SAS Data Management Technologies Supporting a Data Governance Process Dave Smith, SAS UK & I
Agenda Data Governance What it is Why it s needed How to get started SAS technologies which can assist Data Governance programs
Data Governance Defined Formal orchestration of people, processes and technology to leverage data as a corporate asset
Why Govern Data? Regulation Risk Efficiency Opportunity
The interconnectedness of people process and technology Business Technical Owner Relationships Definitions Links Status Importance Requirements Attributes Data Item Physical Location Metadata Lineage Quality
The Data Governance Journey ORCHESTRATING PEOPLE, PROCESSES, AND TECHNOLOGY Unpredictable Managed Controlled Proactive Integrated No awareness Nascent awareness No accountability and Data Stewards Assigned DS Accountability of business owners Proactive behaviours on DQ Recognized DS DG office DQ culture No common language Application-centric Business definitions DQ processes Shared business definitions and rules DG policies DQ embedded in processes Optimised crossfunctional processes No tools DQ tools Metadata repositories Shared metadata Business glossary Analytical MDM DG process tool and dashboards Operational / collaborative MDM Data as a service
Data Management vs. Data Governance Data management is a by-product of data governance Effective data management needs to be governed
Data Governance Buying Warehouse SystemWho owns Mgt the System data? Who can author data and how? New Product Introduction Who can decide Cons. about the Customer Marketing changes? Exp. THE QUESTIONS IT ADDRESSES Customer Registration Emailing Promotions / Marketing What does good POS Campaign data look like? Mgt Digital Marketing Finance & Risk Pricing WEB How is inaccurate Promotion Management Marketing Campaign Call How are conflicting Market Centre needs addressed? insight Disparate needs for data consumption CRM / Loyalty Program Data silos / Application Centric Data Generation information & Manipulation corrected? Unmanaged cross-functional processes
The role of Data Stewards ORCHESTRATING CROSS-FUNCTIONAL COLLABORATION Data Stewards Manage & Monitor Business Users Create & Consume IT Implement, Adapt & Extend
Common Data Governance Challenges Seen as an academic exercise The culture doesn t support centralized decision making Considered an IT issue The ROI isn t clear Definitions and explanations of data governance are varied and contradictory Nervousness about the G word
SAS Data Governance Framework
Where to Start? Top-down Quick Wins Vision & Roadmap Business case / ROI Data Dictionary definition DQ Standards definition Data Stewardship model Prioritization of DM initiatives Organizational framework DG & DQ Processes Impact & root cause analysis Data Quality Analysis Bottom-up
Other Best Practices Understand what s important to management now Work within your culture Understand your current state before making the pitch Choose sponsors based on initiative owners Corrections at source & available to real time processes Treat Data Governance as a project Rely on the big-bang approach Treat all data the same way
SAS Technologies for Data Governance
SAS Data Management Platform
Data Quality Process Update & Improve systems and processes Define the key entities Identify the sources and responsibilities 6 Remediate & Improve Operations and DI Experts DQ Analyst 1 Define the terms and sources Business Owner Measure & Monitor actual vs. expected, identify trends, allocated tasks Qualify & Quantify actual issues with the Data 5 Monitor & Publish DQ measurement Data Steward Business Owner 2 Discover & Profile the Data DQ Analyst Embed the DQ services and business rules into the operating systems and DI processes 3 Design the business rules to enforce data quality and data services Operations and DI Experts 4 Apply injection and execution Design data quality standards Business Owner DQ Analyst
The Relationship service The Relationship Service collects and stores metadata Content from SAS and sources outside of SAS Processes that include resources used in data management, business intelligence, and data integration Consists of Resources and Relationships Resources are metadata representations of data assets or processes Relationships describe how two Resources are related
Relationship Types Is dependent on Is parent of Contains Is synonymous with Is associated with Is equal to
Lineage Viewer Acts as a viewer on the relationships database Allows different views of data lineage including governance and impact analysis
Business Data Network Central definitions of Terms across the organisation Links business and technical definitions to enable collaboration and clarity
Federation Server Create federated views of data from diverse sources Apply row and column access control, data encryption and masking to sources Enable detailed logging of data access
Questions? www.sas.com