get seriously sick.”

Unilever: Challenging Itself to Apply Data Science Widely

Aneesh Chaudhry noted that, as a consumer goods company Unilever’s business and data science applications are relatively simple. “There are some obvious ones—consumer profiling, supply chain optimization, market mix modeling—but these are well embedded and well understood and those conversations are easier,” he said.

The challenge, he said, is often that the demand doesn’t come to the analytics team because it’s not apparent to the line leader elsewhere in the organization that their problem can be addressed with analytics.

“If I have an HR line manager, it takes some work and probing to convince them that there is an opportunity to use big data to help make recruitment decisions,” he said, noting that the key challenge is to identify the opportunity. “And the way we tackle it is we move away from asking what analytics problem they want to solve, and instead start with a business question they’re trying to answer. Then we can discuss collaboratively how to apply analytics, data science and machine learning to this question.”

Do you have similar experiences inside your organization? What are the “day-to-day” uses of data science that you see? Please join the conversation below. And if you missed it, click here to check out the full session of the “Analytics Best Practice Webinar from the Experts at Allianz Group and Unilever.”