Toronto Dominion Bank and its subsidiaries are collectively known as TD Bank Group ("TD"). TD is the sixth largest bank in North America by branches and serves more than 24 million customers in three key businesses operating in a number of locations in financial centres around the globe: Canadian Retail, including TD Canada Trust, TD Auto Finance Canada, TD Wealth (Canada), TD Direct Investing, and TD Insurance; U.S. Retail, including TD Bank, America's Most Convenient Bank, TD Auto Finance U.S., TD Wealth (U.S.), and an investment in TD Ameritrade; and Wholesale Banking, including TD Securities. TD also ranks among the world's leading online financial services firms, with approximately 10.7 million active online and mobile customers.
Consumers expect a seamless customer service experience across channels and flexible personalized service and pricing options. The faster a bank can bring new, innovative offerings to market, the more it will succeed at delighting customers and growing its market share and profit.
Using existing legacy data retrieval and preparation systems, it took six months to deliver data to business users to develop new products, explore customer insights, and pursue new opportunities for operational efficiency gains. Reliance on hand-coded, one-off efforts to extract data from legacy systems and deliver it to business users was driving IT costs, increasing reliance on proprietary hardware and software and encouraging the proliferation of redundant, poorly governed data silos across the enterprise—and that meant TD Bank was struggling to respond quickly enough to meet consumers’ expectations for new products and improved service.
The technology environment had become complex, slow, and fragile.
In 2013, TD Bank set an ambitious goal to transform the use of IT to strategically drive the business. The bank launched an initiative to deliver data to business users with a new model of fast, secure, on-demand access to enterprise data (Data-as-a-Service). This new agile IT 3.0 model would leverage a managed enterprise data lake to pre-position data from hard-to-reach enterprise systems, legacy platforms, and third-party data sources on a securely governed platform. Using an intuitive point-and-click, browser-based application, business users could directly request access to sets of data to support new projects or lines of inquiry or innovation, including access to all required supporting hardware, software, storage, and desktop tools.
TD Bank’s agile banking 3.0 program, the centerpiece of which is the data-as-a-service architecture that leverages a Hadoop-based enterprise data lake, is empowering thousands of the bank’s business people to directly browse, select, and retrieve data from dozens of enterprise applications—without assistance from IT or deep technical skills. Users can even add new data into the date lake using simple point and clicks tools, allowing them to extend the platform with new information to support changing business needs or new data sources. The bank’s ability to innovate and respond quickly to customers’ needs or changing market conditions has been dramatically improved. Business users can now get data to explore new ideas, analyze customers or pursue more efficient operating models in hours.
The IT 3.0 initiative is achieving a number of key results, including:
- Enabling users with self-service on-demand access to enterprise data
- Reducing the time required to deliver data to business users from 6 months to hours
- Reducing the quantity of redundant ETL projects and data siloes across the bank
The new platform is on track to cut TD Bank’s overall data management costs by 40-60%, reduce the time required to deliver new data to business users from six months to hours, and provide a consistent enterprise wide platform for analytics and reporting across more than 2,400 branches. Other benefits include:
Unique Customer Identification at 80% Lower CostIn recent years, TD Bank invested heavily in a proprietary third-party solution to identify unique customers by matching transactional records and customer data across dozens of databases and applications for marketing campaigns. Using its new IT 3.0 data set, TD Bank replicated this process internally and achieved a 39% improvement in match rates—at 80-90% lower cost.
Big Data Enabled Customer SegmentationIn the past, processing constraints associated with traditional data warehousing hardware limited how much data TD Bank could use to build customer segmentation models. Generally 300 million customer transaction records representing one year of data was the maximum volume of data that could be included. This has been massively expanded to include 20+ billion transaction records covering five years. The model was also expanded to include more behavioral attributes and to allow for dynamic calculation of threshold values for each attribute to drive more meaningful customer clusters. The result has been a more accurate segmentation model, revealing new customer insights and more revenue and service opportunities.
- Podium Data Platform and Pipeline Controller from Teradata
- Cloudera Hadoop on Cisco Hardware
Like many banks, TD Bank’s IT projects had focused almost exclusively on reducing costs by automating existing processes, and tooling up new customer facing channels to deliver all of the services previously available in a branch office. With the new agile IT 3.0 model—and now unlike many other banks—IT has moved from being complex, slow, and fragile to simple, agile, and resilient.
The common theme of self service, standardization and reusability continues into the bank’s AAAS vision. There are 1,200 analysts with only a small percentage being data scientists. For the majority of the analysts, the vision has been to give them interactive apps for visualization and analytics built on a standard and agile framework while giving data scientists the ability to explore and build reusable, advanced analytical applications and processes.
Early benefits from the IT 3.0 strategy have been incremental. The wins have been in the areas of planning and execution of the technology. The true shiny moment will come when we fully leverage all patterns of IT 3.0 ranging from apps, APIs, cloud and most importantly data. We have shifted into the leverage phase and the measurement will shift from efficiency metrics to business metrics covering customer service, acquisition and retention - directly impacting business value and outcome.
