Data analytics are all fun and games until Uber accuses you of a one night stand

Customer data is a valuable resource, but what happens when it's misused?

Orlando Machado works at Aviva, but he's previously held posts at the BBC, Wunderman, Dunnhumby and MoneySuperMarket - all using big data analytics. He has two kids, lives in East London and uses Uber.

Why is that relevant? It might not be to you, but if you were an insurance provider like Aviva - whom Machado works for, as the global director of customer analytics and data science - then this information would make it easier to make a decision about his premiums.

Aviva was founded in 1696, but is increasingly trying to put data at the heart of its business. While, like other providers, it has a long history of using risk and pricing science, it is now using ‘customer science' to understand the people who are buying insurance.

"The more we learn about people, the more we can connect with them," said Machado. "We think we can use data science, AI and machine learning to understand customers better than they understand themselves… When we think about technology, we think about bridging the gap between us and our customers."

Aviva segments its 16 million customers "in a way that's quite traditional in one sense, and quite contemporary in another." Each customer is placed into one of seven segments, then 28 smaller segments and finally into 437 micro segments. AI can be used to target them even further.

This enables very granular marketing, testing and learning - but, Machado asked, "Is it too creepy?" Consumers are conflicted. On one hand, they're very protective of their own data; on the other hand, they expect a tailored service.

It is easy to misuse customer data, and a prime example comes from Uber (quelle surprise). In 2012 the company posted a blog called Rides of Glory (long since deleted, but still visible here) in which its data scientists analysed the customers that they presumed to be having one-night stands: taking a taxi from a bar late at night on a weekend and then another from the same drop-off location four-to-six hours later.

Uber's maps, like this one of San Francisco, showed the hotspots for 'Rides of Glory'

Uber included a huge amount of information in its blog post, including maps, hotspots and timelines (Valentine's Day is a low point), but drew massive criticism from customers and the press for sharing very sensitive data.

Insurers are in a similar position: Aviva knows that a customer has a good car, lots of valuable possessions, no window locks and when they're going on holiday. There's a massive responsibility to protect that data, especially under the GDPR.