Application of big data and technology in small business credit in US

Source: IFA

Author: Sanjay Phadke
EVP – Platform & FI Business Head

The United States, apart from being the largest economy, is also home to one of the 3 largest small business populations, the other 2 being China and India. As economy switched gears from agrarian to industry to services over the last few centuries, the number of small businesses have seen a healthy growth. However, the access to capital for small businesses, the lifeline for any business, has not kept pace, especially in the last 10 years post crisis.

This comes paradoxically at a time when liquidity with banks is plenty, and total money supply has doubled over the last decade. The explanation to this puzzle lies in increasing regulation and banking consolidation making the economics of administration of small business loans unattractive. Thus, while debt is hard to come by for Small and Medium Businesses (SMBs), or is quite expensive at 10-50% APRs , the US also does not have a small business listing platform where equity capital can be generated for growth like bigger exchanges for big companies. This is a big issue leading to persistence of inequality, which by some measure has gone up a few times in the last few decades. Technology, with its new business models, was supposed to be the panacea for all problems, including this. This article summarizes the challenges and potential solutions using technology for improving small business lending access and cost.

Fundamentally, in services business like financial services, technology helps by digitizing, standardizing and then automatically processing vast quantums of data a lot more efficiently than humans. The problem with technology adoption by SMBs is the diversity of small businesses which on one hand include a mom n pop shop selling candies or providing laundry services in a small town to a small enterprise vendor in a supply chain of a large business making automobiles or providing communication services in a mid-sized town. The diversity of business types also spawns differences in familiarity with technology such as accounting system, or online order management, or receiving or making payments online enabling automatic tracking. Each different organization means a different data type and needs a different way of dealing with it. The myriad differences mean that large standardized IT systems are not useful and are anyway too expensive. The solution for SMBs is to effectively use technology, like cloud-based pay as you go services with an ability to align with the variety of underlying businesses, and their unique forms and data.

Thus, the problem of small biz financing can be solved for one class of SMBs; the supply chain ones. According to Harvard Business Review, there are 1 million such enterprises. These are important employment generators, providing up to 500 jobs each. The advantage with this group of SMBs is that they would generally use at least basic technology in terms of billing and linked payments. For them, paying for their operational costs is probably easiest done by a cloud-based SaaS service which captures their invoices and provides early payment to the counterparty. This is a very safe credit risk, especially if the trade has been going on for long across cycles, evidencing a strength of ongoing supply chain relationship. Such an approach does not need detailed financial or balance sheet analysis as much as a quick check on industry dynamics. This would take care of Payables/ Receivables and would probably be a US$ 100 Billion market by just taking care of early payment or collection of short term cash flows, essentially revenue items. As for financing for capital expenditure, automated credit approval in the US is particularly difficult as there is no filing required by private companies. That makes the process of risk analysis difficult, and as such growth capital is hard to come by for SMBs, unless they pledge collateral to Asset Backed Lenders, or provide a portion of revenues like MCA loans. The gap in information and trust here makes these loans seriously higher priced, making it nearly impossible for SMBs to scale. This can be addressed by getting access to tax filing, with IRS making APIs available to enable faster information dissemination leading to quicker decisioning. Another approach is creating a plug into the accounting system like Quickbooks or Xero to run a credit score engine, but as the latest experience with social media shows, this needs to be tightly managed in terms of privacy and unauthorised sharing of data.

The other issue for running serious SMB loan portfolios is the high cost of customer acquisition. Historically, it was solved by feet on street, by way of branches of community banks or network of intermediaries. Later, tele-calling came up, and then digital origination; but both are not as effective for enterprise customers unlike with retail consumers. In any case, the cost of running digital campaigns and search optimizations etc is not trivial, especially if the value of every conversion is not large; which would be the case for SMBs, where the maximum requirement is small ticket loans. One useful feature here is the referrals, which also improves the honesty of the credit ecosystem. The last challenge for SMB access to capital, apart from information insufficiency, cost, and access is the ease of use. The clunky legacy technology that is prevalent in large organizations would not work for SMBs. They would need easier web or mobile interfaces, or if possible, the entire transaction flow happening automatically, pretty much in an invisible fashion. The user experience for SMBs needs to be almost like the ease of switching on a light bulb, or the ease with which one can swipe a credit card; oblivious to the myriad processes that happen in the background. They don’t have the time, patience, resources and expertise to fill detailed forms, especially on a repetitive basis. Thus, the use of one time on boarding, use of emails, etc. are the hallmarks of effective technology uses, apart from other factors mentioned above.

Once the SMBs’ capital gap issue is solved by way of reasonable cost of capital in a digital doorstep fashion, all of it thanks to smart yet invisible technology, the next piece would be the use of big data and AI to make this even more efficient. For those techniques to work, we first need enough data for training algorithms; which maybe a simultaneous development along with cracking the SMB capital puzzle.