2. Agenda
• Demand for MDM
• Approach
• Success to date
• Lessons Learned
2MDM & DG Summit - SF 20158/17/2015
3. Agenda
• Demand for MDM
• Approach
• Success to date
• Lessons Learned
3MDM & DG Summit - SF 20158/17/2015
4. Operational Data
LOA’s & Agency
In 2010 my client had many disconnected views of Customer:
Updates in one system did not replicate to other systems
≠
Billing Data
≠ ≠
≠
≠
4MDM & DG Summit - SF 20158/17/2015
5. Future State: A single version of the truth, providing
multiple, consistent views of customer
Customer Master Information
Data Governance – (Cleanse, qualify, process, synchronize, enforce)
Customer Profile
Data Receivers Agency
Billing
Customer History
Revenue & Price Terms
Invoice & Volume Cases/Incidents
Dashboard Views
Committee
5MDM & DG Summit - SF 20158/17/2015
6. 4 Major Themes for 2011
1. Company & Customer Sync
2. Data Governance
3. Agency Management
4. Dashboards
SSO
Industry
reference
Financials
Ticket
Tracking
6MDM & DG Summit - SF 20158/17/2015
7. Master Data Management – Customer Focus
Customer Master Information
Data Governance – (Cleanse, qualify, process, synchronize, enforce)
Customer Profile
Billing
Data Consumers Data Providers
Prospects
Customer History
Revenue & Price Terms
Volume Cases/Incidents
Dashboard Views
• SSO
• Industry reference
• Financials
• Case Tracking
• SP Community
• Agency Relationship
Customer
Profile
• Sales
• Product Mgmt
• Customer Support
• Product Support
• Billing
Dashboard
View
Customer
Relations
• System 1
• Active Directory
• Contracts
Customer
Profile
• Integrate XSalerator
• Lead Management
• Demand Generation
• Lead Nurturing
• Opportunity &
Pipeline Management
• Executive View
• Custom Reporting &
Analysis
Reporting
Phase 1
Phase 2
Phase 3
• Lead Nurturing
• Forecasting
• Campaign Mgmt
• Marketing View
Sales
Automation
7MDM & DG Summit - SF 20158/17/2015
8. Agenda
• Demand for MDM
• Approach
• Success to date
• Lessons Learned
8MDM & DG Summit - SF 20158/17/2015
9. Challenges
• What is MDM?
• Limited funds spread over three years
• Are the four systems the only ones?
9MDM & DG Summit - SF 20158/17/2015
10. Approach
• What’s MDM?
– Limited experience with data modeling in organization
• Bring in consultants to help us size the problem and lay the initial plan and
develop initial data models
• Provide training to staff
• Industry standard party model used
• Limited funds
– Prevented purchase of MDM tools with the exception of a data quality tool
– Used labor to map systems, load data, etc.
– Selected SaaS CRM solution for dashboard
• Only four systems?
– No, many more
10MDM & DG Summit - SF 20158/17/2015
11. Four Major Functions in MDM
• Collect
• Cleanse
• Control
• Use
11MDM & DG Summit - SF 20158/17/2015
12. Customer Master Data
Consolidation• Collect:
– Identify data owners
– Develop logical and physical models
– Profile data
– Map the data
– Extract and transform data
– Load Customer master data base repository
• Cleanse:
– Clean data
– Provide suggested changes to source systems
• Control:
– Synchronize with systems
• Use:
– Provide access and functionality through new CRM solution
– Provide web services
12MDM & DG Summit - SF 20158/17/2015
13. Are these the only four systems? No!!!
Data below for Company information
13MDM & DG Summit - SF 20158/17/2015
14. Are these the only four systems? No!!!
Data below for People information
14MDM & DG Summit - SF 20158/17/2015
15. Master Data Management Model
Party
Person
Employee
Non Employee
Organization
Internal Organization
External Organization
Contact
Method
Party
Relationship
Party Relationship
Role
Party
Address
Party
Contact
Method
Product
Party Product Role
Party
Identification
Invoice
LOA
Order
15MDM & DG Summit - SF 20158/17/2015
Address
16. Customer Master Data
(Golden Record)
Applications
The Benefits
Company
File
Industry
Reference
SSO
Financials
Ticket
Tracking
Other
Apps
Data Access
Layer
16MDM & DG Summit - SF 20158/17/2015
18. How we handled Data Quality issues
• Enough funds to purchase Informatica DQ
• Determined not to update directly the source systems
– Work through the data stewards
– Automate where possible
– Built routing system to handle notifications
• Secured additional funding for one-year DQ project
• Purchased reference geographic data to handle address
issues
18MDM & DG Summit - SF 20158/17/2015
19. Built access to data
• For dashboard and related use, implemented
cloud-based CRM solution
• For OLTP access, Teradata not ideal
– Built Oracle database
– Provided services layer to access data
– Implementing access by systems as possible
• SSO accessing company information
– Data still stored in SSO
– Long term may replace TD portion
19MDM & DG Summit - SF 20158/17/2015
21. Agenda
• Demand for MDM
• Approach
• Success to date
• Lessons Learned
21MDM & DG Summit - SF 20158/17/2015
22. Successes
• Program ongoing
– Eight systems feeding daily into MDM
– Nightly exception routing sends exception information to data
stewards/systems for processing
– Oracle store and services layer built
• Services key to integrate with other systems
– Two silos prevented
– Other areas asking for help
– New roles developed: data owners and stewards, data quality staff,
data architects
– DQ COE organized
• CRM system viewed as key win
22MDM & DG Summit - SF 20158/17/2015
23. Challenges
• Where’s the value?
• Many more source systems to address
• Eliminating silos
• New subject areas such as location
23MDM & DG Summit - SF 20158/17/2015
24. Agenda
• Demand for MDM
• Approach
• Success to date
• Lessons Learned
24MDM & DG Summit - SF 20158/17/2015
25. Lessons Learned
• Not only an IT problem!
– Need active participation from the business side
• Matching models a tough challenge
– Need a data architect. Period.
• There are more source systems than you expect
– Pick your battles and Federation is not bad
• Look for ways to demonstrate value
– And deliver it. Preferably, find ways to generate revenue
• If starting now, consider open source solutions such as Talend
• It is a program not a project: needs continuous attention
25MDM & DG Summit - SF 20158/17/2015
26. How to Do MDM on a Budget?
A Simplified View
• Scope the problem
– Sources of data
– Data models
• Secure expertise if not in-house
• Identify tools and assess WRT budget
– MDM
– ETL
– DQ
• Implement & monitor
• Sell value 26MDM & DG Summit - SF 20158/17/2015