10 Ways Mastering Customer Data Fuels Digital Transformations At Banks
Guide to Mastering Customer Data for Corporate & Commercial Banking
Insights from industry leaders such as Scotiabank, Capital One and Santander
Despite significant investment in core operational systems (FIS, Fiserv, Salesforce, and SAP) and analytics teams hired to put to work the data that’s flowing into the systems, most financial institutions still struggle to gain a true 360-degree view of their customers. Without a holistic understanding of their customers and their activities, banks are missing upsell opportunities, putting themselves at risk of fines from regulatory bodies, and opportunities to delight customers and foster greater loyalty.
With many in the industry investing in a cloud computing strategy and executing digital customer strategies, the time to tackle messy customer data is now.
We’ll discuss how leaders can fix the fragmented customer experience caused by bad data. We’ll show how a machine learning-driven, cloud-native approach to data mastering unlocks complete customer views across business units, countries, and corporate hierarchy. And, we’ll highlight ten tactics industry leaders have used to master customer data and accelerate their digital transformation.
What Customer Mastering Means
Customer mastering achieves true customer 360 - a unified, accurate, enriched view of customer data across systems and sources that can feed both operational and analytical systems, from core banking operating systems, CRMs, ERPs, and marketing automation tools to business intelligence tools, dashboards and Excel.
10 Ways Customer Data Mastering Helps Successful Digital Transformation
Leverage data across departments to drive cross-sell and upsell:
Financial services is a “high touch & high tech” industry, all institutions want to provide a full range of services to their customers/clients. The benefits of producing services across business units include a higher retention rate, and more cross-sell and upsell opportunities (sometimes called the revenue multiplier effects). Reliance on sales to ‘know their customer’ and remember engagement patterns becomes unmanageable at scale. Reaching out with the right product recommendation at the right time is essential to driving account expansion. By enabling a holistic, clean view of an account across business units, product recommendations can be data-driven and automated to drive expansion.
Enrich customer data for faster customer onboarding and KYC:
By deploying a modern approach to data mastering across their huge customer data franchises, organizations can save significant amounts of time for both the business users of the data and the data management staff responsible for delivering the data in usable form, resulting in shorter KYC and customer onboarding times, which meant faster revenue generation and improved customer experience. At the same time, enriching new lead information ensures that Sales is armed with the correct information to act quickly. It provides 360-visibility on existing business with the corporate, and accurate lead contact information for outreach. _Pricing strategy within the industry is complicated. By unifying and tracking pricing information across corporate accounts, customer data mastering enables a centralized view of pricing. Instead of being merely reactive to quote requests, operations can monitor pricing by market segment, business unit and customer segment to proactively engage sales and pricing teams on commercial terms to ensure profitability is prioritized. Without a data-driven pricing strategy, sales reps can go rogue to ‘sell at any cost’ which may not result in a desirable level of profitability. Marginal improvements in pricing can translate to millions of dollars of revenue growth as the strategy scales across accounts.
Automate data pipeline to increase team productivity:
Data preparation, cleaning and categorization are time-consuming processes when tackled with rules-based approaches (“if x=y, then”) or sheer manual effort. The work is typically spread across teams - while data scientists and engineers often dedicate the most time to it, data consumers within the business often manually adjust the data given low data quality. This extra work can range from a sales rep wasting precious hours trying to reconcile data in Salesforce to a financial forecasting team manual merging data sets in excel. Through a machine learning approach to mastering customer data, manual effort can be reduced by 90%. Forrester estimated that the average company saves $4.9M over three years by implementing machine learning-driven data mastering.
Create a single source of truth for compliance and reporting:
The benefits of a single view of the customer extend beyond improving sales and marketing performance to another critical area of the business: risk mitigation and compliance. With the ability to continually clean, cleanse, and consolidate customer data to maintain a single source of truth, financial organizations can improve their know-your-customer (KYC) workflow, operate with greater confidence in the accuracy of their reporting, and reduce the risk of violating compliance and running foul with governing bodies. Making sure the data can be relied upon and can be easily accessible is crucial for regulatory reporting, internal auditing, and budget forecasting.
Deliver better digital experience to achieve better services and retention:
Customers’ appetite for digital banking services is a longstanding industry trend. According to a recent survey, 89% of US respondents say they use mobile banking channels, and 70% say mobile banking has become the primary way to access their accounts. As more commerce is moving to self-serve/online, having a platform that focuses on the end-to-end customer experience will help you stand out from the noise. By providing a holistic digital view of accounts, companies can guide recommendations, ensure fulfillment and create a path for renewal and expansion, which some industry leaders referred to as “optimizing the digital shelf”.
Develop a first-party data strategy for a personalized experience across digital channels:
As digital marketing shifts from third-party identifiers toward a privacy-driven approach, marketers must adapt and invest more in first-party data. This is especially important in the banking industry, where customer data is treated differently based on its’ sensitivity. It’s data that you own and collect with direct consent from customers, through interactions on apps and websites, and in response to marketing initiatives, like email and loyalty programs. When used responsibly and efficiently, first-party data can help companies build direct relationships with their customers, create value, and boost marketing performance.
Have a hierarchy view of customer accounts to improve operation efficiency:
Given the complexity of customer relationships across business units, countries, and corporate hierarchy, it’s critical to focus the go-to-market (GTM) budget on outreach strategies that target decision-making business units. Unifying customer data based on location and parent hierarchy gives marketing a richer view of customer relationships so that marketing budget isn’t wasted on multiple campaigns to the same companies, bounce rate is reduced and promotions are targeted to the right business level. A clearer understanding of customer hierarchy will also help with financial and operational risk management.
Avoid and resolve duplicate records to optimize marketing across key accounts:
By creating a mastered data view of customers, relationship managers and sales reps are armed with accurate account information across CRMs and analytical tools to target customer conversations and avoid duplicate outreach efforts. Sales leaders today are focused on reducing all non-customer engagement activities so that account representatives can focus on what matters - the customers. According to research by McKinsey, the typical salesforce only spends 26% of its time on customer-facing tasks such as meetings, calls, and account strategy.
Focus on the quality of data being consumed more than how it’s being created:
A modern approach to enterprise data mastering assumes imperfect source data and acts accordingly. This is similar to Google’s search indexing infrastructure which assumes that all data on the web is imperfect. In other words, data consumption is what truly matters and it sets the right context for enterprise data mastering. Over time–as data consumers validate the best data as a data organization–you can then go back to remediate the original sources’ data. By applying machine learning strategically to enterprise data mastering, you can create trustworthy, automated models for the mastered entities that matter to the business: customers, products, employees, suppliers, and so on.
Articulate business value and ROI from improved data infrastructure:
According to Gartner, the number one reason Master Data Management (MDM) projects fail is the lack of a structured framework to qualify and quantify data management value creation. Data leaders not only need to be able to identify and solve the technical problem, but also need to be able to articulate the business problem and link the two. Without a strong business case, potential high impact projects can be delayed (or killed) for the lack of support from executives.
Key Challenges Impacting the Quality of
Customer Data in Financial Services
- Large amount of data
In the financial services industry, data volume is often higher than any other industry, with increases in data attributes generated in transactions, and relationships between different entities. On top of that it is also difficult to keep up with changing reporting requirements. Overtime, many data teams attempted piecemeal solutions plus one-off cleaning to try to meet deadlines, which in turn created data governance nightmares and it would take immense manual efforts to clean up.
Many financial institutions continue to struggle with varying degrees of quality and consistency across customer systems. Typical errors include misspelled organization names, wrong contact information, and unintentional duplicate records within databases. In some cases, it can be as extreme as contact names or address information found in fields meant for the company name. Inconsistent data schema and metadata can fuel misunderstanding of what data was intended to be captured. The global nature of many within the industry also leads to language standardization and translation challenges.
Siloed data instances
Because of data sovereignty and financial regulation requirements in different markets, financial institutions had to keep separate data instances. For them to do financial or risk reporting, data teams need to write and maintain data integration rules. The number of rules can grow to wildly unmanageable and team knowledge isn’t always institutionalized when members leave the organization. As a result, rule-based systems quickly become prohibitively expensive (both monetarily and time) to maintain.
Apple Computers vs. Apple, Inc. vs. Beats Electronics
One of the most common, highly visible challenges for any bank is that customer names are often spelled differently or are subsidiaries: Apple Computers vs. Apple, Inc. vs. Beats Electronics - the list can feel endless and accommodating name variations or misspellings via rules (if [nameX] = [nameY] then..) becomes unmanageable. Mastering customer data requires using often imperfect information and judgment to decide if two customer entities are connected, which is why a machine learning approach is vital.
How Machine Learning Delivers Better Customer Data
To achieve high accuracy and carry out the mastering process efficiently, a machine learning approach is needed. The backbone of most traditional approaches to master data management are rules (at the most basic level, ‘if-then’ type statements). Rules do not scale and, in our example, struggle to allow for the nuanced relationship levels of our Apple example. Rules require consistent high manual effort from creation to maintenance and become a complicated web to untangle for any data team once scale is reached. By leveraging a machine learning approach, it allows customer data to be mastered with a fraction of the manual effort, while maintaining team input and influence on how the data should look. Poor underlying data quality can mean piecing together partial information like half-completed address fields - a task built for a machine learning approach.
Three Important Technical Features
for True Customer Mastering
Case Studies: Real-world Impacts of Customer Data Mastering
Below are two examples where customer data mastering solved long-standing data challenges to make a huge impact on their customer engagements.
Santander Cuts Lending Time in Half with Customer Data Mastering
With more than 14.4 million customers and £22.3 billion in corporate loans, Santander UK was able to cut loan processing time in half by powering a single view of customers from its 45 data sources. The data available to, and required by Santander UK for its lending practices had exploded in recent years as a result of organic growth, company acquisitions, and added information required by new regulatory and due-diligence requirements. To capitalize on their wealth of data, the bank embarked on an ambitious digital-transformation plan headlined by a new lending system connecting modern digital processes, data, and SaaS applications in the cloud with APIs.
Santander UK deposits customer records from 45 data sources across these 20+ systems into a data lake. Tamr then uses REST APIs to connect with the relevant data sets and identify potential duplicate records. To keep the project in line, and avoid data drift, the bank utilizes confidence thresholds for the machine learning models. If a model went past the set threshold, an intuitive feedback workflow would start to engage subject matter experts in the bank to correct the data to fine-tune and improve the ongoing performance of the model. The resulting mastered data is returned to the data lake and processed by the credit lending system.
Using Tamr’s modern customer data mastering capabilities, the bank now has continually updated and action-ready 360-degree views of its customers.
- Created complete, up-to-date customer views from 45 data sources and tens of millions of records in under four weeks
- Increased reliability of reporting data with unified data feeds for financial, credit risk & regulatory reporting
- Empowered sales for cross-sell opportunities with holistic view of customers across divisions and products
With the help of Tamr, Santander UK rolled out a dramatically faster lending process by cutting credit decision times in half with the rollout of a new lending system.
Scotiabank effectively cutting regulatory reporting cycle with machine learning
Scotiabank is a Canadian multinational banking and financial services company. It serves more than 25 million customers in more than 50 countries around the world, and offers a wide range of products and services including personal and commercial banking, wealth management, corporate and investment banking. The bank maintains separate data instances with high heterogeneity because of sovereignty requirements. In addition, they inherited data across jurisdictions through inorganic growth over the years. Scotia faces unreliable data quality with wildly varying degrees of data formats and completeness across multiple systems.
By deploying Tamr, Scotiabank was able to create high-scale ML clustering of customer records, taking input and curation by human subject matter experts (SMEs) to train the model with little or no data engineering knowledge needed. They were able to have multiple golden records created with Hive DB integration and ability to cluster and create unified IDs at multiple levels. In the system, they were able to set custom confidence levels that are adjustable to account for complicated business scenarios and inconsistency. As a result, Scotiabank was able to achieve:
- Reduced risk of error, improved compliance by mastering across 35 systems with ~4M entities
- Hierarchies for sales, risk and legal enabled customized views at different levels depending on business needs
- Cutting time to required to onboard a new system from months to days with unified customer data kept up-to-date through daily remediation
Tamr solutions enabled the bank to create a single profile of every customer across dozens of jurisdictions, providing the foundation for regulatory compliance and reporting efficiency.