Sound Data Quality for CRM

<Insert Picture Here>

Sound Customer Data Quality for CRMManoj Tahiliani, Senior Manager, Customer Hub Strategy

The preceding is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions.The development, release, and timing of any features or functionality described for Oracle’s products remains at the sole discretion of Oracle.

<Insert Picture Here>

Agenda

• Data Quality – Pains, Drivers and ROI• Siebel Data Management Solution• Data Quality Products• Best Practices• Oracle Credentials

“Data quality is a Business Issue”

• Virtually all enterprises are experiencing a significant amount of pain directly attributed to data quality issues.

• Significant amounts of wasted labor and lost productivity translate into direct financial losses to the business.

• Some enterprises that have measured the impact have found they are losing multiple millions of dollars each year as a result of poor data quality.

Companies

Velocity of Data Change is Staggering

• 240 businesses will change addresses

• 150 business telephone numbers will change or be disconnected

• 112 directorship (CEO, CFO, etc.) changes will occur

• 20 corporations will fail

• 12 new businesses will open their doors

• 4 companies will change their name

Source: D&B, US Census Bureau, US Department of Health and Human Services, Administrative Office of the US Courts, Bureau of Labor Statistics, Gartner, A.T Kearney, GMA Invoice Accuracy Study; 2 Data: An Unfolding Quality Disaster, Thomas C Redman, DM Review Magazine August 2004, Mintel Global New Products Database (GNPD), 2007. CNNMoney.com 2006, 3Quality is Free, Philip Crosby

• 5,769 individuals in the US will change jobs

• 2,748 individuals will change address

• 515 individuals will get married

• 263 individuals will get divorced

• 186 individuals will declare a personal bankruptcy

Individuals

“If bad data impacts an operation only 5% of the time, it adds a staggering 45% to the cost of operations.”2

“Poor data quality cost business’ 10% to 20% of revenue!”3

Change of Circumstances

• 4.7 Million Marriages• 1.53 Million First Births • 2.04 Million First-time

Home Buyers• 1.9 Million Divorces• 43 Million Residential

Moves• 1.4 Million Work

Retirements

In one hour… In one hour… In one year…

Master data changes at rate of 2% per month.

7 Questions About Your Data

1. Have data initiatives failed or been delayed due to unreliable data?

2. Do you always deliver the right product to the right customer?

3. How many marketing pieces are un-delivered or un-answered?

4. How much time is spent in reworking inaccurate data?

5. Do you face difficulties with regulatory compliance?

6. Is customer satisfaction going down?

7. Do you distrust your data to take critical decisions?

Poor Data Quality is the #1 enemy of CRM Solutions

Out of Date

Rapid changes in a dynamic society: marriages, divorces,

births, deaths, moves

Garbage

Typos, misspellings, transposed numbers, etc.

Fraud

Purposeful misrepresentation of data:

identity theft, wrong information (bankruptcies, occupation, education, etc)

Missed Opportunities

Information that we do not know about (customer

relationships, up-sells, cross-sells)

IT Agility

Ineffective Cross-sell/Up-sell

Lower call center productivity

Increased marketing mailing costs

Reduced CRM adoption rate

Customer Service

Increased data management costs

Increased sales order error

Delayed sales cycle time (B2B)

Mediocre campaign response rate

Operational Efficiency

Risk, Compliance Management

Increased integration costs

Increased the time to bring new projects and services to market

Proliferation of data problems from silos to more applications

Heightened credit risk costs

Potential non-compliance risk

Increased report generation costs

Measuring actual ROI achieved

Example of Customer Data Quality IssueA Simple Customer Table Sample

Name Address City State Zip Phone Email

Bob Williams 36 Jones Avenue Newton MA 02106 617 555 000 [email protected]

Robert Williams 36 Jones Av. MA 02106 617555000

Burkes, Mike and Ilda 38 Jones av. Nweton MA 02106 617-532-9550 [email protected]

Jason Bourne, Bourne & Cie.

76 East 51st Newton MA 617-536-5480 6175541329

… … … … … … …

Mis-fielded data

Matching Records

TyposMixed business and contact names

Multiple Names

Non Standard formats

Missing Data

20 Common Errors & Variation (1)

Variation or Error Example

Sequence errors • Mark Douglas or Douglas Mark

Involuntary corrections • Browne – Brown

Concatenated names • Mary Anne, Maryanne

Nicknames and aliases • Chris – Christine, Christopher, Tina

Noise• Full stops, dashes, slashes, titles,

apostrophes

Abbreviations • Wlm/William, Mfg/Manufacturing

Truncations • Credit Suisse First Bost

Prefix/suffix errors • MacDonald/McDonald/Donald

Spelling & typing errors • P0rter, Beht

Variation or Error Example

Transcription mistakes • Hannah, Hamah

Missing or extra tokens • George W Smith, George Smith, Smith

Foreign sourced data • Khader AL Ghamdi, Khadir A. AlGamdey

Unpredictable use of initials

• John Alan Smith, J A Smith

Transposed characters • Johnson, Jhonson

Localization • Stanislav Milosovich – Stan Milo

Inaccurate dates• 12/10/1915, 21/10/1951,

10121951, 00001951

Transliteration differences • Gang, Kang, Kwang

Phonetic errors • Graeme – Graham

20 Common Errors & Variation (2)

Two Facts about Data Quality

• The Data Quality Challenge is an iceberg• The biggest DQ threats are the ones we do not see.Data Profiling lowers the water line and draws a clear view

of the quality issues

• Data value decays• Data is an asset which value decays over time• Business events can make this worse

• M&A, new applications, new products, new contact files, etc

• Quality is not a one shot process but a constant effort in the enterprise processes.

Data Quality needs to be pervasive and continuous.

<Insert Picture Here>

Siebel Data Management Solution

Data Management - Deployment Options

Middleware Application Integration Architecture

Web site

Call Center

SFA PartnerFusionApp

FusionApp

Call SCM

ERP2 LegacyERP 1

MDM

Middleware Application Integration Architecture

FusionApp

Call SCM

ERP2 LegacyERP 1

PartnerData Mgmt Layer

CRM

Trusted Customer Data

Trusted Customer Data

Components of Siebel Data Management

Web ServicesLibrary

Web ServicesLibrary

Publish &SubscribePublish &Subscribe

Transports &Connectors

Transports &Connectors

AuthorizationAuthorization

RegistryRegistry

Profile & Correct

Profile & Correct

History& AuditHistory& Audit

PrivacyMgmt

PrivacyMgmt

Events & Policies

Events & Policies

Import WorkbenchImport Workbench

Identification & Cross-Reference

Identification & Cross-Reference

Source DataHistorySource DataHistory

SurvivorshipSurvivorship

ParseParse

Cleanse & StandardizeCleanse &

StandardizeEnrichEnrich

Manage Decay

Manage Decay

Match & Merge/ Unmerge

Match & Merge/ Unmerge

Roles & Relationships Party

Vertical VariantsRelated Data

Entities

Roles & Relationships Party

Vertical VariantsRelated Data

Entities

Hierarchy ManagementHierarchy

Management

<Insert Picture Here>

Data Quality Products

Data Quality Functionality in a Glance

ProfilingProfiling

CleansingCleansing

MatchingMatching

EnrichmentEnrichment

Understand data status, deduce patterns

Tel# is null 30% LName + FName (Asian Countries); FN+MN+PN+LN (Latin);

Addr = #, street, city, state, zip, country; St, Str = Street (ENU/DEU);

Spot and correct data errors; transform to std format/phraseIdentify and eliminate duplicates

Haidong Song = 宋海东 =

Attach additional attributes and categorizations

Haidong Song: “single, 1 child, Summit Estate, DoNot Mail”

Functionality Customer Data example

Comprehensive

data quality

Comprehensive

data quality

Feature

Batch and Real-time

New Data Quality Products

Introducing New Products to provide full spectrum of information quality functions:

• Oracle Data Watch & Repair• Ongoing Discovery of Actual state of Master Data

• Data Governance

• Oracle DQ Cleansing Server: • ASM (Address Standardization Module)• Integrated single engine– supports all countries

• Oracle DQ Matching Server: • Full Administration Access and increased level of support

• Improved performance and enhanced tuning capability

New Data Quality Products

Matching EngineData Quality

Matching Server

Data Quality

Cleansing Server

Administration UI / Rules Editor

Improved performance

18 Languages

52 Languages

Address Standardization Module

240 Languages support

Data QualityProfiling

Profiling Console & Engine

OldOffering (SSA)

NewOffering

ProfilingProfiling

CleansingCleansing

MatchingMatching

EnrichmentEnrichment

Comprehensive

data quality

Comprehensive

data quality

Oracle Data Watch and Repair• Ongoing auditing prevents data

decay, ensures continuous quality• Non intrusive profiling across

existing applications/databases • Quickly narrow in on anomalies• Generate rules to repair problems• Edge Application (no Upgrade

impact)• Out of the box connector to Siebel

CRM

ProfilingOngoing Discovery of State of your Data

• Advanced Validation and Standardization of addresses in more than 240 countries

• Scalable high performance• Integrated single engine– supports all countries• Edge application (no upgrade impact)

ProfilingProfiling

CleansingCleansing

MatchingMatching

EnrichmentEnrichment

Comprehensive

data quality

Comprehensive

data quality

>240 Countries

One API

Oracle DQ Cleansing ServerStandardize & Validate against References

Proven Performances•Not just number of records but also volumes of Txns

•In use on systems with > 800 million records

•> 250,000 txn/hour on large credit systems

•> 1.5 million txn/hour on screening app

•11,000 million index entries on one database

•30,000,000 real time transactions in an hour

Flexible & Adaptive•Smart indexing & fuzzy logic to emulate expert reasoning•Highly configurable•Edge application (no upgrade impact)

Unprecedented Global Coverage

•52 Languages/locales•Cross script matching

Oracle DQ Matching ServerRecords linked to Same or Related Entity

ProfilingProfiling

CleansingCleansing

MatchingMatching

EnrichmentEnrichment

Comprehensive

data quality

Comprehensive

data quality

Hybrid Algorithm Industry’s Best Matching Technology

Heuristic

Probabilistic

Deterministic

PhoneticLinguistic

Empirical

• Best Solution: Hybrid• “Which algorithm is the best in solving

my searching and matching needs?” • The answer is “No single algorithm is

capable of compensating for all the classes of error and variation present in identity data.”.

• In order to achieve a consolidated view of your identity data, you will need a combination of these algorithms, and more, each one addressing a particular class of problem,

• Oracle Matching Server uses a variety of techniques, including the six mentioned here and many more, to address different classes of error and variation in identities

Oracle Data Quality Matching Server

Siebel UCM / CRM

ApplicationObject ManagerUser Interface

Data Admin

Oracle DQ Matching Server

Loader & Utilities

Rule Manager

Key & SearchStrategies

MatchPurposes

Search Server

Update Synchronizer

Console Server

Console

AdministrativeClients

PopulationOverride Mgr

Edit RuleWizard

• Acxiom, D&B Integration

Data EnrichmentAdd Details from External Sources

SCM

Marketing

Web site

legacy

Call Center

SCM

SFA

Acxiom Knowledge Base

Batch or Interactive Delivery

Acxiom Customer Data Integration Services

Clean Recognize Enrich Protect

Integration LayerIntegration Layer

Oracle MDM Schema

Oracle MDM Web ServicesO

racle

M

DM

S

erv

ices

ConsolidateCross

ReferenceAudit & Control

Manage Events GovernPublish

Integration LayerIntegration Layer

ProfilingProfiling

CleansingCleansing

MatchingMatching

EnrichmentEnrichment

Comprehensive

data quality

Comprehensive

data quality

Prospect Mastering with Knowledge-Based MDM

Perform segmentation within Siebel Marketing application

Generate prospect selection criteria

Campaign Planning

Load selected prospect records into Oracle MDM-CDI solution

Consolidate existing customer info with prospects from other sources

Oracle EBSAcxiom/D&BAcxiom/D&B

DataDataProductsProductsMDM-CDIMDM-CDI

SiebelSiebelMarketingMarketing

Load

Loading & Matching

Siebel CRM On Demand Plug & Play Market

Campaign Execution

Campaign Execution

Send criteria and list of existing cust/prospect to Acxiom/D&B etc

Acxiom/D&B produces the net new prospect list and send to customer

Contact informationDemographic dataWealth/income classificationsSegmentation groupingsLifestyle indicators

Prospect Acquisition

Next Generation Data Quality

• Best of Breed Data Quality

• Matching – uses “fuzzy” logic and a unique two-stage approach to overcome the limitations of traditional techniques for 52 languages

• Cleansing – Contains postal address information for 240 countries and territories

• Profiling - discovers the quality, characteristics and potential problems of source data

• Enrichment – integrate with 3rd party content providers for business & consumer data

Embedded best in class Data Quality Open framework & connectors

• Universal DQ Connector

• End to end connector available for selected vendors

<Insert Picture Here>

Best Practices

LeadershipLeadership

Policy DefinitionPolicy Definition

Planning and CoordinationPlanning and Coordination

Execution and Decision-MakingExecution and

Decision-Making

Compliance Monitoring and

Enforcement

Compliance Monitoring and

Enforcement

Master Data

Data Management Governance

Record Definition

Data Quality Assessment

Initial Data Quality and Load

Ongoing Data Cleansing and Conversion

Data Management Processes

• Central executive leadership• Enterprise steering committee

to arbitrate issues and enforce the rules

• Coordination and compliance• Define & communicate data

quality expectations • Establish policies, procedures,

success metrics and processes to maintain quality data

• Identify all business and application stakeholders across the enterprise – data owners

• Conduct audit and control• Communication and change

management

Formalize a Governance Framework

Closed Looped DQ

A Day in the Life of a Data StewardData Stewardship is a critical component of DQ Process

1. Runs profiling routines to monitor overall DQ within application• Inspects most crucial or known problem areas

• Gains deep-level understanding of data (e.g. min, max, # nulls..)

2. Creates and applies new data rule based on profiling results

3. Resolves duplicates and creates links

4. Reviews history and audit trail

5. Defines compliance rules and policies

6. Defines event and policies for ongoing monitoring and management

7. Executes corrective action: recover, unmerge, etc.

8. Performs ongoing monitoring of data quality

• Information Completeness– Do we have complete profiling information for our accounts / contacts?– Where are the information holes?

• Information Validity– Does the customer have valid address, phone number and email?– Have we been able to communicate to the customer using stored contact

point information?

• Information Uniqueness (Duplication)– What is the duplicate rate in our accounts and contacts? What is the trend

over time? – Which systems creates the most duplicates?

• Information Accuracy– Is the information still up to date– Does the information have the proper integrity based on available sources

and/or defined business rules?

Data Quality Scorecard

<Insert Picture Here>

Credentials

Case Study - Lead Telco

CHALLENGES / OPPORTUNITIES• Drive improved customer experience &

satisfaction• Consolidate customer information from

disparate systems and multiple lines of businesses

• Improve customer data quality• Complete understanding of customer

hierarchies and relationships

SOLUTIONS – Oracle MDM & Data Quality• Enterprise wide customer master to provide

a single view of customer• Match, deduplicate, and consolidated

customer information from multiple systems into the customer master

• Built out customer hierarchies and relationships

RESULTS• Consolidated ~30 mil customer

records from 10+ applications into customer master

• Improved customer data accuracy and completeness

• Provided consistency and integrity of data across multiple operational systems

Selected Oracle Data Quality Customers

Human Resources