The boom in peer-to-peer (p2p) lending we have seen over the past few years has provided us with a fascinating look at how an arguably new industry can develop so quickly.
What is less apparent at first glance is all the work entrepreneurs and their companies are doing to develop and continually refine small parts of the whole that have allowed so many of them to establish a niche in this rapidly evolving sphere.
Ben Duran of Borrower’s First[/caption]
Ben Duran is the Chief Risk Officer of BorrowersFirst, an online, peer-to-peer lending platform. Their approach looks to both simplify and accelerate the traditional lending process. Borrowers get fast decisions on eligibility which grants them access to those willing to invest in them. Lenders, if all goes well, have access to credit-worthy borrowers.
Mr. Duran brings 15 years of experience to BorrowersFirst, including stints in risk assessment at PayPal, Dell Financial Services and most recently at Teled Analytic Solutions where he worked to refine methods of predictive modelling.
Mr. Duran spent some time with BT to discuss predictive modelling, risk assessment, and the rapidly evolving world of p2p.
1. More and more companies are getting into peer-to-peer loans. It will get tougher to differentiate. Where will that come in? Quicker screening? More predictive modeling?
Having the experience to develop a basic lending product is an important first step to setting up a p2p lending platform. Having the ability to adjust quickly as the industry evolves is also important. Borrower acquisition is and will continue to be a challenge for any platform. So it’s important to differentiate your product and services so borrowers believe it is the best.
This means a streamlined application and fast approval and funding process, a personalized experience and excellent customer service. Creating and promoting brand loyalty, meeting online borrowers where they are via the right channels and partnerships, and providing fast access to the money they need, will be the hallmarks of a successful platform.
2. With so many institutions buying up peer-to-peer loans, are we running the risk of taking the peer out of it? Is that bad?
The p2p space is relying more and more on institutional investors capable of committing large dollar amounts to fund loans. Retail investors are increasingly disadvantaged as sophisticated investors with faster computers and sophisticated decision-making tools take over the loan acquisition process. While retail investors will likely always have a place to acquire loans on traditional p2p platforms, the sale of loans in larger quantities to institutional investors is happening with greater frequency. More streamlined sales of loans to institutional investors will result in greater availability of loans to borrowers.
3. It was reported by Zillow that 9.7 million Americans have underwater mortgages and 18.8% of all homeowners were underwater during the first quarter. Have we learned anything from the recession?
We learned that even the strongest institutions can suffer long term damage if they disregard a borrower’s ability to pay back a loan. Lenders should provide credit only to applicants who can afford it, regardless of the temptations for increased profits.
4. Is such a climate more conducive to p2p’s long term survival because people continue to get the same treatment from the traditional financial system and they continue to spend?
The recession has created an unprecedented pullback from traditional consumer lending that created a vacuum in which p2p lending was born. BorrowersFirst is well positioned to meet the borrowing needs of consumers, regardless of the type of product – installment loans is just the beginning. Even as traditional banks evolve, BorrowersFirst is built on a streamlined platform meant to anticipate where borrowers will be as opposed to where they have been.
5. Where are we in the development of predictive models for lending and credit, near the beginning, maturation, or somewhere in the middle?
Predictive modeling has been available for credit decisioning for decades, and the process is well known and understood. New modeling “techniques” are announced every few years, but the approaches of the early years still provide an extremely robust toolset for model development.
Mistakes are made in believing that anyone can create predictive models with the proper automated tool and knowledge of a p-value. The skillset to properly build models is still extremely valuable and highly sought after. Experience is required to avoid the mistakes that can lead to huge downstream problems.
6. If there is still room for growth, where will that growth come?
Current p2p lending has only touched the surface of the potential opportunity in consumer lending. There is currently $11 trillion dollars of consumer debt, including over $700 billion in credit-card debt alone. There is a massive opportunity for online lenders to tap into this market for excellent returns.
Now that p2p has proven that online lending is secure and viable, investors of all types are flocking to buy the paper. Online e-commerce marketplaces have transformed retail, attracting more and more consumers to buy goods online with choice, value and a fast, easy experience. Online credit marketplaces are transforming borrowing. Borrowers want what we at BorrowersFirst call “Connected Credit” — fast and easy access to money via an online platform, in the same way that they effortlessly shop, socialize and connect online.
7. Do well-honed models run the risk of freezing out certain groups of people who are a good risk because they fit a pattern? How do you prevent against that?
The goal of any model developer is to minimize the effect on “good” customers who fall into the “bad” score range. Models are the best way to manage the situation, but some good risk customers will always be adversely affected in a score cutoff approach. A risk manager’s job is to continually develop and enhance models to make the best decision possible and reach more “good” customers who are traditionally considered poor risk.
8. What are some of the things you have learned about whom to call and when to call them that contribute to a more efficient and therefore productive collection process?
Collections actions have definitely changed over the years, with optimization software providing the best time and location to call. It’s a well known fact that constantly calling to collect will not provide the best results. Traditional collections tactics might alleviate the symptoms, but don’t resolve the root issue. Most people want to pay their bills, but often are unable because they get into a situation they can’t handle or experience an unexpected event that affects their ability to pay.
Lenders can hopefully begin to do something different to help implement a more effective collection experience. How can we provide support for borrowers when an unexpected event occurs that diminishes their ability to pay? BorrowersFirst will be a leader in our industry in developing creative solutions to solve the root problem, instead of a band-aid approach that might have a positive short-term impact, but a negative long term result.
9. By the fact it’s done online, these processes can lead toward the impersonal. What can be done to humanize online lending?
Building products consumers really want helps to create that personal touch. Online lending was created because of the demand generated by the consumer – the desire to have the option to do everything from home, including apply for loans. Creating a friendly brand, engaging tone and thoughtful interactive process can humanize online lending.
There are many online brands that imbue their experience with warmth and playfulness. And because operational costs are far less than in traditional banking, there should be a focus on investing in the type of call center, loan servicing and customer service operations that go the extra mile in responding to customers and creating a human connection. In any business, the goal should be to recognize your customer and find ways to connect.
10. What factors contributed to your decision to sell whole loans to institutions and fractional ones to retail investors? Was it solely risk minimization or were other concerns involved?
P2P started as a retail investment product only. It was only later on that institutional investors began to buy the loans, which provided the P2P players the ability to quickly grow their capacity. Institutional lenders who can commit are capable of committing significant funds to buy whole loans. Making the process more efficient. We also have on-balance sheet capital available so we can instantly fund all loans at all times. For this reason, and the others I outlined, we don’t sell fractional loans to retail investors.
11. Who is bowing man?
Seeing the complexity and friction in platforms that attempt to serve borrowers and both retail and institutional investors, we narrowed our focus to the borrower—to create the best online “Connected Credit” experience for them. Having institutional investors buy whole loans from us, combined with our on-balance sheet funding, makes for much less heavy lifting.
Our name—BorrowersFirst—perfectly reflects our focus and business model, so we designed the “f” in our logo in the image of a person bowing in service to the borrower. Our brand promise is to deliver on the golden rule to borrowing: to treat every borrower in every way as we would want to be treated.
12. What happens to the industry when rates go back up?
Rates are based on the risk associated with an individual. If unemployment goes up or other events suddenly make lending to borrowers more risky, then rates could go up, but our job at BorrowersFirst is to provide the most competitive rates available for each individual – that will be true regardless of credit environment or the economic outlook.
13. You have the fastest decision time I have seen as of yet, what was the model development process like for that?
Decision times in general have been reduced as technology has provided faster and faster solutions. Our system has been optimized to deliver a quick response, but can be attributed to the experience of our technology staff. The data pulls are the most time consuming, once a credit model has been built, the implementation of that model within the system and time to process is minimal compared to the data pulls.
Although others might eventually utilize similar techniques to speed up the lending process, having the ability to originate loans that we will hold on our own balance sheet will make our lending process quicker than any process that requires auctioning loans to the public.
Contributors
