How machine learning can improve SME financing

(Image credit: Image Credit: Razum / Shutterstock)

Machine learning is increasingly the technology of choice to streamline internal business processes, improve data analysis, and even source job candidates. The power and value of machine learning in the financial industry is becoming more evident every day. A KPMG report indicated that in the next 15 years, robots could perform up to 75 per cent of financial service jobs. This will have a significant impact on hiring and operational costs. However, it may also allow financial institutions to put more time and effort in other areas of their businesses.

Machine learning involves computer programs that analyse data and learn from it in order to draw insights and make predictions. Machine learning solutions have the ability to learn without being explicitly programmed, and the more data there is to feed to the models, the more accurate the results are.

Financial institutions are home to immense amounts of data, and therefore, there are many opportunities that machine learning can help banks uncover to stay on top of the fast-changing financial landscape. Though many financial institutions already use machine learning to perform data analysis and improve their internal operations, the technology also provides unique opportunities to overcome the many challenges of lending to SMEs and meet their increasing financial needs.

Current SME lending challenges

Financial institutions are often hesitant to provide financing to SMEs with limited or no credit history. For businesses that lack a credit history, it can be difficult to prove their ability to repay a loan, and therefore, they are denied the opportunity even to try. Traditional financial institutions that limit their services are also putting themselves up against a growing number of industry challenges, including:

Small and medium-sized enterprises (SMEs) represent more than 95 per cent of registered businesses worldwide. SMEs also account for more than 60 per cent of jobs and in many emerging markets, contribute up to 40 per cent of Gross Domestic Product (GDP).

The success of SMEs is important for economic growth, especially in emerging markets. Despite all of the benefits SMEs provide, they remain significantly underserved by financial institutions. The difficulty in obtaining financing is one of the top three constraints for SMEs to survive and thrive. In many emerging markets, it is frequently the top constraint.

 

  • Growing competition from fintech startups - Financial technology startups are disrupting every sector, from payment collection and processing to borrowing and lending.
  • Increasing regulatory and compliance costs - Dedicating more resources to compliance can mean less time and resources for customer support and/or improving the customer experience.
  • Cybersecurity and consumer digital privacy concerns - Financial institutions must rethink and improve their customer data security practices.
  • Evolving consumer expectations - Financial institutions that require unnecessary in-person visits or excessive documentation instead of digital experiences and the convenience of at-home banking are at risk of becoming irrelevant.

Whether it’s increased competition or rising consumer expectations, machine learning can help financial institutions overcome all of these challenges and serve more SMEs.

Identifying more creditworthy SMEs

Traditionally, banks have an advantage when it comes to determining which applicants to approve for a loan because they have years of historical data on which they can base their decisions. However, these traditional lending models take into account a very narrow range of data points to form a decision.

Machine learning allows lenders to analyse additional information and paint a more accurate picture of an SME’s creditworthiness. Everything from a business owner’s personal credit history and tax payment history to social media and product shipping activity can go into a machine learning model for review. By evaluating more comprehensive data, and in real-time, lenders can potentially expand their customer base to include more “thin-file” SMEs in a less risky and more effective way.

Speeding up the loan process

Respondents to a Small Business Credit Survey by the Federal Reserve Banks overwhelmingly ranked long waits for credit decisions as a top reason they were dissatisfied with lenders. It can take banks anywhere from a few days to a few weeks to respond to a request for a loan. After the initial response, it can take even longer to complete the approval process and deliver the capital for business use.

Machine learning models can streamline the entire loan process, dramatically decreasing the time lenders spend on manual loan processing tasks. The result is not only a reduction in the total time it takes to provide SMEs with a loan but lower costs and potentially bigger profits for lenders who can serve customers faster as well.

Monitoring loans and payments

Financial institutions hold enormous amounts of data about their customers. Machine learning provides new ways to analyse that data and monitor loan payments and behaviours in order to flag borrowers that show signs of trouble. Lenders can automate payment reminders and step in before a business defaults on a loan. Machine learning can help lenders identify issues and resolve them before they happen, ensuring that their SME lending activities remain profitable.

Recommending loan opportunities

SMEs require different financing solutions depending on their industry and individual needs. Therefore, it can be very beneficial for financial institutions to provide a wide range of lending services and to recommend those services at the right time to the right customer.

Machine learning programs can analyse large amounts of data to make recommendations regarding which lending services are best for each type of business. These recommendations may be based on factors such as the business’ location, industry, transaction sizes, cash flow, etc. By monitoring customers’ behaviours, machine learning models can also predict the best times to approach a growing business with new lending services.

The future of SME financing

The current credit gap for formal SMEs hovers around $1.2 trillion. For both formal and informal SMEs, that number is estimated to be as high as $2.6 trillion. Improving access to finance for SMEs is crucial for economic growth, especially in developing markets. Thankfully, technologies like machine learning are enabling financial institutions to unlock new financing opportunities for SMEs and support their growth. Machine learning isn’t the solution for every challenge traditional lenders face to serve more SMEs; however, incorporating it into the financing process is a significant step toward ensuring traditional banks continue to evolve and better serve these key contributors to the economy.

Diego Caicedo, Co-Founder and CEO, OmniBnk
Image Credit: Razum / Shutterstock