Machine learning. The futuristic term sounds like something from the Matrix or Terminator movies. As it turns out, the future is now. Machine learning is already at work, improving our lives in numerous ways.
What is it? Machine learning is a field of computer science where computers analyze data and learn from it in order to make more accurate predictions about the future. For instance, Netflix uses machine learning to examine what movies you’ve watched in order to suggest other films you might like. Amazon’s recommendation engine operates along similar lines, suggesting products you might want to consider based on what you’ve already purchased or searched for. Likewise, machine learning is behind LinkedIn’s “People You May Know” tool.
In all cases, the machine “learns” based on a constant feedback loop. If you watch the recommended movies and rate them highly, that counts as a successful prediction and informs future predictions. If you don’t like the movies, however, the algorithm learns from that mistake and refines its next wave of predictions. Multiply that effort by thousands of other shoppers, datapoints and patterns they generate, and the entire system is designed to increase accuracy exponentially over time.
There’s much more to machine learning than just bettering our entertainment selections and social media interactions. Traditional lenders can now use machine learning to improve their lending process and profits dramatically. How so? Read on.
Machine learning can help banks identify more creditworthy borrowers.
The reputation of small businesses: Not creditworthy. If you’re judging SMBs on their PAYDEX scores alone (which dropped from 53.4 to 44.7 during the Recession), this is technically a correct way of thinking. But how true is that really now? Traditional lending models only take a narrow range of data points into consideration. Machine learning has the capability to review much more data and review many more scenarios which can inform a widening of criteria to include additional information like Yelp scores, social media activity, and real-time shipping trends — creating a more accurate, nuanced portrait of creditworthiness. More enhanced, real-time data can help lenders evaluate “gray area” loans more effectively and expand the pool of creditworthy borrowers with limited risk.
Machine learning speeds up the loan application process.
One of the top complaints of small businesses seeking credit? Long waits for credit decisions, according to a 2015 Cleveland Federal Reserve survey (PDF). A traditional bank needs 3 to 6 days to respond to a small business borrower’s loan request. But from there, it takes even more time to determine whether the company is a good credit risk and then even longer before the owner gets cash in hand. A high-tech model that uses machine learning reduces the total amount of time — from loan application, to approval and eventually gaining access to capital — introducing dramatic cost efficiencies that improve profitability. (See our post Quantifying Speed’s Value in Small Business Lending for more information.)
Machine learning improves on-going loan monitoring.
Banks generate a large amount of data on their customers. But without an advanced system to analyze it, important knowledge goes unknown. Pinpointing early signs of a loan in trouble and stepping in before a borrower defaults can help make SMB lending more profitable. Likewise, identifying “non-traditional” borrowers that banks aren’t lending to, but should consider, can help them grow their business.
Machine learning makes it easy to adapt your credit policy.
Perhaps the best aspect of machine learning isn’t big-data analysis, but rather, the smart capabilities of its algorithms. Once banks establish credit policies, they can use machine learning to quickly determine which borrowers meet that criteria. Then, machine learning can scrutinize those existing loans to see how they’re performing and fine-tune the bank’s approval criteria if need be.
Digital marketplaces are already capitalizing on the benefits of machine learning, and borrowers are responding by opting to borrow from these alternative lenders instead of traditional financial institutions. (Morgan Stanley reports that digital marketplaces have doubled their loan origination every year since 2010, most recently lending $12 billion in 2014.) To offer customers the 21st century lending experience they’re looking for, traditional banks should incorporate machine lending into their existing small business lending. Are you ready to embrace the future of lending — machine learning — today?