Customers should share employees' view of their data, says Revolut's Lead Data Scientist

If data is oil, Revolut is giving everyone a drilling rig

Data is "beautiful". Data is "gold". Data is "the new oil". Perhaps no other word in technology is the subject of so many metaphors, and they all mean the same thing: data has value. But, much like oil, data can also cause serious damage.

Fintech firm Revolut, which describes itself as a 'digital banking alternative', knows exactly how important data is; not only to business success, but also to its reputation. That is why the company emphasises giving its customers a nearly-unprecedented level of access to the data it holds on them.

Abhi Thanendran

"It allows the customer to understand stuff about them that we are able to understand as well," said Lead Data Scientist Abhi Thanendran. "In a lot of cases, if you use any software or any program, most of the mobile apps, the data science teams at those companies probably have a really, really strong understanding of who you are - more so than you may do, based on your actions on the application.

"A good example would be Facebook: Facebook has a really good understanding of the network of people you know, and what the stuff they like is and how many people there are, etc. That's the kind of information you probably don't think about on a day-to-day basis. So at Revolut, while we are able to understand similar types of things, we want the user to be able to know the same stuff that we know. That's what we're working towards."

We want the user to be able to know the same stuff that we know - Abhi Thanendran

Thanendran's opinion is that "Customers should be able to see whatever is relevant about them," making Revolut's approach quite different to Facebook's. The social media giant has faced criticism (and, recently, punishment) for how it collects, stores and shares data.

The move towards a more open approach to data is growing slowly in the fintech space, but Thanendran expects it to spread to other companies - "once we start rolling out a few of the other things that we've got planned," he teased. Under the auspices of the GDPR, whereby customers can request any and all data about themselves, this trend is likely to be especially quick to catch on in Europe.

Tackling data bias

Customers can view large amounts of the data that Revolut holds on them and so, of course, can employees. Like any modern organisation, the company leans on this to improve its service.

Firms like Amazon have been accused of making poor decisions based on bad data. Revolut, on the other hand, attempts to use data to challenge and mitigate peoples' preconceptions.

"It's easy for people to make decisions based on some bias that they have. The thing is, whether people want to be unbiased or not, they're always going to be biased; it's just inherent. So the way to mitigate this is to actually use numbers to justify your bias or disprove your bias.

"Naturally every single employee is going to have to make decisions. To make sure they're making the right decision, we use numbers to back it up, and basically, to get numbers you need data.

"So, everyone has access to data; it's just that they don't have access to all of the data, so people only have access to what's relevant for their role. We allow them to query the data and answer the questions that they're going to be asked when they do have to defend their decision to their manager or whoever."

Every Revolut employee thus has some experience in data analysis. While every team does have a data scientist, this approach avoids the need to also hire dedicated analysts; a role which, Thanendran admits, is not easy to fill.

"The data analyst is...not something that a lot of people would want to stay in for a long time, so you'll naturally have a lot of turnover if you do start hiring those people. It's just a problem we don't want to deal with."

Revolut is not immune to bad data, although Thanendran notes that missing data is a more common issue. "We have a lot of code that deals with taking data from various sources and putting it into [analytics database] Exasol," he says. "Sometimes the code fails and you'll have missing data, which can make someone's analysis seem inaccurate. That does happen, but it's easy to notice; it'll be highly anomalous."

If data really is analogous to oil, Revolut's data practice is equivalent to giving everyone - customers and employees alike - their own tiny drilling rig.