According to some estimates1, the world generates 2.5 quintillion bytes – or one billion gigabytes – of data per day. With 90 percent of all data created and stored in the world today having been generated in the last two years alone we are seeing a tremendous increase in the volume and richness of data. The proliferation of technology in everyday life, combined with cheap widely available commodity hardware, highly accurate data measurement and storage technologies. The world steadily becoming more connected with an ever-increasing number of electronic devices will result in continuous growth of data generation and storage in the future.

The mantra for many companies today is “mass customization” which may seem like an oxymoron. However, these massive repositories of data are a major asset for companies that can tease out actionable insights from them, enabling them to customize product and service delivery on a mass scale. We see examples of these in our daily lives today – whether it is Amazon tracking customer’s search and buying habits to customize offers or a Comcast assessing a user’s viewing patterns and preferences to suggest content. Banks and specialty lenders similarly, have access to vast amounts of customer data that they collect such as customer transactions, account balances, and personal preferences and spending habits of their customers.

Extracting actionable insights from these large datasets to better serve customers, involves the use of highly complex data analytics programs that can comb through vast amounts of data in seconds. These data analytics programs can work around the clock, are very accurate and need little maintenance once they are developed – traits that make them ideally suited to benefit banks and specialty lenders.

Financial institutions recognize the potential that data analytics represents as is illustrated by the following2:

  • JPMorgan Chase recently invested in analytics programs that can analyze legal documents and extract important data and clauses in a few seconds as compared to over 360,000 man-hours taken to analyze the same documents manually.
  • Bank of America is using chatbots with predictive analytics and cognitive messaging to interact with and engage the company’s 45 million customers.
  • Bank of New York Mellon Corp. is using automated bots that run on highly sophisticated analytics programs for everything from correcting formatting to handling customer requests, with 100 percent accuracy, an 88 percent improvement in processing time, and over $300,000 in annual savings.

Many of America’s banks and specialty lenders have invested heavily in data analytics to streamline operations, cut costs, improve processing times and more effectively serve customers. It is clear that the future of finance will continue to be heavily influenced by emerging financial technology and analytics applications.

However, when it comes to Mid-Tier banks and lenders, the data seems to indicate that the adoption of sophisticated analytics solutions and machine learning applications have been impeded by the high cost of these solutions and the scarcity of skilled workers who can develop and implement such systems.

Mid-Tier banks and lenders are faced with the same challenges that larger institutions like the ones profiled above, are trying to address. Below are some of the ways data analytics solutions are changing the way banks and lenders operate3.

Improving commercial and corporate banking

Corporate clients expect premium finance, risk, payment and working capital service on a global scale and around the clock. They also expect fast and transparent access to these services. Inconsistencies in sales, client onboarding, product delivery, and service management limit the ability of banks to efficiently cater to their customers’ needs. Data analytics can be used to design a single, seamless, and interconnected corporate banking experience across all relationships and services, including trade finance, supply chain, lending, cash management, and treasury operations.  This experience will help boost sales while providing consistency across products and pricing, improving the agility of your departments, and deepening client relationships.

Deposits and product personalization

Insights extracted from historical customer data can be used to generate general guidelines on what kinds of products and returns suit specific types of customers. By offering a high level of personalization along with rates that meet or beat customer expectations, banks can grow deposits as well as their customer base and their portfolio of products and services. All it takes is a simple questionnaire to arrive at the right package for the right person.

Consumer lending & credit cards

One of the most time-consuming of all banking activities is conducting customer background checks. Data analysis programs can be designed to conduct such checks and assess individuals as well as businesses for creditworthiness and grant approvals or issue rejections for loans based on a set of customizable parameters such as credit score, ability and willingness to pay,  education level, and occupation – all within a few clicks.

The benefits of data analytics in banking reach far beyond improving products and growing deposits. Data analytics can be used to redesign and improve internal banking operations as well.  Here are a few examples:


Data analytics can be used to comb through vast amounts of data very quickly.  Automated analytics programs that run around the clock and can work autonomously are used extensively in anti-fraud and banking compliance. Complex machine learning systems that are constantly monitoring behavior can identify and reduce fraud. By focusing on the real-time and most critical fraud issues, teams will likely see fraud rates diminish.

Chatbots and advice bots

A number of banks have launched automated advice programs that use data analytics to accurately assign their customers and their desired portfolio a risk category within minutes.  Automating this process saves the bank the cost of hiring individual fund managers for each client, not to mention the time it takes to assess and categorize individual customers and portfolios, thereby lowering the cost of investing, enabling the bank to attract more customers.

The functions of these “robo-advisers” extend beyond portfolio and risk assessment since they can also be used to give customers portfolio support and advice in place of human customer support reps, thus further reducing operational costs.

Improving customer service

Data can be intelligently used to improve customer hold and servicing times by evaluating the costs, revenue, and inefficiencies of specific approaches to customer care. For example, data on ATM queuing times versus customer booth wait times can be compared across different cities, times, and banking departments, giving insight into things like customer drop-off rates and the average value of a walk-in versus a regular customer for each mode of contact. Having insight into what worked in the past and what current trends are today can help banks minimize operational risks, develop new strategies, and plan where to place technology in the future.


In December 2019, the Financial Account Standards Board (FASB) will implement the largest change to financial reporting in decades. The new standards require periodic estimates of lifetime expected credit losses for all financial assets that are within its scope. The change affects how banks calculate their reserve requirements (ALLL) for their lending and leasing portfolios.  The Federal Reserve defines the Allowance for Loan and Lease Losses (ALLL) as:

Estimated credit losses are estimates of the current amount of loans that are probable that the bank will be unable to collect given the facts and circumstances since the evaluation date (generally the balance sheet date). That is, estimated credit losses represent net charge-offs that are likely to be realized for a loan or group of loans as of the evaluation date4.

This is an accounting requirement for all banks no matter the size. This will require a large effort to gather the relevant data and build loss forecasting models across the all loan and leasing portfolios.

So, how can smaller banks and lenders avail themselves of the capabilities to address the above challenges? Fortunately, new paradigms are evolving that preclude the need for institutions to invest heavily in large teams of data scientists and technology. There are two models that particularly relevant – Data Science as a Service; and Industry-Academia Partnerships.

Data Science as a Service, is a managed service where a third-party takes on the model design, development, validation and execution on behalf of a bank or lender. This can be a more cost-effective way for these institutions to leverage these solutions than developing their own capabilities in-house.

The Industry-Academia partnership model is one where a financial institution partners with an academic institution with deep data-science expertise, like Temple University. Academics are willing to address complex modelling challenges in exchange for being able to publish research papers based on relevant findings from the project.


All banks and specialty lenders are looking to serve their customers in a more specialized and targeted way while taking care of operational realities such as fraud and regulatory changes. Data Science and Analytics are key to addressing these challenges effectively.

While these solutions can have significant, measurable payback, affordability remains an issue. Newer commercial models are making these sophisticated capabilities more accessible to smaller institutions.

DecisivEdge helps businesses find, develop and sustain competitive advantage using data-driven analytics, process optimization and strategy consulting. Talk to us about how our Data Science team is helping banks and specialty lenders address their challenges. Read more about our Data Science as a Service offering.

If you enjoyed this article, you might also like this – ‘Data Science As A Service’ Is Almost Here, Making It Even More Important To Understand Data