Improved Digital Banking through Data Analytics

Financial institutions can opt to digital banking as it is cheaper than keeping brick-and-mortar branches. Legacy banks have been shrinking branch networks in the US for over a decade. The COVID-19 epidemic slowed down in-person services, and encouraged the shift to digital banking. Customers can access their accounts, products and services via a website or mobile app.

It’s not surprising that banks are spending more on technology. Improving customer experience and service delivery are their top priorities. Increased spending on customer service centers has meant that financial institutions are not reaping the greatest return on their digital investment.

A 2020 survey from the management consulting firm Capital Performance Group found that, from the end of 2019 to December 2020, online banking activity–including transactions and other interactions–increased as much as 30% and mobile banking activity surged as much as 80%. It also showed that some contact center volumes increased by as much as 50% at certain institutions, and that mobile banking activity rose as high as 80%. Cornerstone Advisors conducted a 2021 survey which showed that customers call their banks more often because they cannot find the answers online or because their financial institutions aren’t providing the support necessary to solve their problems.

These statistics highlight the unfortunate truth that many of the digital products and services banks have been offering for years–from payment systems like Zelle to authentication–still fail to meet customer expectations, often because using them doesn’t feel intuitive enough. Zelle has also been plagued with disputes, particularly regarding fraud and unauthorized transactions.

As consultant to severallargest commercial banks in the United States I have seen firsthand the challenges of digital transformation. Although banks have a lot of data that can help them to develop strategies for customer activation and retention, they are not using it enough. The data analytics use of most institutions is not robust enough to capture the necessary depth and breadth to understand customers’ needs and to determine how best to serve them. One bank I know of uses very small sample sizes for recording and listening to calls regarding quality assurance-11%. Extrapolating findings can lead to misleading results.

These problems can only be addressed if banks develop more comprehensive, holistic customer data analytics at a larger scale, including all phone calls. They must then use the patterns of customer behavior to help create and enhance digital functionality that meets customers’ needs. This article will show you the steps I take in helping banks achieve this goal.

Create a Strategy Team

To increase consumer use of digital banking, and improve customer satisfaction, the bank should establish an internal team made up of product experts and analysts. This will allow them to access all data across all channels and products. Because there is too much data to analyze, study and draw conclusions from, this team is vital. To ensure that its findings are implemented at the enterprise level, this group must actively collaborate with department heads.

This team should be divided into smaller, cross-functional teams for each product. My work with Commercial Bank required me to tell each team how many customer calls it could reduce by implementing certain functions or features, so that it could prioritize its work. To determine the potential reduction, I used a proprietary customer-journey analytics platform that we had developed to analyze user flows and identify friction points. You can also use Google Analytics and Tealeaf for the same purpose.

Identify Data Categories and Set Your Goals

The second step is to identify and gain access to disparate data sources across all platforms and functions. Data sources in a typical legacy bank can be divided into two categories and several subcategories. These were the data sources I used to set up teams at banks.

Products and Businesses

• Retail, such checking and savings accounts
• Credit cards
• Mortgages
• Finance for automobiles
• Wealth Management

Contact

• Telephone calls to the contact center
• Contact center interactive voice response (IVR) communications
• Retail branches offer in-person interaction
• ATM transactions
• Desktop application
• Mobile application
• Notifications/alerts for outbound notifications

My analytical work focused mainly on call statistics at contact centres. This is the most important area I focus on. This is where the vast majority of support requests are made. Banks don’t accept email customer service inquiries anymore, even if done offshore. Chat has replaced email, but it accounts for a very small percentage of customer service interactions at top banks – less than 5% in the case I was able to see. Live agent calls can be very costly because of the volume of requests. This is where I will focus as I explain my process.

After data sources have been identified and accessed, financial institutions can start defining key measurable goals that will help define the project’s scope. This can be used to set the stage for their problem-solving strategies. These are the objectives I established at the banks where we consulted:

• Customer experience can be improved – measured by Net Promor Score (NPS), which is a key indicator for customer satisfaction and assesses people’s likelihood of recommending a company across all channels.
• Digital adoption and engagement should be increased
• Calls to contact centers should be reduced if they don’t add much or any value
• Branch interactions with basic banks at low margins are reduced
• Service operations can be improved while risk reductions can be reduced.

The data revealed that digital was the most popular channel for customer interactions, as expected. Surprisingly though, I found that highly active digital customers were more likely to seek help than less active traditional and digital banking users at the banks where they worked. The number of inquiries and calls to contact centers for digital banking was more than double that of traditional banking.

Understanding Why Customers Call

After identifying data goals and data categories, the team must consider the types of queries that the bank’s data can help it evaluate the nature and circumstances for customer support requests. The banks I worked with focused on contact centers and customer interactions within a specific time frame. We then came up with these questions:

• How many customers have spoken with a live agent?
• These callers were who? What were their service profiles, such as how they interacted with customers across channels and products, and what was their enterprise-level customer value?
• How many of these callers were digitally active
• What banking activity, if any? Has it occurred before the call?
• Which channel broadcasted a bank activity that occurred before the call?
• What were the customers talking about?
• Are they calling you more than once? Did they call more than once?
• What was the duration of the calls?
• How long did it take for customers to make multiple calls?

Although banks often keep track of the number of calls received by their contact centers, but they only look into a few ancillary statistics. This is consistent with the financial service experience. While banks do track calls, they don’t usually look at ancillary statistics. Call center workers would benefit from knowing that a customer spent 20 minutes trying to resolve a dispute or activating cash-back rewards.

My strategy teams and I collaborated with banks I advised to help them document every call through a system that records the purpose. We could see that a customer had tried unsuccessfully to close their accounts online minutes before calling us. This was the reason we called. Each call was assigned a label indicating its purpose and a date stamp. By identifying the events occurring around the primary catalyst, we were able to identify secondary and tertiary causes for each call. This allowed us to create a complete picture.

The call-to-contact spread was a key metric that we used to evaluate the experience of all callers. It was also used as a benchmark to improve efficiency.

• Call Rate: The total number of calls made as a percentage of the entire customer base
• Contact rate :The number of customers who make calls to you, expressed as a percentage of all customers
• Call to Contact Spread: The call rate is less than the contact rate

We set out to lower both call and contact rates. They were higher than industry benchmarks for top banks. These benchmarks are usually around 20% and 10% for a month. Based on my experience and reports by third-party benchmarking companies like Finalta, and McKinsey, our first goal was to reduce them. We wanted to equalize the metrics, which meant that we would eliminate repeat calls. We had achieved an initial call resolution. Consumers only need to make one call to resolve their problems. This is a critical benchmark in customer service management.

We also looked at other metrics and hoped to reduce them.

• Call durations
• Transfers by phone
• Escalations
• Complaints

While most contact centers use call reasons from customer relationship management systems or call recordings to identify callers, very few use systems of record to improve these metrics. The method of record’s upstream conditions can help eliminate calls or match callers with suitable agents based on their previous activities, customer profiles, and the level of service required.

Many financial institutions use customer experience management software, which surveys customers after each interaction and produces a Net Promoter Score. The banks I worked for set a goal to increase the score of contact centers to 55%. We achieved it.

To create profiles and analyze call patterns.

Next, analyze call patterns using different systems of recorded data points to determine what drives customers to phone. We used extensive customer-level data sets where I was consulting:

• Monetary transactions include large point-of-sale (POS) fees and payments or reversals.
• Transactions that are not monetary, such as card declines, address changes, disputes, etc.
• Channel interaction events include phone calls, IVR communications, and desktop activity. Branch visits, ATM transactions, and outbound alerts are all examples of channel interactions.
• Customer profile/segmentation, such as tenure, high-value flag (indicating a high level of engagement), and several authorized users.
• The active vs. inactive product hold of customers, i.e., the number of products each customer owns and whether they use them. This indicates how likely they will interact with the bank to receive service. Note: We focused on active customers to measure call and contact rates.

These categories allowed us to extract key data elements that were easily accessible and helped us better understand the following:

• Channel confinement: Customer propensity to stay within the channel for short periods, usually 15 to 20 minutes
• Self-service channel activity: What the customer did to self serve, via digital, ATM or IVR
• Channel preference or mode: primary channel the customer used

After analysing these upstream conditions, our focus shifted to the downstream effects. We measured and analysed what was happening with agent calls. We excluded calls lasting less than 60 seconds. Instead, we looked at call patterns and types. We also calculated time ranges and identified how calls were made. We also compiled the data to see which channels customers were banking through.