This article was taken from the June 2011 issue of Wired magazine. Be the first to read Wired's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online.
Moneylending site Wonga is booming on eye-popping interest rates. It's controversial -- but could it's data-led approach transform personal finance?
In an office in St John's Wood in London, not far from Lord's Cricket Ground, samedaycash.co.uk was about to spring to life.
Within ten minutes of the site going live, a customer logged on an took out a loan for around £100. With just a £50 investment in Google Adwords, samedaycash.co.uk had attracted its first customer. And without ever realising it, that person had just become the net's first recipient of a fully automated loan: no human had approved the transaction.
Errol Damelin, one of the ten or so people sitting working at computer screens in the office, had not expected a customer to appear within minutes of the site going live. The moment was pivotal for him. "For me the epiphany was right there," he recalls three years later. "People were online, looking for a solution to a problem."
The second significant event took place a few days later: the site had its first defaulter. The nonpayer was not a type associated with bad credit, but was an apparently respectable man working for a well-known bank.
One successful loan; one failure. The yin and the yang of the credit business. If the first event proved to Damelin, now 41, that consumers are actively looking to borrow cash online, the second may have seemed like a setback. But, in fact, the failure confirmed something else that Damelin has learnt: when it comes to people, you can't trust appearances. All you can trust is data.
Over the coming weeks, the pattern repeated itself: for every successful loan, there was a defaulter. If this were all there was to the business plan, SameDayCash would have been an expensive catastrophe.
But its strategy over this period wasn't just to disburse money -- it was to accumulate facts. For every loan, good or bad, SameDayCash gathered data about the borrowers -- and about their behaviour.
Who were they? What was their online profile? Did they repay the money on time? The site was feeding an algorithm that would form the basis of Wonga, launched a year after the beta experiment that was SameDayCash.
Within a year, Wonga had issued 100,000 loans, worth £20 million, earning about £15 million by charging interest at an eyewatering headline rate. Eighteen months later, the company had issued over a million loans. By disrupting a market the banks had underserviced -- small, shorter loans -- Damelin has found a highly profitable niche. But his ambitions go beyond high-margin money-lending.
With £90 million of venture capital behind it, Wonga may be the most radical financial-sector disruptors in years: using the internet to transform the relationship between customer and lender, Damelin wants nothing less than to reinvent retail banking.
But not everybody sees Wonga as an exciting technology startup.
Its growth has been accompanied by ever louder assaults by those who see it as predatory. The Daily Mirror castigated the company for charging "4,214% interest" (though that headline APR is, by law, clearly displayed on the website). Labour MP Stella Creasy has launched a campaign against Wonga after denouncing it as "legal loan-sharking". When Wonga sponsored free public transport in London on New Year's Eve, a BBC Radio 4 documentary, The Report, suggested that the firm was responsible for luring customers into a cycle of expensive loans. "The fact that previous free New Year's Eve travel sponsors include NatWest is a sign that the likes of Wonga are cashing in on a market that has been abandoned by high-street lenders," it claimed.
Damelin, who talks at an alarming speed in a soft South African accent, argues that data should be trusted over prejudice. To answer his critics, he points to a 2011 Wonga customer survey conducted by Populus. From a sample of 1,500 borrowers, he says, 95 per cent said that they were "satisfied" or "very satisfied" with their experience of his company. These are numbers that conventional, regulated, high-street lenders would be extremely happy to achieve. (Which?'s recent customer satisfaction ratings put the Bank of Scotland at 41 per cent. The industry average is 61.) Damelin points out that the Wonga numbers are better even than
Apple's customer ratings -- 86 per cent, according to the American Customer Satisfaction Index.
Sitting in the conference room of Wonga's HQ -- a pair of immaculate stucco-fronted houses built by Regency architect John Nash overlooking Regent's Park -- Damelin wears a black jumper, jeans and Timberland boots. Wonga's HQ doesn't look much like a bank. The rooms are filled with artworks by the company's cofounder and chief technology officer, Jonty Hurwitz. He and Damelin, both South Africans, are close friends. But in many ways, they are opposites: Damelin is a sports fanatic who claims that he absorbed all he knows about business on the rugby field and the athletics track. "That's where you learn about pressure and how to deal with it," he says. "Doing the 800 metres, that's where you learn about long-term goals and short-term goals."
Damelin is still extremely active. He's a keen marathon runner,
mountain biker, sea kayaker and skier. Hurwitz, on the other hand, is an artist who describes his own sports career wryly as "playing thirteenth batsman on the cricket team". A mathematician and physicist turned software entrepreneur, he has spent the last five years -- while at Wonga -- reinventing himself, with some minor success, as a professional artist. Hurwitz's sculptures fracture heads and bodies into thousands of pieces that come together only when looked at from the correct angle, much like a complex mathematical problem. In the small reception area, a blue-and-red sculpture of two half men becomes whole when seen in a pair of mirrors. "The best partnerships," Damelin explains, "are based on people who overlap as little as possible." Their overlap was as entrepreneurs; both had had a successful track record. But Damelin decided it was time to do something truly ambitious. He describes how he wanted to move into what he calls "a great space" -- great in the sense of exciting and big. That meant the consumer market. "I got to thinking about consumer credit," he says. "The thing I loved about it was that it was so badly provided for. People were getting a raw deal. This is crap." His sense of what consumers want from a loans company is straightforward. Damelin thinks they look for three things that banks and conventional lenders are incapable of giving. Firstly, simplicity -- the ability to borrow what they want, when they want.
Secondly, speed -- the transaction needs to happen fast. Thirdly, they want to know exactly and clearly what the loan is going to cost them whether they pay on time, pay early or even miss their deadline.
Banks, he says, have no incentive to meet these needs. The internet has changed retail by offering customers speed, choice and value, but banks, Damelin believes, have failed to offer these advantages, choosing instead to remain an analogue, conservative oligopoly. "Banks love regulation," Damelin says. "They have been better than anyone else at co-opting it to suit themselves. They love embedding themselves into the messy greyness of how policy is created. And that makes it harder for them to innovate. They're all wink-wink. They don't compete. When did you last hear of a bank competing to bring down the cost of CHAPS payments? When was the last time you saw an interesting new Barclays product? I don't think ever."
But he believes that the sector is about to be disrupted by innovation and that, within ten years, the banking establishment will be broken. "Soon it'll be totally clear there are alternatives," Damelin says. "And within 30 years we're going to find the idea that we even had to have a bank hysterical."
In 2006, when he and Hurwitz first started looking for backing for Wonga, potential investors saw the short-term, small-loans business as an unprofitable, risky backwater best left to pawnbrokers and loan sharks. But the Facebook generation, Damelin realised, have a different attitude: they want flexibility and choice - and they expect things to happen fast. "The web changed expectations," Damein says. "Amazon changed our expectations, as did Google and so did Facebook after that. Things that used to take ten minutes, and seemed fast, now seem mundane."
I first met Errol Damelin in a café in Hampstead," says Robin Klein of The Accelerator Group (TAG), which specialises in early-stage investment in startups. It was June 2006. "He put [his] idea to me.
Here was this guy who had no background in this kind of technology yet he was clear there was a significant opportunity to disrupt the market."
Klein has an impressive track record as an entrepreneur and investor. His mailorder company, Innovations, was the first UK business to allow payment transactions on the internet and the first to hit £1 million in online sales. An early investor in Lastminute.com, Klein advises the startupstimulator programme Seedcamp, run by his son Saul. The father and son are partners in TAG, which invested early in bit.ly, LOVEFiLM, Moo.com and Tweetdeck.
Robin Klein admits to being initially cautious about the ethics of short-term loans. But as he learned about the business model, he became reassured that it wasn't based on exploiting the needs of the desperate. It was about meeting the expectations of a generation that had long expected flexibility and speed.
Less of a stumbling block was the fact that Wonga, at this stage, was just an idea from two people with zero experience of retail banking. Damelin already had two multimillion-dollar startups to his name. His first was Barzelan, a steel-wire company that he cofounded in 1997 in Israel, where he lived for eight years. Then came Supply Chain Connect, a business that operated in the UK and US producing software to monitor inventory. "I have enough pattern recognition from everything I've done,"
Klein says, "to know Errol was going to be a great entrepreneur."
Working with one developer, then a second, Damelin and his team started to map out what the algorithm would look like before a single line of code was written.
The team's objective was to do something disruptive: to make the process of borrowing money completely automatic. Applying for a conventional bank loan -- especially for unsecured loans -- is slow and bureaucratic, partly because it requires human intervention. A sophisticated algorithm would be able to make an instant decision, lending at the kind of speed that consumers expected from other transactions. That means that all the information that the algorithm uses to make its judgment has to be available online.
The other reason Damelin was attracted to automation runs deeper. "People are all different," says Damelin. "People are complex. People can't be put into little boxes."
But data doesn't care who you are.
Wonga's timing has been fortuitous. The financial squeeze has meant that household budgets are stretched and high-street banks are reluctant to lend. "It's interesting when you look at the critiques of the crisis,"
Damelin says. "What you hear a lot of is, 'Oh, I remember when I was young and I used to have my own banker and he knew me and he knew my parents and I used to have to dress up in a suit to get a loan. That's what we should get back to.'"
That is the wrong response to the crisis, Damelin says. When it comes to finance, metrics should always supersede managers. "Name a person who doesn't think they're good at reading people," Damelin says. "Everyone thinks they're good at it, even though we're not."
Trusting data over appearances, he says, is partly down to his background in South Africa: "Prejudice and generalisation are something I grew up with."
As a student at the University of Cape Town during the apartheid era, Damelin found the politics of the time inescapable. University campuses were cornerstones of anti-apartheid activism, and Damelin became a member of the student representative committee.
The apartheid era left its mark in a way that would become crucial to the inspiration for the Wonga algorithm; if Damelin's background had taught him one thing it was that people are easily prejudiced, and that prejudices make for bad decisions. If you could remove the human factor, you could make better decisions. "I think when people are saying that a good old bank manager should make the decisions," says Damelin drily, "what they're really saying is some middle-aged white guy should make the decisions."
The Wonga algorithm first has to discern if the person applying for the loan really is who they say they are. In a world in which
credit card fraud costs the established UK banking industry an estimated £440 million a year, this is no easy task. Then it has to work out if the person asking for the loan is able and willing to repay it when they say they can.
Damelin started talking to high-street banks: if he wanted to make this as instant a process as possible, he would have to find a way to get money into borrowers' accounts as quickly as possible.
They were dismissive of Wonga's business plan. "They told us we wouldn't be able to verify customers accurately," he says. "There was unanimity about that. We could never be totally satisfied that the customers were who they said they were without seeing some kind of actual document from them."
A new bank customer needs to show both a photo ID and a letter from another bank, a utility firm or government agency proving that they live where they claim. On its own, a mobile-phone account proves little. Wonga's hunch was that the net -- as a giant data source -- would provide deeper insights into consumer behaviour.
What if you crossreference a standard form of ID against an email address, plus the organisations that the address is linked to, plus a plethora of other data sources, from the electoral roll to Facebook? The information that consumers consciously gave was just a starting point.
The Wonga home page features two horizontal sliders. The first allows borrowers to adjust the amount of money they want to borrow -- up to £400 on a first loan, as much as £1,000 for returning customers. One slider moves in increments of £1. The second slider shows the period of the loan in days. As the sliders move, the site calculates the cost: a loan of £250 for 21 days would cost £58.42.
Even though the loan is for three weeks, the site must also legally show that eye-popping annual-percentage rate (APR): 4,214 per cent.
There are no other costs. "We do small, short-term things, and the cost of delivering that service is high," Damelin says. "Catching a cab might be expensive, but it's convenient and nobody complains that being charged £15 for getting across London is immoral."
Once a customer has decided on the size of a loan, he is taken through a series of questions. Within about 15 minutes, Wonga has retained around 30 pieces of simple information about a potential borrower. And, from those pieces of information, Wonga has found it can access a further 6,000 to 8,000 online data points that relate to the applicant. What exactly are those data points? Damelin won't say; the decision engine lies at the heart of Wonga's IP and he is reluctant to offer clues.
Following first-stage funding of £3.7 million by Balderton Capital in 2007, Accel Partners joined Balderton and Greylock Partners in offering second-stage funding of £14 million in 2009.
Hinting at the originality of the site's data capture, Accel's Sonali De Rycker, now on Wonga's board, says: "They use a lot of social media and other tools on the internet you don't even think about. That's where the magic is."
The crux of the algorithm is less about the individual pieces of data -- your postcode, the colour of your car, how large your mortgage is -- but how these pieces of information relate to one another. Crucially, the data points are stacked against the other pieces of information gleaned from past Wonga clients. By the time Accel came aboard in 2009, Wonga had issued 100,000 loans. That's 100,000 data sets contributing to an ever growing net of information, and each comprising 6,000-8,000 pieces of information about a borrower. "You build the story by joining up lots of data,"
Damelin says. "We pay for that data, but we need it. It's about computing thousands of combinations to look for things that look wrong - or right."
When it launched in 2007, the decision engine wasn't efficient.
Half of the clients defaulted: investors were paying to allow it to fail. "We were happy to underwrite losses because if the engine didn't learn, then you'll never have a good business," De Rycker says. "The question was how much to underwrite. That was the decision."
Until Wonga, the benchmark formula for estimating credit risk has been FICO scoring, which awards applicants a score between 300 (high risk) and 850 (low risk). Named after the company that developed the algorithm in 1956 -- Fair, Isaac and Company -- the FICO score is widely used in the US. Like British banks, American institutions take their information from credit bureaux such as Equifax and Experian in order to calculate whether a potential borrower is likely to default. "I'll tell you something very interesting," says Damelin with a conspiratorial smile. "When we started our business, we were having to rely on credit scoring from agencies because we didn't have our own data."
It was that conventional scoring that produced the 50-per-cent default rate in the first few weeks of the operation. Wonga soon found that FICO-type scores - which were designed for longer, larger debts - were next to useless for predicting behaviour of short-term debtors.
While the data that Wonga accumulates becomes ever more complex, it no longer has to rely on the same systems as do the banks. Today its default rate for what is assumed by the banking industry to be the most high-risk market for loans - short-term lending - has fallen to single-figure percentages. The banking industry typically estimates around ten per cent of credit-card debt as uncollectable. (Damelin refuses to reveal his business's figure, but insists it's "definitely industry-leading".) "So we've built an engine that is dramatically more predictive for what we do than FICO," says Damelin. "Dramatically on a scale that's unbelievable. And that was always our thesis that we could."
The company had started with same-day payments. "Same-day money was the bare minimum," says Damelin. "We wouldn't have started this business unless we could do that. As soon as we got it there the same day, it was, 'Oh, can we do it the same hour?'" Damelin remembers. He still believes that he can shave time off the current guaranteed time of 15 minutes. "We're aiming for instant," he says.
The company has been collecting user data to feed its algorithm for four years now, and because the loans are typically for only around 16 or 17 days, there is plenty of new data always arriving.
Wonga's real potential disruption lies not so much in the profits from the short-term-loan business it has built, but in what that business is telling Damelin and his colleagues about consumer behaviour. What Damelin believes he is in the process of building is not just a small-loans company, but an Amazon for the financial-services market.
If Wonga proved that it can outperform the high-street banks in the short-term loan sector, where else can it achieve the same thing? Damelin believes that there's massive scope. "I can't imagine us moving into investment banking, or commercial banking," he reveals. "But in consumer banking, think where else do speed, convenience and transparency interest the customer? So this vision is bigger than short-term loans. Where precisely? I would rather not say."
Around 60 people work from the London office, alongside new hires in the US -- where Wonga is looking to expand -- and a customer-service team in South Africa. It has a total staff of around 100.
The weekly senior-management meeting gathers in Damelin's office. A mix of developers, marketing people and lawyers discuss strategy. Some wear T-shirts and Reeboks, others suits and brogues.
Upstairs is where the developers code.
They're still the biggest single team in the company. Though the simple user interface has looked constant over the last four years, the back end is constantly changing. Since launch the software has gone through over 30 releases.
In the basement, the five-strong customerservice team deals with queries and collections. They're supported by the 15-strong team in South Africa. The system is not yet fully automated: each day, a handful of applications have to scan documents for verification. On one phone line a couple who have filled in their bank details incorrectly are trying to borrow £50 for seven days. They whoop when the loan is approved.
Sitting at a computer, customer-care executive Tarik Abdellah deals directly with clients. There are, inevitably, casualties. A man who borrowed £1,000 before Christmas has stopped repayments. He has been too scared to talk to Wonga. Abdellah calms him: he tells him that he has frozen the repayments and deducted £200 from his account, bringing the debt down to £1,500. "Right. How much can you afford to pay each month?" he asks.
By the end of the conversation the customer appears almost grateful. At no time has he been threatened with court or chided. "I appreciate that," he tells Abdellah. Although he has paid hundreds for a £1,000 loan, there is no anger or bitterness in his voice.
But many borrowers appear to prefer not to talk to a representative. They are more likely to express concerns on the site's forums or its Facebook page. There, a customer writes: "5 emails and countless times on hold with wonga today and no answer?
Can you give me information on how to contact help desk?" "The way we looked at it," Damelin says, "was: 'Where do people want to engage with us on their terms?' Well, they're on Facebook already. For us that's a risky business because our customer-service stuff is all in the public domain. But we're being super transparent." On Facebook, a customer writes: "how come your only offring me silly amount of money when ive been with you for over a year and always payed you back on time????" (sic).
Wonga is not the only firm attempting to disrupt the industry's dead hand. Alongside it is one of the largest investors in peer-to-peer microfinance: Kiva (see 06.10), which has provided $204 million (£125 million) in loans. Peer-topeer initiatives offer low-cost loans, sourced by funds from individuals. The US-based company Prosper.com has so far lent a total of $224 million. In banking terms, that's peanuts.
These initiatives excite Damelin but he's also critical because he considers them "impure" in their decision-making. "These peer-to-peer guys are not innovative technically," he says. "Their verification is manual. I like Prosper. The CEO, Chris Larsen, is a friend. What's great is they put their data in the public domain.
But their model is still, 'I know John. He went to a good university, blah blah. I'll give him £10,000.' Prosper have got 40 percent bad debt. The reality is that John defaults on them nearly half the time."
Laurie Azzano of Prosper tells Wired: "The statement by Errol Damelin regarding Prosper's credit decision model is incorrect:
Prosper's credit decision engine is 100 percent data driven and delivers defaults closer to 5 percent, not 40 percent."
Damelin's business is growing rapidly: last December, it passed its millionth loan. It won't say, but it seems reasonable to calculate its total loans so far at around £200 million. In February this year, the company received a third stage of investor funding of a further £73 million from Oak Investment Partners, Meritech Partners and, notably, the second-largest medical-research charity in the world -- the Wellcome Trust. It's as if a controversial moneylender were starting to be seen as almost respectable.
This article was originally published by WIRED UK
