Last Mile Execution · PDF fileLast Mile Execution Technology solutions that deliver deeper...

Last Mile Execution Technology solutions that deliver deeper customer insight into your decisioning systems

Lee Grice, James Blue

May 2013

Table of Contents

1 Introduction: The Incredible Shrinking Bottom Line

3 The Demand for Customer Insight

5 The Case for Alternative Data

6 Data Drives Decisions and More Data Means Better Decisions

7 Technology: Old challenges. New opportunities

9 Suggestion #1: Simplify to grow

10 Suggestion #2: Gain efficiencies from bureau engine technology

13 Suggestion #3: Augment your legacy systems

14 The Equifax approach

About Equifax

Equifax is a global leader in consumer and commercial information solutions, providing businesses of all sizes and consumers with information they can trust. We organize and assimilate data on more than 500 million consumers and 81 million businesses worldwide, and use advanced analytics and proprietary technology to create and deliver customized insights that enrich both the performance of businesses and the lives of consumers. For more information, please visit Equifax.com.

www.equifax.com

Equifax Inc. | Last Mile Execution | ii

Introduction: The Incredible Shrinking Bottom LineBanks are looking for new ways to improve customer insight to grow their bottom line. To survive and thrive, banks are striving to incorporate customer insight that drives better business decisions and profitability. This paper is a must-read for marketing, risk, IT, analytics and credit professionals that are involved with the mission-critical processes of improving interactions with the bank’s customers.

Meaningful statistics regarding the challenges the banking industry has faced over the last five years are not difficult to find. The following table summarizes key statistics regarding banking industry growth over the last 5 years (data compiled from FDIC.GOV).

A quick review shows bank profit has suffered as loan performance and diminished fee-income have negatively impacted performance. The industry as a whole has seen reasonable growth in deposits, which does create some opportunities for banks to increase lending as consumer economic factors improve. The graphic below visually represents the table above.

As banks look hard at internal data on account and product profitability to see where opportunities exist, it is not a surprise that Identifying, acquiring, and retaining “good” customers is the best way for banks to help ensure success. Consequently, banks are seeking to improve lending and account opening decisions in order to efficiently build multi-product relationships with customers.

Last Mile Execution

Table 1. All institutions - National. Totals in thousands.

All Institutions - National

12/31/1997 9/30/2002 9/30/2007 9/30/2012

Number of institutions

Total assets

Total deposits

Income earned

10,923

$ 6,044,717,755

$ 4,125,799,743

$ 26,782,450

9,415

$ 8,272,829,759

$ 5,391,981,629

$ 41,302,722

8,559

$ 12,706,149,708

$ 8,180,225,813

$ 70,734,222

7,181

$ 14,222,940,859

$ 10,504,235,135

$ 46,762,105

Equifax Inc. | Last Mile Execution | 1

However, lenders intent on re-establishing consumer credit as a key bottom-line contributor face two significant challenges. Think of them as two chasms that must be crossed.

The first is an information chasm. Banks need more timely and relevant information about new applicants and existing customers. Access to information is a significant challenge for many banks as depicted in this chart, from a 2011 AITE Group survey of 100 US Community banks.

Rolling 5 year CAGR for Key Banking Statistics

� Total assets � Total deposits � Loan Income earned

15%

10%

0%

5%

-5%

-10%

1997-2002 2002-2007 2007-2012

Figure 1: 5 year CAGR.

Community banks struggle to access data

� Challenging � Very Challenging � Extremely Challenging

Lack of access to internal data for inclusionin analytics, modeling, segmenting

Lack of access to qualify externaldata and analytics

Q. To what degree are each of the following challenging your commercial lending group?(Those responding “challenging” to “extremely challenging”; n=100)

Lack of analytics to measureperformance of asset based loans

Lack of access to real-time transactiondata to monitor business performance

0% 10% 20% 30% 40% 50% 60% 70%

Figure 2: Source: Alte Group survey of U.S. community banks, May to July 2011.

23% 19% 19%

25% 15% 15%

24% 15% 15%

22% 17% 17%


The second is a systems chasm. Existing software systems are inhibiting growth through their inefficiencies, costly maintenance, cumbersome interfaces, and inability to react readily to change. When Aite Group surveyed CIOs and senior IT executives at 80 large banks in 2011, researchers found that the top challenge to customer satisfaction is data incompatibility. The two greatest opportunities, Aite Group reported, are leveraging data in real time enterprise-wide and replacing aging software and analytics.

A good summary of the systems chasm is made in an article titled “The Case for ‘Progressive Renovation’” by Bank Technology News. The article describes how many banks are constrained by legacy technology which is expensive to maintain and often doesn’t meet the demands placed upon them.

Most banks are not accessing the latest customer insight, because their legacy systems are unable to consume new data sources and analytics. Before we discuss best practices around upgrading legacy systems and processes, let’s look at how the market and the need for deeper customer insight has evolved.

The Demand for Customer Insight With the stroke of a pen, the Dodd-Frank reforms of 2010 forced US banks across the board to re-examine their business practices. Not only did banks have to accommodate new compliance obligations, but their traditional fee and income models were rendered obsolete. This table shows how changes in the compliance environment and marketplace have impacted components of bank strategy.

Table 2. Before/After banking crisis.

Before banking crisis After banking crisis

Free checking pervasive

Opportunistic marketing

Profitability managed at the product level

Customer acquisition dominant strategy

Channel-specific products

Fee-checking more the norm, selective free

Targeted marketing

Profitability managed at household level

Multi-product relationships more important

Customer management across multiple channels


Banks will have an increased appetite for information to develop the roadmap to accommodate the changing strategy. Relationship and performance information, often locked away in legacy systems, must be unencumbered and made available in support of product design and policy development activities. As existing customers enter into servicing channels, banks that cannot leverage what they already know about these clients to better serve their needs are missing a huge opportunity to build deeper, more meaningful relationships.

External information also plays a tremendously important role in helping banks convert prospects into customers. External information supports such important processes, such as authentication, decisioning, cross-selling, and on-boarding throughout the enterprise. Banks have long-since used credit information to understand risk, but other opportunities exist to better understand customers and prospects. More customer insight, both internal and external, will help banks better respond to customer needs and create profitable, meaningful, multi-product relationships.

In summary, customer insight will be essential for banks to create smart growth. Smart growth is essential, as historically low interest rates are diminishing the income available from increased deposits. And, cost per account has increased due to relatively fixed infrastructure costs along with increasing variable costs due to regulatory changes.

How the Banking Crisis Affected Account Profitability

� NSF Income � Other Non-Interest Income Average Cost Per Account

NSF Transactors(0$ balance assumed)

<$500in balances

$500 to $1,000in balances

$1,000 to $3,000in balances

>$3,000in balances

Figure 3: How the Banking Crisis Affected Account Profitability

NSF revenue reduced (mostly due to regulations)

Average account expense increased due to lowerrevenues in high profit segements

Overall lossof wealth reduced

Non-Interest income


The Case for Alternative Data

Traditional credit information provides an important view of consumers' true financial obligations and their ability to satisfy them. Additionally, there is a significant opportunity to use “alternative data” to help assess risk for accounts that may not have sufficient information to score using credit file data alone.

Alternative data sources expand the perspective beyond traditional credit histories into new areas, such as income, assets, and payment performance of accounts that won’t show up on the credit file. Using new alternative data sources, for example, modelers can create predictive value from payment data, such as consumers' wireless phone, satellite television or utility accounts.

Equifax's new Insight Score for Retail Banking (ISRB) solution uses payment data to help financial institutions boost new account acquisition rates at the point of sale. ISRB delivers fresh, unique visibility into consumer DDA risk, allowing financial institutions to identify and serve more accounts.

ISRB includes more than 25 million consumers with no credit scores and as many as 70 million people with thin-file and no-file credit. By understanding the payment behavior of these consumers — particularly payments not on the traditional credit file — banks can expand their reach into new segments.

Banks understand there is significant value to adding new data sources, but they struggle with integration and data management issues, both of which are magnified by the prevalence of legacy technology platforms.

If these issues can be overcome, using alternative data sources can help deliver valuable benefits to support growth strategies.

In working with several banks to integrate alternative data with conventional approaches, Equifax has often seen:

• Lowerbadrateswhenthenumberofaccountsremainsconstant

• Moreaccountsapproved,withlittlechangeinthebadrate

• ReductionsinDDAaccountcharge-offs


Data Drives Decisions and More Data Means Better DecisionsHow can banks turn massive amounts of data into insights that help enable them to make better decisions? Let us look at the process financial institutions use to create scores, the key driver of decisions.

Transforming data into insightful scores for decisioning seems straightforward, but in practice the art is fraught with layers of complexities that arise from:

• handlingbureaus’differentscoringsystems• managingmodelsinmultipletechnology“silos”thatsupportproductlines and channels • assuringthatdecisionsarecompliantwithregulatoryrequirements• lackofcontrolinmanagingmodelsfrommultipleproviders• inabilitytore-usemodelsamonglegacysystems

Ultimately, what typically happens is that each of a bank’s product lines ends up with a labyrinth of support systems that may result in inefficiencies and organizational risk. It’s not unusual for a bank to have:

• 3ormoreuniquedataaccessplatforms• 4ormoredifferentanalyticalsystemsfordeliveringmodelsandanalytics• 4ormoredifferentdecisioningengines• 3ormoreloanoriginationsystems

Turning Data Into Decisions

Figure 4: Turning Data Into Decisions

Structuredbut not

assembledinsightfully

Structured,often aggregated;

can be simpleor derived

Algorithmsthat calculate

predictivescores

Rank orderingbased on

dependentvaiables

Resultingactions to apply

bank policyto scores

DecisionsScoresModelsAttributesData


Technology: Old challenges. New opportunities

To get the high degree of sophistication needed for real-time decisions, banks must find better ways to bring timely, relevant data into their decisioning processes.

The chain of events for decisioning can be extremely complex; so, there are a myriad of reasons why bankers may not be getting the results they expect from their current processes. A few examples show why it’s difficult to add data sources, as well as creating and managing attributes in their business processes:

• Thesoftwareusedtoaccesscreditandotherdatasourcesusedindecisioning,often referred to as a bureau engine, might not be up to date on new and niche data sources—and won’t support them. If one system can’t support what the bank needs, other methods must be found to get data into the system they are already working with.

• Those older systems are difficult to update and require special resources.Banks either have to depend on professional services or maintain development staff to support the bureau engine maintenance, which may be dependent on older technology, creating other resource challenges.

• Once access has been established, then attributes and models must bedefined by systems analysts based on user requests, then coded, tested and deployed. Building out the analytics to derive value from the new data can take months to complete.

Disparate Technology Creates Problems

• Each Product Line has its own technology driving data access and attributes.

• Cost of change enormous, and little re use possible.

• Cost to add incremental data intelligence prohibitive.

Figure 5: Disparate Technology Creates Problems

Attributes 1

Vendor 1 data access

Attributes 2


Attributes 3


Attributes 4


Consumer lending

Credit card

Business lending

Consumer deposits

Multiple pipes and lines of businessThe Problem

?


Other factors adding more complexity are the frequency of compliance changes and the proliferation of technology silos.

• Compliance changes often affect the information used in decisioning, and the cadence of change remains at a high level. Often the resulting requirements affect the decisioning systems. And, changes can be initiated by banks themselves or pushed upon them by updates at the data sources, which are also regulated entities. For example, if a credit report adds new narrative codes in support of a regulatory requirement (a relatively minor change), it may affect multiple attributes within a bank’s policies.

• Banks often have a proliferation of incompatible technology silos in theirproduct lines and in their channels. It’s not uncommon for banks to have different bureau engines in support of credit card, direct deposit, auto, home equity loans, and other lines of business. Each may have its own definition of attributes—and they may not align across product lines and across the organization’s mobile, branch banking, and other supported channels.

As a result, maintenance costs are compounded for multiple platforms, and customization projects may inhibit future upgrades or further increase support costs.

Consequently, users and IT staff must modify each of the product silos to get new data, create new attributes, and deliver new scores. Annual costs of such changes for all decisioning processes can be significant—even before the organization tackles new cross-selling initiatives. These “hard dollar” costs don’t take into account opportunity costs, such as the inability to use models at full strength, forcing users to settle for less-predictive models.

But in their current legacy systems environments, pulling in more data is best achieved when banks are simultaneously addressing technology problems. Two additional ingredients are recommended to augment traditional risk models with alternative data sources, thereby progressing toward the goal of turning raw data into intelligence:

1. Centralizing the management of attributes and models. With data for credit and onboarding decisions in one place, systems supporting products in any line of business have centralized access to all relevant data.

2. Applying expertise in analytics to improve the predictive qualities of attributes used in decisioning. Decisioning requires more than modern tools, alternative data sources and well-integrated systems. Highly sophisticated expertise is also essential for transforming data into meaningful measurements; that is what analytics does.


What’s the best strategy for overcoming these challenges?

The answer may vary for every bank, but there is a recommended “common ground” that should serve as the starting point:

Define global best practices throughout the organization’s decisioning processes for adding new data sources, creating attributes and models, and integrating results into runtime environments.

Suggestion #1: Simplify to growOrganizations trying to grow in a mature, highly competitive, highly regulated business need every advantage they can get. In this environment, technology must be an asset, not an albatross. That’s why progressive bankers are working to simplify how they get the right data into their decisioning systems and allow their IT to focus on more strategic initiatives.

Steps to simplification:

Leverage a common data integration engine company-wide

Savvy modelers are asking, why not build common attributes just once—ready to reuse for all their lines of business—so that banks can respond to readily changing market needs? Modern technology tools help make this fast, feasible and affordable to accomplish. Every time you reuse a common attribute, you are adding speed, efficiency and profitability by subtracting a cumbersome, time-intensive process.


Centralize analytical attributes and scorecards out of decisioning

Don’t stop with the benefits of a common data integration engine. To help cut costs and gain efficiency across all product lines, centralize the management and maintenance of the attribution and analytics management process. Quality assurance can be simplified and improved by centrally managing attribute definitions, when they are used, and how they are used in decisioning processes across product lines and channels.

Deploying attributes within a bureau engine helps simplify development and maintenance of models. For example, there would be no need to define “debt burden ratio” differently for every model that uses it, when it’s accessed as a centralized attribute. New tools now enable a bank’s non-IT personnel to define a common attribute one time, keep it in an extensive library of standard and user-defined attributes, and leverage its reuse as needed throughout the organization.

Suggestion #2: Gain efficiencies from bureau engine technology Enable business users and bypass IT bottlenecks

For many banks, it can take six months to add new attributes and modeling to a legacy system. This puts a severe strain on the bank’s ability to respond to changing market conditions. Banks are now moving towards using newer technology, delivered through an outsourced “software-as-a-service” (SaaS) format that can make the same changes in as little as one month (and that period involves only one week of personnel activity; the remainder is driven by the bank’s software testing cycle). Best of all, changes are made by letting business users manage deployment, often without being dependent on IT resources.

Benefits of Shared Solution

• Centralized data access and attribution.

• Change can be managed across all product lines.

• New data assets are ubiquitous

Figure 6: Benefits of Shared Solution

Consumer lending

Credit card

Business lending

Consumer deposits

Centralized access to data and attributes

!

The Solution

Equifax

Bureau Engine


Intelligent orchestration across multiple data sources

Orchestration can be greatly simplified by outsourcing data access management. Orchestration refers to the business rules for “when and where” to access specified data. For example, orchestration can employ the common technique of “waterfalling.” (Example: a bank wants to obtain income data from Bureau A and debt data from Bureau B. If there is no hit with Bureau, request debt data, too, from Bureau A.) Another form of orchestration is using preference tables to determine what data to pull. (Example: A bureau may wish to first request a credit file from a set of zip codes because it gets preferential pricing in that region.)

Another application of orchestration appeals to banks concerned with fraud control. A bank can access as many as 15 different fraud mitigation products from Equifax alone. Through intelligent orchestration, the user can set up seamless, automated access to a complete portfolio of fraud solutions. The outsourcing services select appropriate fraud solutions from provider of choice based on user-defined criteria. Once the chosen solution makes an initial determination of fraud risk, further data sources could be requested if needed.

Orchestration Pattern: Customer Waterfall

Transaction “orchestration” is:

• If no-hit or thin file, access alternative data no-hit solution.

• If score indicates low risk, access wealth information for additional customer cross-sell segmentation

• Calculate credit attributes and custom score

Figure 7: Orchestration Pattern.

Equifax

Experian

Trans Union

Wealth

No-hit solution

Analytics Calculation

Data Deduplication

Orchestration

Data Access

Calculate thinfile condition

Access No-hitsolution

AccessCredit Bureau No-hit? Thin file? Score

Low Risk?

Access WealthInformation

Calculatecustom

attributes

Return data,attributesand score


Data Normalization

Normalization is the process of smoothing over the slight variations in how different data sources, in this case from credit bureaus, define a particular attribute. Take debt burden ratio, for example - most banks would prefer to define it one time, and it would thereafter be calculated the same way, regardless of the data source.

The goal is analytical accuracy. If the three credit bureaus define an attribute differently in a tri-bureau reporting scenario, that attribute should not be used interchangeably within a credit policy, or it could produce three different results. By normalizing data, banks can use “waterfalling” rules to access data from all three credit bureau sources. Data normalization ensures that scoring will work the same way, with the same predictive accuracy, regardless of whether the attribute is from the bank’s preferred or secondary credit bureau.

Normalizing techniques can also be used when two individuals apply jointly for a loan. They have shared accounts, such as a mortgage, in addition to their own unique credit histories. The bank can use normalization to identify what is common to both, and then create a single credit view of the couple.

Normalization is an arduous task for banks to perform internally – often banks will need to set up their specific attribution with each individual bureau. With an outsourcing approach, that burden is transferred from the bank to the provider, and a single set of attribute definitions can be managed in one place.

Multiple Applicants

Figure 8: Multiple applicants.

Applicant File

Trade A Balance: 5000 Open Date: 03/10/2005 Report Date: 02/12/2009Trade B Balance: 6000 Open Date: 03/10/2005 Report Date: 12/12/2010

Trade C Balance: 5000 Open Date: 03/10/2005 Report Date: 02/12/2009Trade D Balance: 6000 Open Date: 03/10/2005 Report Date: 12/12/2010

Co-Applicant File

Applicant Deduped File

Trade B Balance: 6000 Open Date: 03/10/2005 Report Date: 12/12/2010

Trade D Balance: 6000 Open Date: 03/10/2005 Report Date: 12/12/2010

Co-App Deduped File

Combined & Deduped File

Trade B Balance: 6000 Open Date: 03/10/2005 Report Date: 12/12/2010


Suggestion #3: Augment your legacy systemsIdeally, most banks would like to start afresh with an entirely new data access management approach. Strategically, the fresh-start approach has immense appeal. It can help enable the organization to achieve the full range of benefits discussed in this paper. However, banks need to move forward quickly; they want to tap the benefits now of alternative data and better solutions for modeling and analytics.

For most banks, this kind of “rip and replace” approach is impractical. As cumbersome as their legacy systems are, they work—and replacing them is not an easy, fast, or inexpensive task. Without completely reengineering their decisioning processes, banks need a sound way to get major incremental benefits soon from using alternative data and better solutions for modeling and analytics

In this case, Equifax suggests employing an augmentation strategy: Don’t throw out what you have. Make what you have work better. Bureau engine solutions can work with your existing interfaces to simply pull in additional data sources and package it in with your legacy formats. Minimizing the effort needed to take advantage of these within your existing policies. Users can then integrate newer information sources, along with derived attributes and scores, into their decision processes.

Banks can use the augmentation strategy to extend their general capabilities but leave legacy technology in place for now—as an interim tactical move before ultimately graduating to a more sophisticated approach to data access management. The augmentation strategy approach captures significant upside value until such time as the bank can phase out its legacy systems and phase in an enterprise-wide, centralized solution.

Large bank calls in Equifax to revamp its data access management

A top-30 bank wanted to gain insights into customers and make better lending decisions. To meet its growth and efficiency goals, the bank would have to overcome some daunting challenges:

•Theexpenseofsupportingandchangingmultipledataaccesssystemsservingdifferentlineofbusinessesandchannels•Theinabilitytoaddnewdatasourcestolegacysoftware•HeavyITinvolvementinchangemanagement•Lengthydelaysinaddingnewdatasources,attributesandscorecards

The bank implemented Insight Gateway™, a data access management solution from Equifax. According to the bank’s Strategic Analytics specialist, “With Insight Gateway, we are able to centralize attributes and access new data sources, and do advanced decisioning around credit data. Equifax provides full project management support. Using Equifax credit experts to set up this solution provided best practices around credit decisioning.”

Change management processes are now managed across all product lines, and data assets are ubiquitous across multiple legacy platforms. Costs for deploying a new score or model with Insight Gateway are estimated to be $250,000 less than similar deployments before the bank engaged Equifax. The bank’s time to market for new scorecards is much faster because business users can now design and configure attributes without involving a programmer for hard coding.

To view a video that further explains these best practices, please visit our Recorded Webinar


The Equifax approachBanks don’t have to face these challenges alone. Equifax has a highly experienced team of professionals to help advise customers on suggested best practices, recommend process changes, and help customers deploy the right solutions for their requirements. We are experts at helping customers integrate Equifax solutions with their existing legacy systems.

Equifax offers deep customer insight, including the Insight Score for Retail Banking™ solution discussed earlier. We also offer Insight Gateway, a powerful data access management system that provides a single, common interface for integration, attribution and scoring of disparate data assets used in decisioning models. Ask us to demonstrate how large banks are using Insight Gateway to help reduce the costs, delays, and hassles of adding new data and analytics into the bank’s automated decisioning processes. Some banks can achieve an estimated 75% efficiency improvement from using Insight Gateway to augment their existing data access solutions.

SummaryTo empower their business users with better data, banks need to deploy best practices around data integration, decisioning, offer management, fraud mitigation and compliance. When new modeling and analytic techniques are supported with the new data assets, modelers can help deliver more predictive and powerful intelligence to risk and product teams.

The new solutions centralize access to the latest data and analytics, so mission critical systems can deliver deeper customer insight across the enterprise. This helps enable business users, modelers and risk analysts to:

• Gainimprovedcustomerinsightviafastaccesstonewdataandanalytics• Improveriskmanagementbydeployingnewmodelsandscoresmorequickly• Reduce IT workload by streamlining complex data management and

compliance tasks

It is a certainty that automated decisioning will become more complex and more regulated. Progressive, growth-oriented banks are taking advantage of new data sources, which are constantly being introduced. And banks are upgrading their data asset management systems to accommodate centralized libraries of reusable attributes. Ultimately, a bank that embraces be practices for data access management will achieve smarter growth.

Equifax can help you prepare for success. To learn more, please visit http://www.equifax.com/technology/acquire-customers/acquire-insights/en_us


About the authors

Lee Grice

Lee Grice is Assistant Vice President of Product Marketing and Sales Enablement for Equifax Technology and Analytical Services. He is responsible for developing and executing the marketing strategy for Equifax’s portfolio of application processing, credit risk, and cross-sell products. Lee has 25 years of industry experience in software development, technology marketing and product management.

James Blue

James Blue is a product manager for Equifax Technology and Analytical Services. He is responsible for developing multiple products in the InterConnect suite of solutions. James has deep experience providing data and decisioning software solutions to businesses. James has over 10 years of industry experience in both B2B and B2C software product management.

There are references throughout this publication/website to various trademarks or service marks and these, whether registered or not, are the property of their respective owners. Equifax is a registered trademark of Equifax Inc. Inform > Enrich > Empower is a trademark of Equifax Inc. Insert all other trademark, registration, text here. Copyright © 2011. Equifax Inc., Atlanta, Georgia. All rights reserved.

For more information, please contact:[email protected]

Contact Us Today


EFX-ABC-0000-05/28/13

www.equifax.com

Last Mile Execution · PDF fileLast Mile Execution Technology solutions that deliver deeper...

Documents

Transcript of Last Mile Execution · PDF fileLast Mile Execution Technology solutions that deliver deeper...