FOUNDER & CEO, ZESTFINANCE
Financial tech company.
Douglas Merrill is the founder and CEO of ZestFinance, a financial technology company that uses its unique credit-decisioning platform to help lenders effectively predict credit risk so they can increase revenues, reduce risk and ensure compliance. As the former CIO and VP of Engineering at Google, Douglas led an organization of 1,500, and drove multiple strategic efforts, including Google’s IPO auction in 2004. He also served as COO of New Music at EMI Group, as SVP of Infrastructure and HR Strategy at Charles Schwab, and as an Information Scientist at the RAND Corporation.
I founded ZestFinance in 2009, in the wake of the Great Recession, to make fair and transparent credit available to everyone. It was a time when using machine learning in banking was becoming part of the conversation, but nothing significant was actually happening. I had previously served as Google’s CIO, and because search and underwriting are both fundamentally math problems, I decided to take Google’s mathematical approach of identifying, analyzing and using data — and apply it to credit decisioning, or the process of giving loans. At ZestFinance, we apply data science and machine learning to help lenders effectively predict credit risk so they can increase revenues, reduce risk and ensure compliance.
Lenders are increasingly adopting new credit appraisal techniques based on “alternative” or nontraditional data, to increase their loan approval rates, while simultaneously reducing risk and ensuring compliance. This is our bread and butter; most traditional underwriting uses 50 data points, not nearly enough to make informed decisions about millions — if not billions — of unserved potential customers. Users of Zest’s underwriting platform can look at tens of thousands of data points about a borrower to paint a more accurate picture of their ability and willingness to repay a loan.
The credit appraisal system is antiquated; traditional underwriting systems have been in place for decades, so there is a tremendous opportunity to modernize those processes and expand credit on a massive scale. Through machine learning, we’re able to replace these outdated systems with more precise, automated methods of determining a potential borrower’s creditworthiness. Our aim is to use our advanced technology to help lenders offer financing to consumers who otherwise might have gone overlooked by traditional credit appraisal methods.
For example, in China, where there is limited data on consumers and in turn, limited access to credit, we’ve been able to work with Baidu to turn its search, location, and payment data into credit scores. JD.com also uses the Zest platform to turn its transaction data into credit decisioning information about Chinese consumers. The opportunity to expand credit to millions of consumers around the world is vast, and we want to continue playing a role in making that happen.
There will always be challenges when it comes to disrupting an entire industry; in our line of work, these could include regulatory changes, entrenched underwriting methods, and even fear of change. That’s why we’re on a mission to educate businesses on the benefits of using machine learning for underwriting and the potential to reach millions of underserved consumers.
There are so many people in the world who either can’t access credit or only have access to expensive and abusive loans; this is especially true in the U.S., where people like millennials are often denied credit because they lack financial history, and in emerging markets where credit is limited or non-existent. At ZestFinance, our vision is to make fair and transparent credit available for everyone.
Right now we’re focused on helping lenders transition from traditional to machine learning-based underwriting. Using Zest’s ZAML platform, we want to expand credit to new borrowers in the U.S., as well as in new global regions where lenders have difficulty scoring “thin-file,” “no-file,” and other hard-to-score borrowers.
We’ve been fortunate to work with some very prestigious partners who are committed to expanding credit access for traditionally un- and underserved consumers. Baidu increased its approval rates for small-loans by 150% without a rise in losses; a major American credit-card issuer can now safely extend an additional $1B in credit to borrowers it had historically denied, without experiencing an increase in risk; and one of the Big 3 American auto manufacturers now saves $50M annually through increased automation of its underwriting process. We look forward to collaborating with more partners seeking to achieve similar gains.
Machine learning has the potential to transform the financial services industry. But the adoption of machine learning in consumer lending has been slowed by one major challenge: the inability to explain the reasoning of machine learning models. Historically, these technologies have been considered “black boxes”: you don’t know why a machine model made a decision, you just know that the model decided something. That “black box” quality is unacceptable in consumer loan underwriting because lenders have to explain to applicants why they’ve been denied a loan. Black boxes make giving those adverse action notifications to consumers impossible. The ZAML platform includes explainability tools that empower lenders to understand the reasoning of their machine learning models in ways that can be explained to loan applicants and regulators. Machine learning explainability has been one of the most difficult things to address in the field of artificial intelligence, but it’s something we’ve been able to overcome.
Our ideal partners are lenders of all types — banks, credit card issuers, auto financiers, retailers, and technology companies — looking to expand credit to thin-file, no-file and hard-to-score borrowers like millennials. The difficulty financial institutions face underwriting these borrowers is their limited credit histories. Traditional underwriting works well when evaluating borrowers with long credit histories, but when there is limited data, it can’t differentiate between creditworthy and high-risk applicants. ZestFinance’s tools helps fill those gaps by analyzing a vastly broader set of data.
I’ve built ZestFinance with high horsepower colleagues committed to expanding credit in fair, transparent, and innovative ways. Our recruiting efforts focus on identifying employees, whom we call “Zestians,” that share and manifest these values. Our hires are incredibly smart and talented people who thrive in environments where they are willing to learn from others, speak their mind and do good work.
Zestians are encouraged to maintain work/life balance across many aspects: the flexibility to attend school functions for their kids, and community functions during work days; unlimited vacation; and a new family leave policy that provides six months paid family leave and an additional six-month part-time option.
Among the many other perks and activities we use to motivate our employees: volunteer opportunities in the community; a weekly happy hour with a dedicated bartender; a fully equipped on-site gym; ping-pong, pool and foosball tables; free lunch daily, and more.
Our office culture supports creativity, collaboration, and healthy fun — and I think we’ve done a great job: in January, we were awarded 13th place in Best Workplaces for Technology by Fortune, and #37 in Great Place to Work’s Best Workplaces for Women. In November 2016, ZestFinance was named to Forbes Fintech 50 list and was named one of the LA Business Journal’s Fastest Growing Companies.
Try to find smart, interesting people to learn from, and create relationships with them. Mentors and coaches, whether formal or informal, are key to helping you grow. Also, read a lot, across many subject areas.