We know that as we browse the internet, we leave behind a trail. Search results reflect our browsing history; usernames and passwords are remembered on long-forgotten websites; and personalised adverts increasingly seem to follow us around.
But what if our browsing habits, and how we access different websites, came to form part of how we are assessed for credit? And what if this were combined with an assessment of us, as individuals, using personal information gleaned from our social networking accounts?
Actually, this is already happening. Not to everyone, but to those using the short-term, high-cost loans being offered by a select group of online payday lenders. In the UK, it is happening to customers of the payday lender Wonga.
To understand exactly what Wonga is doing, you need to appreciate two crucial points. The first is that, like many lenders, Wonga relies on credit scoring when making lending decisions. Its model is not conventional, however. It does buy third-party credit reference data – provided by familiar credit reference agencies such as Experian – but this does not form the basis of its system. The reason, claims Wonga’s founder, is that its own scores are “dramatically” more predictive.
The second point is that speed is a crucial part of Wonga’s pitch to borrowers, with decisions made within six minutes. This means that Wonga has to use data that is available instantly and online. Given that Wonga claims to rely on 8,000 different data points in its decision-making process, where is all this information coming from?
Well, as is increasingly understood and as I’ve written about in the past, Wonga and similar sites rely on a huge range of information “leaked” by internet users, fed through sophisticated algorithms, to make their credit assessments. We know this could include the time of day a user accesses the site, locational data (probably gleaned from your IP address), the device you are accessing the site on, your browser type, how you behave on the site, to name but a few. If you’re curious about some of the data you’re leaking right now, there are plenty of sites that can quickly tell you.
This information might seem mundane, but when fine distinctions are needed to separate debtors who it is suspected will repay from those that won’t, each additional data point could be crucial.
This thirst for information is also driving Wonga and similar sites towards social media. During the application process, users are given the option to “connect” to Facebook (this actually means installing an app). This part of Wonga currently appears to be under development: when you try to connect, you get an error message. That said, the URL itself tells you all you need to know, providing a complete list of the permissions being requested.
These permissions, if granted, would give Wonga access to information that would help confirm the identity of a user, including birthday, hometown, and location. The app also seeks information that might verify or undermine the income level declared by a potential borrower: educational history, work history, as well as relationship details.
Perhaps more surprisingly, Wonga is also interested in seeing “softer” information. This includes the user’s “likes” and listed interests, games activity, religious and political views, any subscriptions they might have, groups the user is part of, and their personal website.
While Wonga is not allowed to copy details out of Facebook’s databases wholesale, it could search this information against a potentially infinite variety of terms and test the predictive quality of this analysis as part of its own scoring models.
There is one further particularly powerful permission buried in the request, called “read_stream”. This not only means Wonga could see all of your Facebook posts – a highly intimate level of access in its own right – it also means it would be able to see what your friends are doing, as shown in your news feed.
As researcher Bernhard Rieder writes, what Facebook in its description breezes over as just “posts in the user’s News Feed” might be more accurately translated as “a minute account of your friends’ activities”. Here it is worth noting that an app run by credit scorer Kreditech is currently operational and, in fact, requests permission to see even more data. What’s more, there’s a discount for borrowers who grant access.
Much debate around payday lending has focused on the cost of the loans. This is understandable: Wonga’s current Annual Percentage Rate (APR) stands at an eye watering 4,214%.
But high interest rates shouldn’t distract us from the issues of privacy arising from lenders’ use of often unwittingly leaked online data. Is this a new instance of what some sociologists have called “consumer surveillance”? And what are the social justice implications, given that it is payday lenders who are driving the use of these techniques rather than more upmarket credit providers?
For, even if some of these technologies are unfinished, what is clear is their ambition: to know far more about your online self than you probably thought possible.