How Banks Can Compete More Effectively Using In-Database Analytics Increase your information turnaround speed exponentially while achieving vastly superior data integrity and accurate results
White Paper
How Banks Can Compete More Effectively Using In-Database Analytics
Table of Contents Executive Summary......................................................................................................1 How can you get more timely insights from ALL your data to drive better decision making?......................................................................................2 Business Benefits ........................................................................................5 Customer Case Studies and Successes: Demonstrating the Power of In-Database Analytics ..................................................................................................................6 Commonwealth Bank of Australia (CBA)........................................................6 Discover Financial Services...........................................................................7 Hellenic Bank................................................................................................8 A Large, Multinational Bank...........................................................................9 A Large, Multinational Latin American Bank ...............................................10 A Top International Bank ............................................................................10 Unlock the Potential of Your Enterprise Data Today ...................................................11
i
How Banks Can Compete More Effectively Using In-Database Analytics
Executive Summary
“[Information and knowledge management] professionals are adopting an emerging best practice known as ‘in-database analytics.’ Under this practice, data mining, predictive analysis and other compute-intensive analytic functions are migrating to the EDW platform. In-database DW analytics can either replace or supplement traditional analytics execution approaches.”
Consider the number of decisions made within your business each day, and the time it takes to make them: • Hours to decide which customers and what channels are the best ones to target for up-selling a new product, based on results from the last campaign. • Minutes to profitably price the risk of offering credit to a potential new customer who is looking at a competitive rate at another bank. • Seconds to determine if a long-term customer on hold with your contact center or visiting your Web site qualifies for special pricing that will incentivize them to buy the service or product, as well as keep them coming back. Until now, most companies have found it difficult to deliver the reliable, accurate and near real-time insight required to support critical decisions like these. This paper explains how you can use in-database analytics not only to accelerate your time to insight, but also gain greater confidence in the accuracy of your decisions and maintain the security and integrity of your data.
James G. Kobielus Senior Analyst Forrester Research “ Massive But Agile: Best Practices for Scaling the Next-Generation Enterprise Data Warehouse”, Forrester Research white paper, June 2009.
Predictive analytics – sometimes referred to as predictive data mining – is a branch of business analytics that uses historical data to make predictions about future events through sophisticated modeling techniques. Increasingly, predictive analytics is being used for critical business purposes, such as: • Predicting when a customer might be about to move to a competitor. • Predicting if a customer would be a likely up-sell opportunity. • Predicting loan default rates and identifying customers in danger of defaulting. • Predicting future stock prices. • Identifying fraudulent activities by identifying transactions that deviate from predictions (e.g., credit card transactions that deviate from past behaviors can be flagged and investigated for possible fraud). • Determining risk of illness either to determine insurance premiums or – more helpfully – to arrange preventative medical programs. Guy Harrison, “Crystal Ball-Gazing With Predictive Analytics,” Database Trends and Applications, July 2009.
1
How Banks Can Compete More Effectively Using In-Database Analytics
How can you get more timely insights from ALL your data to drive better decision making? Even in a strong economy, over a five-year period, it’s not uncommon for a business to lose up to half of its customers. Equally daunting is the fact that acquiring a new customer can cost six to seven times more than retaining an existing customer. Without near real-time customer insight and updates to your enterprise data warehouse (EDW): • How can you ensure that your new credit card customers are offered the most profitable rate based on their existing banking behavior and relationship? • How can you accurately assess and report each customer’s risk and reward potential as it changes daily, without creating bottlenecks in your credit processes? • How can you detect fraudulent activity as soon as it occurs in your portfolio of mortgages, loans and credit lines – before these events cause significant losses? • How can you obtain a quality portfolio of credit card customers – such as a customer at an electronics store who needs an instant loan approval and wants to take his purchase with him? • How can you efficiently execute campaigns to make the most relevant offers to the potentially profitable customer at the appropriate time, using the most appropriate channel (e.g., branch, call center, ATM, Web)?
The Vision • What if you could identify, through branch visits or online transactions, certain home mortgage customers who are likely to purchase an equity line of credit, and then immediately flag them to receive an offer? • What if you could better manage campaign costs by eliminating guesswork and assigning offers based on data accurately detailing the most current and up-to-date customer behaviors and trends? • What if you could offer a better package while you’re talking on the phone to a customer who is about to refinance his mortgage through a competitor? • What if you could measure active campaigns to ensure they are meeting objectives and modify them in real time to maximize effectiveness?
2
How Banks Can Compete More Effectively Using In-Database Analytics
“In the securities industry, certain types of risk, like market risk, happen second by second. Being able to have an architecture that copes with that is going to be vital.” David Furlonger VP and Investment Industry Analyst Gartner Inc. “E-Trade Turns to Agile Response Team for Rapid Risk Analysis,” Securities Industry News, April 13, 2009
“Enterprises have substantially completed their adoption of core BI, enterprise data warehouse, and enterprise content management platforms and will increasingly turn to powerful predictive analytics, data mining, statistical analysis and text analytics tools to leverage that information for business optimization. One consequence of this trend will be the growing adoption of in-database analytics techniques, under which users will process these compute- and data-intensive functions inside the enterprise data warehouses, taking advantage of that platform’s massive parallel processing.”
Aggregating, analyzing and processing large volumes of data quickly enough to support critical decisions that must be made in hours, minutes or seconds is difficult without using in-database analytics and an EDW. Ensuring the security of corporate enterprisewide information is even more challenging, because IT has to extract, move and replicate data from an EDW for the analytic team to use it for data preparation, model development and deployment. This process can take days or weeks. By the time decision makers receive what they need, the opportunity to make the most efficient use of it has passed. As a result, people at all levels of a business are forced either to make decisions based on the analysis of outdated, untrusted data or work-educated instinct. To address this challenge SAS and Teradata Corporation have joined forces to support in-database analytics, which enable you to analyze data directly within a Teradata database without having to extract and move it first. The joint solution reduces data movement and latency; improves security and data consistency; and by leveraging the database platform parallelism, accelerates the processes of analytic data preparation, model development and scoring by running these processes inside the Teradata EDW. By optimizing how your analytics and database technology investments work together to perform their unique but dependent functions, you can improve performance and make better decisions based on the reliable and latest set of data – helping you reduce fraud, mitigate risk, optimize pricing, and attract and retain your best customers. Our Approach SAS has tailored its software to run directly inside a Teradata database, so you get more from your data – faster and at a lower cost. Rather than having to pull a massive data set from your EDW and move it to another server for analysis, now you can run and optimize select SAS® software directly within your Teradata database. Your data is secure, properly managed and you get results with far fewer errors than when using the old method of extracting, moving or copying, and analyzing data outside the database. Key SAS analytical functions are now able to exploit the core Teradata parallel architecture, giving you significant improvements in speed and performance. Equally important, you can trust the results of your analysis either for a given analysis and across lines of business because the data warehouse provides the most current, consistent and accurate data available to the business.
James G. Kobielus Senior Analyst Forrester Research “ Experts forecast business intelligence market trends for 2009,” SearchDataManagement.com, January 7, 2009
3
How Banks Can Compete More Effectively Using In-Database Analytics
As a result, you can get reliable answers to your urgent business and customerrelated questions in hours, minutes or even seconds. For example, deploying the SAS model-scoring logic directly within the Teradata database engine significantly increases processing speed and performance, and reduces model-scoring rewrite and revalidation man hours. Results that used to take days or weeks can be completed in an hour. This means you can potentially: • Reduce costs and time to make intelligent decisions from months down to hours. • Make the next most relevant offer while customers are still in front of sales people. • Measure potential spending and profitability before customers leave your site. • Receive alerts regarding criminal or fraudulent activities before transactions are complete.
The Vision • What if you could better predict default and delinquency on your throughthe-door population, as well as get new risk models into production in days vs. months? • What if, on a daily basis, investigator queues contained a list of prioritized alerts based on risk scores that incorporate KYC data, transactions, historical behavior and expected activity across all accounts? • What if you could consistently predict the impact different scenarios would have on your risk exposure, identify portfolio concentrations on multiple dimensions, and know with certainty that your capital reserves were commensurate with the risk that you face? • What if you could be confident that the risk models you have in production were performing optimally throughout their life cycles and that you would know as soon as model performance began to degrade? • What if, with the click of a mouse, you could see your risks at multiple levels (branch, product, geography) and drill down to identify areas and loans that pose the greatest threat, based on daily updated behavior? • What if you could respond to company board and management inquiries by drilling down through a report and decomposing your risk calculations to show exactly how you arrived at your numbers – in a matter of minutes or hours vs. weeks or days?
4
How Banks Can Compete More Effectively Using In-Database Analytics
Why Teradata? Teradata is recognized by both Gartner and Forrester as the leader in EDW solutions. The Teradata DW and EDW platform family offers: • A single, integrated view of the enterprise-wide data needed to make smarter, faster decisions. • Consistent and accurate data at both the detail and summary levels. • Linear, scalable parallel database processing that supports nearly unlimited amounts of data and users, as well as simple to complex queries, mixed workloads and high concurrency. • An affordable, flexible and proven Teradata purpose-built platform family to meet diverse business needs from departmental to across the entire enterprise.
These are clearly the kinds of capabilities that customers have been asking for and analysts are excited about. Dan Vesset, Program Vice President for Business Analytics Research at IDC, had this to say when the SAS and Teradata in-database analytics initiative was announced:
“ Today’s businesses are challenged to manage huge data volumes while optimizing analytical model development and deployment environments. In-database analytics enable IT and analysts to be more productive and responsive to the growing demand by business decision makers for analytics-based strategic and operational decision support. Today’s announcement shows that SAS and Teradata are closer to delivering on the promise of in-database analytics, which is to give customers an efficient, powerful means for implementing predictive analytics and information analysis in one location.” Business Benefits Our combined efforts give you both the powerful in-database analytics for quick results on a reliable set of data to improve business performance and required process and data governance. In banking, data quality and data governance initiatives are often driven by external regulations such as Sarbanes-Oxley, Basel I and Basel II, and the need to reduce risk exposure. The risks can be financial misstatement, inadvertent release of sensitive data or poor data quality for key decisions. Using in-database analytics can help you: • Decrease the risk of regulatory fines through improved data integrity, data quality and reporting ability to regulatory authorities on the newest, consolidated set of information. • Maximize the income generation potential from your data. • Gain near-instant insight needed to drive personalization and other customercentric strategies. • Gain insight into trends, risks and opportunities, and respond swiftly to improve your competitive outcomes. • Free up staff to focus more time on value-adding analytic modeling activities vs. data extractions, movements and conversions.
5
How Banks Can Compete More Effectively Using In-Database Analytics
Why SAS? SAS is the leader in business analytics software and services, and the largest independent vendor in the business intelligence market.
Customer Case Studies and Successes: Demonstrating the Power of In-Database Analytics As the following customer examples illustrate, the vision of in-database analytics can be a reality today, enabling faster, more-informed decisions that help you retain customers, manage risk, grow your business and compete more effectively.
• More than 2,900 financial institutions worldwide are SAS customers.
Commonwealth Bank of Australia (CBA)
• 96 percent of banks in the FORTUNE Global 500® use SAS.
Situation:
• SAS has more than three decades of experience working with financial institutions all over the world. Since 1976 SAS has been giving customers around the world THE POWER TO KNOW®.
The bank needed to reduce the high expenses associated with fraud detection and the difficulty of quantifying and reducing actual fraud losses. The bank wanted to implement an anti-fraud system. Alternative approaches and processes to solving this need often include a silo approach – a hodgepodge of databases and systems scattered across different business units and channels producing data with high latency and questionable relevance. CBA wanted to identify data patterns to generate the insight necessary to combat the sophisticated types of fraud.
Solution: • CBA migrated all of its siloed information into the Teradata warehouse in order to analyze transactions and customer activity. Results: • The joint SAS and Teradata solution helps the bank access customer information as it changes – in real time – to quickly identify suspicious behavior and act on it as it is happening. • The bank improved efficiency and performance by fine-tuning existing models and developing new models to predict fraud before it happens. • The bank has detected twice the level of check fraud than in its legacy system and increased Internet banking fraud alerts by 60 percent. • Check and Internet fraud loss-to-turnover ratios have improved by 50 and 80 percent respectively.
6
How Banks Can Compete More Effectively Using In-Database Analytics
Discover Financial Services
Situation: Hurricane Katrina struck the Gulf Coast of the United States. Discover needed to provide significant financial support to help affected people when they had nothing but their wallets and a few things they could carry from their homes. For that to happen, Discover had to quickly identify a specific population of cardholders and their active status.
Solution: • Within 90 minutes of the hurricane hitting, Discover analysts were supplying metrics to decision makers to determine the cost and risk of extending credit to cardholders in the affected areas, bringing together sales, activation, performance and bankruptcy data. • Discover analysts used SAS and other tools to perform cross-line-of-business (cross-LOB) analysis on data stored in the Teradata EDW. Results: • Customers in the disaster area were extended additional credit within a day of losing everything. Discover’s IT group plans to continue to integrate new advances in SAS and Teradata technologies over time so: • Company sales representatives will be able to provide their customers with specific products and services tailored to their individual’s needs, based on the most up-to-date behavior and demographic data vs. data that is weeks and months old. • Phone representatives will have next-best-offer information available on the spot, which will help them provide even more effective customer service and drive increased revenues. “With Teradata and SAS, Discover Financial Services’ IT group can better serve the business. We have clearly seen benefit in the form of higher revenues, expense takeout and improved customer satisfaction,” says Glenn Schneider, Senior Vice President and Chief Information Officer of Business Technology. Cheryl D. Krivda, “Cache Back: Enterprise Data Warehouse Helps Drive Innovation at Discover,” Teradata Magazine, March 2009.
7
How Banks Can Compete More Effectively Using In-Database Analytics
Hellenic Bank
Situation: The bank needed to comply with Basel II requirements in a very short span of time. Key to doing this was the bank’s ability to analyze risks across multiple business processes and automate workflows.
Solution: SAS and partner Teradata used Credit Risk Advantage, a part of the SAS and Teradata Risk Advantage Program, to integrate risk management into daily business activities in order to catch risky behavior as soon as it occurs. • The joint solution leverages an integrated architecture providing advanced risk management for regulatory compliance and the optimization of risk-adjusted pricing and returns. • The joint SAS and Teradata solution contributed to increased efficiency and improved customer service. • With improved data management and governance, using all of the bank’s customer data, the bank can create models for more advanced capital calculation methods. Results: • Implementation was complete in less than eight months. • Automated results reduced analysis time from four days to two hours. • The bank achieved Basel II compliance, along with more effective analysis of risks across multiple business processes. Petros Ioannides, Manager of Group Credit Risk at Hellenic Bank, said, “Capital adequacy calculation was a manual process that required extensive effort to complete. Now, with a press of a button we know exactly where we are in terms of capital within two hours.” Moving forward, the bank expects that other branches of the bank can become familiar with the returns of funds from interrelated and crossrelated areas. The bank will be able to quickly calculate capital consumption by sector, which could be particularly useful for evaluating new services and products.
8
How Banks Can Compete More Effectively Using In-Database Analytics
A Large, Multinational Bank
Situation: A large, multinational bank needed to combine data across business lines, increase marketing response rates, reduce bad debt on credit cards, and help business analysts work collaboratively and more productively. Past response rates to blanket marketing offers were less than 1 percent. Eighty percent of an analyst’s time was spent pulling all the data together.
Solution: To increase marketing response rates to credit offers, the SAS and Teradata integrated framework segmented customers, not just by how they use their current cards, but by what drives their spending. They are now able to more effectively do this by using new weekly data from third-party marketing partners and its own constantly changing internal data. • With current behavioral and attitudinal segments, the bank ranks customers by lifetime value and chooses offers to better match its customers’ needs and ability to pay. • The bank is able to analyze each customer’s risk profile, along with their needs and propensity to respond, prior to sending out offers. • The bank also segments late-paying customers to determine methods to help those customers reduce their debt. This approach has enabled the bank to reduce bad debt by 5 percent. Results: • Increased marketing response rates twentyfold from 1 to 20 percent. • Reduced bad debt on credit cards by 5 percent. • The integrated solution eliminates the need to copy and replicate data from one solution to another and allows the bank to use analysts more effectively.
9
How Banks Can Compete More Effectively Using In-Database Analytics
A Large, Multinational Latin American Bank
Situation: The bank’s complicated, manual information deployment processes took months to complete. As a result, decisions were based on inaccurate and out-of-date risk models. The bank makes thousands of credit decisions every day. They have hundreds of credit officers working on originating and approving loans. Many situations require instant loan approval. For instance, a customer at an electronics dealer wants to take his purchase with him immediately. In order to continue to compete profitably, the bank realized it needed to make more accurate instantfinancing decisions.
Solution: The bank built a credit risk infrastructure using SAS in-database analytics with Teradata. The new integrated system: • Puts risk management into daily business activities while satisfying a complex mix of Basel II requirements. • Interfaces to a credit bureau report, validates the customer’s identity and runs the application through various models, such as propensity to default. The result is that the bank now delivers an answer to the requestor within eight to 12 seconds, and it is used to help drive new sales to desirable customers and enable better management of others. Results: • Identified and managed nonperforming assets in near-real time, thus reducing losses. • Assessed current and future risks by simulating and stress-testing portfolio and balance sheet performance. • Lowered capital requirements by using Basel II approach for calculating credit risk – such as IRBA – across all portfolios. • Millions of dollars saved through better credit decisions – using trustworthy credit risk data in Teradata.
A Top International Bank
Situation: The bank was spending more time planning how to market their services than actually marketing them, due to the lag time and inability to accurately include changes in their customer behavior as they occurred. In order to more profitably compete, the bank needed to up-sell its customers with the most relevant offers at any point in time.
10
How Banks Can Compete More Effectively Using In-Database Analytics
Solution: • Assign propensity to buy, propensity to spend and (for credit cards) propensity to activate, based on daily updates to customer data in the EDW. • A current, daily 360-degree view that shows the credit card account, fixed deposits and asset accounts – the totality of their relationship. Results: • As a customer’s behavior moves them to the highest-performing, highest-margin stratum, the bank now can immediately offer an almost concierge-like experience that encompasses aggressive, attractive pricing and multiple cross-selling offers. • Today, more than 35 percent of the bank’s credit card portfolio has resulted from cross-selling standard liability products (such as savings accounts) to existing customers.
Unlock the Potential of Your Enterprise Data Today For organizations whose success or failure depends upon accurate and timely decision making, integration of database and analytical capabilities can make all the difference – and lead to greater profitability and growth. SAS and Teradata Corporation have taken the investment in their strategic partnership and in-database analytics to the next level by creating the SAS and Teradata Analytic Advantage Program. This program provides competitively priced, integrated packages that enable you to quickly and cost-effectively implement and deploy SAS Business Analytics with the Teradata Enterprise Data Warehouse platform. The SAS and Teradata Analytic Advantage Program includes three different packages – Express, Advanced and Enterprise – to meet your growing analytic needs. Each of the three packages includes SAS software with consulting and implementation services, and is paired with the appropriate Teradata data warehousing platform solution to meet specific customer analytic processing requirements and direction. SAS and Teradata have also created the SAS and Teradata Center of Excellence, which features a dedicated team of solution architects and technical consultants to help you use our solutions to optimize the performance of your existing and future infrastructure around in-database analytics. Services include architecture assessments and recommendations, proof of concepts, benchmarking and sizing analyses, and customized consulting. Are you ready to transform business decision making through the power of indatabase analytics? For more information, please see the SAS and Teradata Analytic Advantage Program Web site: www.sas.com/technologies/analytics/advantage.html.
11
Copyright © 2010 by SAS Institute Inc. and Teradata Corporation. All Rights Reserved. SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. Teradata and the names of products and services of Teradata Corporation are registered trademarks or trademarks of Teradata Corporation in the USA and other countries. ® indicates USA registration. Other brand and product names are registered trademarks or trademarks of their respective companies. 104244_S46080.0310 (SAS)
