Informatica's Velocity Methodology

Data Governance
Description
Although its value is not represented on the balance sheet, data is one of the most important assets in an organization. Data represents an organization's customers, employees and suppliers; its activities and transactions; and its outcomes and results. Managed correctly data can become an organization's most valuable asset; helping it to remain competitive and agile, to proactively meet customer needs and to keep costs in check. As a practice with roots in corporate and Information Technology (IT) governance, data governance is defined as the processes, policies, standards, organization and technologies required to manage and ensure the availability, accessibility, quality, consistency, auditability and security of data in an organization.

Business Drivers and Objectives

The most common business drivers for data governance initiatives are: Growing revenue Lowering costs Ensuring compliance Mergers and Acquisitions (M&A) Partnering and Outsourcing While these are among the basic goals of any business, it is useful to examine them in more detail and review the integral role data governance can play in achieving these goals.

Growing Revenue
One of the most important goals of almost any business is to grow revenue; and one of the most effective ways to grow revenue is to increase cross-sell/up-sell rates and improve retention among existing customers. To do so, organizations need a broad and deep understanding of their existing customers. They need a "single view of the customer" in order to be able to provide superior service and to better target campaigns and offers based upon a specific customer's needs. Customer data is often scattered across dozens or even hundreds of different business systems. To resolve these data issues, companies must address the underlying organizational,

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

process, and technical issues related to the data. Data governance provides the framework for addressing complex issues such as improving data quality or developing a single view of the customer at an enterprise level.

Lowering Costs
While the pressure to lower costs is not as intense as it has been in recent years, increasing operational efficiency is still a major priority for most organizations. One of the important ways organizations can reduce costs and increase operational efficiency is to automate business processes. For example, organizations may automate their procurement processes to lower purchasing and administration costs. While business process automation increases efficiency, problems with enterprise data prevent companies from capitalizing on the full potential of operational efficiency initiatives. Streamlining business processes across multiple financial, human resource, sales and other business systems requires that the structure and meaning of data be reconciled across those systemsa task that has often been an afterthought in operational efficiency initiatives. The need to lower costs is driving projects such as supplier or product master data management that enable companies to streamline core business processes (e.g., inventory and supply chain management) by rationalizing, cleansing, and sharing key master data elements. Data governance plays a critical role in the success of such projects, providing a structure for addressing the organizational and process issues around master data.

Ensuring Compliance
Doing business today requires compliance with a growing number of external regulations as well as with internal corporate governance policies designed to increase transparency and prevent corporate malfeasance and fraud. To ensure compliance with regulations such as Sarbanes-Oxley, Basel II, the U.S. Patriot Act, and the U.S. Health Insurance Portability and Accountability Act (HIPAA) and with internal policies and controls, companies must streamline the collection of reporting data. For many regulations they must also document the sources of the data being reported, certify its accuracy and implement specific governance policies. When it comes to how it handles its data, complying with these regulations and policies can be a burden to a company. Data governance is an essential foundation for ensuring compliance. It establishes the rigorous data standards, policies and processes that are required by regulations and corporate governance policies, and it helps to automate compliance-related tasks (while lowering costs). It also helps to ensure auditability and accountability for the data.

Mergers & Acquisitions

As M&A activity picks up, organizations are faced with the need to rationalize and reconcile the IT environments from merged or acquired entities. Typically these IT environments have very different systems, data models and business processes. Post-M&A, IT organizations are often pressed to meet very tight timelines for integration. The goal is to accelerate the promised synergies from the merger, both in the form of cost reductions from eliminating redundancies

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

and revenue growth from increased cross-selling. The process of migrating and consolidating the data after a merger or acquisition is a huge task one that is often underestimated initially. IT groups must deal with unknown systems, resolve quality issues and provide detailed documentation on how the information has been merged. And the task involves much more than technical integration. IT organizations must not only reconcile different data definitions and models, but processes must be put in place to ensure alignment of the various entities. A data governance framework provides significant value in managing the organizational and technical complexities of M&A consolidation and accelerating positive business results.

Partnering and Outsourcing

Another broad market trend is the increasing use of partners and outsourcers to manage parts of the value chain. Organizations are focusing on core competencies and handing off non-core functions and processes to partners and outsourcing providers. Some examples of these include: High tech equipment companies rely on contract manufacturers for production. Manufacturers turn to UPS and FedEx for logistics and warehouse management. Pharmaceutical companies rely on third-party clinical trials management firms. IT departments outsource application development and network management. HR groups outsource administrative functions such as payroll or benefits management. As business processes and IT systems shift to outside providers, the data associated with those processes and systems is relocated outside the boundaries of the organization. Organizations must ensure that the data is correctly migrated to the outside provider. The data must be complete and accurate and it has to be restructured to work in the third-party system. It is important to note that although the data has been moved to a third party, it remains a core asset of the organization. Even though it sits outside the firewall, the organization cannot relinquish visibility into and control over that data. A robust data governance framework is critical to managing data that is fragmented across the extended value chain; especially when it comes to defining the standards and processes for interaction and collaboration with external partners and outsourcers.

Industry Perspective
To understand how data governance is evolving as a practice today, let's review its roots in the broader context of corporate and IT governance.

Corporate and IT Governance

As an established discipline for many years, corporate governance can be defined as the policies, procedures and rules that affect the way the organization is controlled and administered. One of the practice areas under corporate governance is IT governance. IT governance focuses on the IT processes and structures that provide accountability and ensures that the organizations overall strategies and objectives are followed. Like corporate

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

governance, IT governance is also a mature discipline. Industry bodies like the IT Governance Institute and International Organization for Standardization (ISO) have established detailed IT governance standards and processes that many companies and government organizations have put into practice. However, many of these IT governance practices have evolved in a pattern based on common IT organizational structures and deployment architectures which traditionally have centered around software applications (e.g., Enterprise Resource Planning (ERP), Customer Relationship Management (CRM)). From mainframes to client-server architectures to Web applications, IT departments have focused on the applications and the user interfaces to those applications, rather than on the data that drives the applications. Typically, data is managed in the context of how it serves a particular application rather than how it serves the strategic interests of the overall organization. As a result, IT governance practices are designed primarily around an organization's applications not its data. In most organizations there are clearly defined owners, processes and policies in place to manage enterprise software applications. For example, when a company implements a complex and far-reaching ERP system like SAP, the SAP implementation is typically managed and governed by an SAP program office with clearly delineated responsibilities for managing and developing the SAP environment; both from a business and from an IT perspective. However, many organizations lack an equivalent governance structure for their most valuable core data assets such as customer, employee or product data. Additionally, due to the IT driven nature (as noted above) organizations often lack the collaboration between IT and business. If a shift to a business process-driven approach were to occur within the organization, it would ensure both the collaboration of IT and the business and the recognition of the context and usage of the data.

The Rise of Data

An application-centric approach to IT governance no longer serves the best interests of today's rapidly evolving business and IT environment. As shown in the figure below, rather than force-fitting business processes to the constraints of existing applications or systems, organizations are striving to align their IT systems with their business processes. As most business processes span multiple applications, IT groups as well as software vendors are busily decomposing applications into components or services that can be mixed and matched in a way that is flexible enough to support business processes. Moreover, it is becoming increasingly clear that enterprise data such as customer, employee or product data also needs to be shared across multiple applications and business processes, rather than being tied to one specific application. To get the most value out of their enterprise data, organizations need to pay greater attention to the data itself and how to manage it as an asset.

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

Implementation Approach
At the highest level, the recommended best practice is to use a 4-step implementation approach: 1. Understand the organization's overall business objectives and related data challenges. 2. Select one high-impact business initiative and build a data governance pilot project around it. 3. Evaluate the data integration technology's ability to support a data governance program. 4. Launch a formal data governance program based on a common framework and methodology. Step 1 - Identify the critical business objectives for the organization, understand the related requirements for data and review the data challenges that act as current or potential future barriers to achieving those objectives. Are there any regulatory or compliance requirements that demand immediate action? Do users have difficulty accessing or finding data? Do managers have the data they need to make decisions? Is the data consistent across different functions and business units? Can the validity of the data be certified? If there are data challenges, document the extent of the issues and the estimated impact on the business. This information can provide the justification for investing in a data governance program. Step 2 - After understanding the organizations overall business objectives, select one key business initiative on which to focus. This initiative should be one that poses significant data challenges, but also is expected to have a visible impact on the business. A pilot project for data governance (tied to this business initiative) will increase the likelihood of success for the business initiative, while also proving out the business value of a data governance program. The pilot project is also a good opportunity to build a strong collaborative relationship between business and IT groups and to gain executive sponsorship. Step 3 - Evaluate the existing data integration technology infrastructure and its ability to support data governance practices. While the technology by itself wont assure the success of a data governance program, it is an essential enabler for all six of the key data attributes accessibility, availability, quality, consistency, auditability and security. To support a data governance program in a scalable and consistent manner across the entire organization, the organization needs a unified enterprise data integration platform that offers: Broad access to all enterprise data, regardless of type, structure, or source from mainframe and midrange systems to XML documents and spreadsheets.

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

An open and platform-neutral architecture designed for ever-changing, heterogeneous IT environments. A single unified architecture to simplify and accelerate the entire data integration lifecycle. Enterprise-class security, scalability, reliability and availability. An approach based on shared, reusable components and services for transparency, interoperability and flexibility. An enterprise data quality strategy. If the existing technology infrastructure does not ensure that the enterprise data meets the six key attributes and is therefore not ready to support the data governance practice, the pilot project is a good opportunity to begin implementing a new and more robust data integration technology foundation. Step 4 - Implementing a data governance program requires addressing the organizational, process and policy concerns. Once a decision has been made to launch a formal data governance program, the following section will provide elaboration on the steps needed to establish and sustain it.

Data Governance Methodology

The figure below shows the 10 step methodology for initial implementation and ongoing execution of a data governance program. This methodology is powerful and flexible enough to support a range of enterprise strategies, including: Data Integration for Customer Data Integration, Product Data Integration or other similar Master Data Management Strategies. Data Governance, Risk and Compliance initiatives to address regulatory needs or corporate privacy, security or confidentiality policies. Service Oriented Architecture from a business perspective (as opposed to a technology perspective). Data Rationalization in support of M&A or legacy consolidation initiatives. Enterprise-wide data quality programs.

Some of the key features and advantages of this methodology in comparison to other industry methods are:

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

It is a practical methodology based on 15 years of experience in applying enterprise architecture, program management and data integration methods. The methodology explicitly recognizes the need for enterprise perspective from an ongoing needs basis and not just from a program or project perspective that has a finite scope and time-frame. The methodology recognizes that evolutionary or grass roots techniques can be a very effective precursor to a formal enterprise information architecture and a way to initiate the process but it doesnt try to impose enterprise models or governance processes until the key step of enterprise buy-in to the strategy has been accepted. It is a complete approach that includes an architecture framework, a methodology and a metamodel for architecture objects and their relationships. It makes more extensive use of holistic reference models than other methodologies and it does so by separating the what from the how. This results in very stable reference models that help to avoid traditional gap chasing and provide a solid framework for planning and the use of a common business language. The service function modeling technique, in conjunction with the information subject modeling and the resultant create-use matrix, produce a mutually exclusive & comprehensive (ME&C) picture of the enterprise. This facilitates effective scope definition and eliminates much of the ambiguity that exists in traditional methods. The methodology is top-down rather than bottom-up. It begins by first modeling the operational view of the business before modeling the systems and technology dimensions. This helps to avoid the paving the cow-path syndrome of some other industry methods. It views service orientation as a business concept rather than as a set of standards and technologies as many SOA methods do (i.e., services is the way a business interacts with its customers and suppliers and is also how internal functional groups interact with each other). These service interactions can be modeling from a business perspective without regard to any technology implementations and only after they are modeling does the methodology drive out the information exchanges and highlight which services are good candidates for re-use. Architecture is viewed as a complementary and collaborative discipline of integration, along with program management and software engineering in other words, it takes an integrated view of the methodology. The table below provides a high-level description for each of the 10 steps: Data Governance Activity Steps 1. Organize Governance Committee: Identify the business and IT leaders that will serve as the decision group for the enterprise, define the committee charter and business motivation for its existence and establish its operating model. 2. Define Governance Framework: Define the what, who, how and when of the governance process. Document data policies, integrations principles and technology standards that all programs must comply with. 3. Develop Enterprise Reference Models: Establish top-down conceptual reference models including 1) Target Operating Blueprint, 2) Business Function/Information Matrix and 3) Business Component Model.

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

4. Assign Organizational Roles: Identify data owners and stewards for information domains, functional owners for shared business functions in an SOA or compliance coordinators in a Data Governance program. 5. Scope Program: Leverage the Enterprise Models to clearly define the scope of a given program and develop a plan for the road-mapping effort. 6. Assess Baseline & Data Quality: Leverage the enterprise models and the scope definition to complete a current-state architectural assessment and identify business and technical opportunity candidates. 7. Develop Target Architecture: Develop a future-state data/systems/service architecture in an iterative fashion in conjunction with Step 6. 8. Plan Migration Roadmap: Develop the overall program implementation strategy and roadmap. 9. Develop Program Models: Create Business data models and information exchange models for the defined program (logical and physical models are generally created by discrete projects within the program). 10. Implement Projects: This is a standard Project and Program Management discipline with the exception that some data governance programs have no defined end. It may be necessary to loop back to step 5 periodically and/or provide input to steps 2, 3 or 4 to keep them current and relevant as needs change.

Step 1 - Organize Governance Committee

Experience has shown that the most successful data governance programs have the commitment and continuing involvement of executive managers who serve to steer these complex efforts and to remove barriers. The data governance committee is designed to promote the necessary involvement of senior managers in data governance programs. Activity Checklist for Step 1 Identify the chair of the governance committee. Identify the governance committee members. Define a formal (written) charter and business motivation. Establish the governance meeting agenda and schedule. Nature of Data Governance Programs A data governance program is a set of activities that substantially changes the way an organization works - often requiring alterations in structure and processes as well as the

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

incorporation of new technology and new computer-based systems. The goals of a data governance program might be to improve regulatory compliance, provide better and faster service to customers, to enable the organization to respond more quickly to changes in the marketplace, to improve organizational efficiency or to reduce operational risks. Rationalizing and consolidating redundant or duplicate databases in support of a merger or acquisition is an example of a data governance program. In todays rapidly changing world, most organizations have not only the opportunity but often the necessity to undertake data governance programs. Such programs always start off with high expectations from their sponsors or they would not begin. However, data governance programs always involve risks and often such programs do not meet their goals. The causes of these risks relate to the nature of data governance programs: Data governance programs, by definition, cut across organizational lines involving more than one functional or line-of-business organization as well as several information technology organizations (and sometimes staff groups such as finance and security). Data governance programs compete with day-to-day operational activities and business imperatives for time and attention. Data governance programs involve many individuals within an organization. Front-line staff are often required to change the way they work and to learn new processes. Data governance programs often involve the development and deployment of new computer technology. Role of the Data Governance Committee To attain the programs business goals and to meet the expectations of its sponsors, the role of the data governance committee is to ensure that action is taken on a timely basis and in the appropriate part of the organization. A second function is to gain the understanding and the commitment from key stakeholders that is necessary to accomplish positive change in any organization. To achieve these ends, the data governance committee provides continued, high-level attention to: Maintaining alignment with goals and expectations Managing risk Providing essential resources and expertise Maintaining Alignment with Goals and Expectations The DG program sponsors have expectations for scope, time, cost and business results. It is important to make these expectations explicit and to make continuing decisions about them in order to stay focused on the goals. This is harder than it seems. For example, pressures to broaden the scope of a program commonly develop and often for seemingly excellent reasons. However, whenever scope is broadened, time to completion, cost, and risk can increase. Similarly, it may be that some benefits expected from a program are no longer valid as time progresses - perhaps because market or regulatory conditions have changed. Thus, it can be beneficial to cut back or reorient the scope and goals of a program and the charter of the committee.

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

It is likely in most organizations and for most programs that over the course of a year decisions must be made on scope, cost, and schedule that could materially affect the programs alignment with business goals. A data governance committee is the mechanism to make decisions in these areas in order to maintain alignment with the overall corporate strategic direction. Managing Risk Every data governance program involves risks. A risk is anything that threatens the business goals of the program, including the ability to achieve the improvements, gains or expected business results. Risks can arise from a wide range of sources, including: Organizational resistance Unexpected complexities Disagreement on priorities Obstacles to necessary changes in the organization Technological performance or integration problems Lack of adequate or complete data These risks ebb and flow during the course of a program. The actions required to mitigate them sometimes require senior-level intervention by members of the data governance committee. Providing Essential Resources and Expertise The success of a data governance program is highly dependent upon the quality of the key members of the program team that are responsible for organizing and performing the work. The skills required can vary over the course of the program and might require additional or different skills from those anticipated at the programs inception. For example, although a data governance program might be focused on establishing new data quality processes for customer service, it could turn out that greater than anticipated skills in handling customer issues is essential for the new process to work in the broader support environment. Furthermore, increased empowerment of front-line service staff during startup could require more than expected training and handholding. Part of the role of a data governance committee is to understand the trade-offs and to make decisions accordingly. Composition of the Data Governance Committee To ensure that timely action is taken, the data governance committee should include senior executives from each part of the organization that has a stake in the success of the program. It is essential that the participants be at an organizational level in which their perspective is that of the overall business goals of the organization and that they have sufficient authority to make resource and financial commitments. An example of a data governance committee is provided below:

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

Meeting Agenda An effective data governance committee structures the agenda for its meetings to focus on the issues that need discussion and action. It is easy to fall into the trap of spending most, if not all, of limited meeting time on status reports and updates, rather than on understanding issues and taking action to mitigate risks. If conducted correctly, a DG committee meeting should take no more than two hours and should result in a clear understanding of the issues requiring resolution and the actions needed to move forward. The agenda below illustrates the course of a typical DG committee meeting. The initial meeting of the committee would follow a different agenda, focusing on a review of program objectives, the expectations of the sponsors, the charter of the committee and the procedure for future meetings. Administrative Techniques To ensure full participation, it is important to set the meeting schedule up to six months in advance. Meetings should normally be held at intervals of four to six weeks during the initial start-up phase and may be reduced to as few as once every three months after the program is well established and operating smoothly. Since the timely identification of items requiring action is a key purpose of the committee, long intervals between meetings or meetings that are frequently rescheduled or canceled defeat this purpose. For any data governance program, an inability to set and stick to a regular schedule of meetings attended by senior participants should be considered a serious risk by itself, and should lead the sponsor to question whether the program has the necessary executive-level support to succeed. Finally, while it is useful to have people with relevant information and expertise attend as needed, the DG committee meetings should not be allowed to grow too large. Regular attendance should be limited to the senior executive participants and key program managers. In general, it is not acceptable for an executive to send a representative, since the representative would not have the authority to make decisions and commitments at the meeting. Critical Success Factors

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

The effective functioning of a data governance committee might itself be considered a critical success factor in a data governance program. Experience indicates that, even when officially set up and organized in an appropriate way, it is often difficult to make such a committee fulfill its real role - timely identification of actions required for the program to achieve its goals. Therefore, the critical success factor is the ability of the committee and its supporting program team to focus on the real issues and risks on a timely basis, to have a substantive and open discussion about what to do, and then to assign tasks to the persons who have the authority and capacity to take effective action. Such a process often runs into cultural barriers, which in turn reflects the organizations inadequacies for mounting a successful data governance program. The cultural difficulty of working across organizational lines can be a major barrier. In some organizations there is a history of distrust and lack of communication across functional or departmental lines. While never beneficial, this barrier can be extremely serious or even fatal for a data governance program that requires coordinated and effective change and problem solving that crosses many organizational lines. The classic example of a horizontal barrier is a rivalry between functional or line-of-business groups and the information technology organizations that support them. In data governance programs, several line-of-business or functional groups might be involved and the rivalries or lack of communication among them can also be serious. The result is finger-pointing (instead of effective problem solving) greatly impeding progress toward the goals of a program. A functioning data governance committee can be an effective vehicle for overcoming these horizontal barriers when the senior members of the committee exhibit an exemplary positive, problem-solving attitude and indicate their clear disapproval of tendencies toward counterproductive behavior within the program team or other affected parts of the organization. If the data governance committee itself falls into the trap of indifference, or worse yet, a constituency-oriented political approach, the effect can be extremely negative. The ability to establish an effective data governance committee is an indicator of the organizations readiness for a successful data governance program. When things go wrong or threaten to go wrong with a program, a common and very human tendency is for team members to avoid getting tagged with blame, while managers try to affix the blame. Yet, the purpose of the data governance committee is to get the program done so that it meets business goals. This goal requires looking forward to determine what needs to be done based on a realistic and accurate assessment of what has already happened. To accomplish their mission, the members of the data governance committee must discipline themselves and provide appropriate signals to the program team to maintain a forward-looking, problem-solving attitude. Summary A cross-functional committee structure can make exceptionally important contributions to the management process that is needed to accomplish data governance programs. Such a committee can help make decisions along the way that keep the program focused on real business objectives - recommending appropriate trade-offs when necessary. It can also assist in identifying risks early enough to do something about them and in enlisting the help of people who have the organizational position to take appropriate action. A data governance committee

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

can provide these benefits only if it is composed of the right people; meets regularly with an agenda highly focused on issues, risks, and actions; and overcomes certain cultural barriers that are common to many organizations. The ability or inability to set up an effective data governance committee is often a good indicator of an organizations readiness to undertake a data governance program.

Step 2 - Define Governance Framework

Enterprise data must have the following six attributes: Accessibility: Ensuring that all enterprise data can be accessed, regardless of its source or structure. Availability: Ensuring that data is available to users and applications, when, where and how needed. Quality: Ensuring the completeness and accuracy of data. Consistency: Ensuring the meaning of data is consistent and reconciled across systems, processes, and organizations. Auditability: Ensuring there is an audit trail on the data. Security: Ensuring secure access to the data. Data governance manages and develops these key data attributes, enhancing the overall value of the data as an asset to the organization. To ensure that enterprise data is all of the above, an effective data governance framework involves four key components: Standards -A key function of data governance is to establish the standards for data in an enterprise. Companies need to establish data definitions and taxonomies, define master data, develop enterprise data models and enforce development and technical standards related to data. Policies and Processes - Establishing and enforcing policies and processes around the creation, development, and management of data is the foundation of an effective data governance practice. Companies need to define data and data-related business rules, control access to and delivery of data, establish ongoing monitoring and measurement mechanisms and manage changes to data. Organization - Arguably the most important issue that companies must address when launching a data governance initiative is how to design the organizational structure. Companies need to define the roles and responsibilities within the organization that are accountable for the data. The organization may include several different roles at different levels, involving both business and IT personnel - from executive councils to day-to-day implementers, such as data stewards and data analysts. Addressing training and organizational change management issues is also critical if data governance programs are to succeed. Technology - Hypothetically, companies could embark on a data governance initiative without an underlying technology infrastructure. Indeed, many organizations launch their initial data governance programs using manual tools (e.g., spreadsheets, Visio, and Word documents) to capture data definitions and document processes. However, most quickly realize that this kind of manual approach is severely limited. It is difficult to ensure high data quality and availability, security is at risk given the ad hoc nature of the approach, and

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

maintaining detailed documentation is an almost insurmountable task. Indeed, it is nearly impossible to achieve the ultimate goals of data governance using a manual approach. Technology can help to automate and scale the development and enforcement of data governance standards, policies and processes. Specifically, a data integration technology platform that provides built-in capabilities to access, cleanse, transform, deliver and monitor data is ideal for data governance. Just as a business process management solution helps to streamline business processes, a data integration platform helps to automate data-related processes. The figure below shows the building blocks of an effective data governance framework. Such a framework takes into account the standards, policies and processes, organizational structure and technology infrastructure required to make data accessible, available, of high quality, consistent, auditable and secure across the enterprise.

The following sections examine the four data governance components in more depth. Standards A key function of data governance is to establish the standards for data in an enterprise. Organizations need to establish data definitions and taxonomies, define master data, develop enterprise data models and enforce development and technical standards related to data. Core principles of effective standards include the following: Develop Enterprise Reference Models. First and foremost, an organization requires a common language to describe its business functions and business data. This model must be independent of organizational structure and technology implementation. In short, it should describe what the enterprise does (not how it does it) and the information that is used/created/modified by the functions and it should do so in a mutually exclusive and comprehensive functional decomposition. This enterprise model serves as the cornerstone of a common language across the enterprise and must therefore represent all business areas and all functional groups at a level that is meaningful to executives. Differentiate Between Data at Rest and Data in Motion. The standards that apply to data that is persisted in a database are different than the standards that apply to data that is being exchanged between applications both inside and outside the firewall. The standards should be clear on when they apply to which category. Support Domain Specific Variants. This principle recognizes that data adapts to meet local needs and that rigid specifications should be imposed only on very selective master data. Invest in Re-useable Code for Standards. This principle suggests that a more effective

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

way of implementing many standards (such as security or monitoring standards) is to not just publish the standards as a text document, but to provide application development and integration teams with approved versions of re-useable software that has the standards embedded in them. Unified Technology Standards: While it is common for large organizations to have a collection of data quality and data integration from more than one vendor, it is critical to have a clearly defined technology standard that presents a unified view of data and metadata. The metadata from the various tools should at a minimum be accessible and transparent. Policies and Processes Establishing and enforcing policies and processes around the creation, development and management of data is the foundation of an effective data governance practice. Organizations need to define data and data-related business rules, control access to and delivery of data, establish ongoing monitoring and measurement mechanisms and manage changes to data. Critical success factors include the following. Solicit Expert Opinion. Seek input from business, legal, finance, risk and architecture experts when establishing policies. This activity may take some time, but it is critical to solicit broad-based input to ensure that the policies are addressing critical needs. Develop a Data Governance Management Manual. This document outlines the phases and activities (descriptions, objectives, roles, tools/aids and tasks) required to execute the DG methodology. The DG methodology encompasses processes such as the enforcement of data governance policies and procedures, issue escalation and metadata management. Create a Data Quality Management User Manual. This document outlines the phases and activities (descriptions, objectives, roles, tools/aids and tasks) required to execute the DQ methodology. The DQ methodology encompasses processes such as data domain definition, scorecard construction and data quality issue identification and remediation. From an ongoing perspective, these procedures involve performing trend analysis on the data and the rules in place to ensure the data continues to conform to the rules put in place through the data governance initiative. Identify and Measure Data Quality. This is a key activity since the ability to understand the data within the confines of the six dimensions of data quality (completeness, conformity, consistency, accuracy, duplicates, and integrity) will form the foundation of the business rules and processes that will be put in place. From an ongoing perspective, the data quality assessment will allow an organization to see how the data quality procedures put in place can impact the quality of the data. In addition, as new data enters the organization, the assessment will provide key information for making ongoing modifications to the data quality processes. Design Quality Improvement Processes. This activity involves the manipulation of the data to align with the business rules put in place. Examples of potential improvements include standardization, removing noise, aligning product attributes, measures or classifications. Monitor Data Quality versus Targets. The ability to monitor the data quality processes is critical, as it provides the organization with a quick snapshot of the health of the data within the organization. Through analysis of the scorecard results, the data governance committee will have the information needed to confidently make additional modifications to the data quality strategies in place, if needed. Conversely, the scorecards and trend

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

analysis results will provide the peace of mind that data quality is being effectively addressed within the organization. Organizational Success Factors Organization is one of the most important aspects of data governance. To achieve the goals of data governance, ownership of the data must be assigned, standards must be defined and policies must be enforced all of which can trigger tricky political situations and fierce territory battles. This strategy document does not attempt to propose an ideal organizational structure to implement or support data governance. However, certain organizational principles are consistent across companies with successful governance programs: Clear Definition of Roles and Responsibilities. A fundamental requirement for any type of governance program (Corporate, IT or data) is clearly defined roles and responsibilities. To enforce accountability, companies need to be rigorous about defining roles and assigning specific responsibilities to individuals involved in data governance. Business and IT Involvement. A common factor among companies with successful data governance initiatives is that executives and staff are involved from both the line of business and the IT department. In most cases, the business assumes ownership of the data and takes the lead in driving data governance. This is appropriate, since the data ultimately exists to serve the business, and the business is the primary beneficiary of effective data governance. IT then partners with the business to implement the technology aspects of the data governance program. IT and the business collaborate to establish specific business metrics associated with the data governance program and track them over time. Demonstrable business benefits are critical to the success of the data governance program; raising visibility of and increasing support for the program throughout an organization. Executive Sponsorship. As a corollary to the principle above, successful data governance programs are sponsored, ideally, by a senior business executive. Senior executives must recognize the value of a data governance program in managing data as an asset and be able to tie the program to specific, concrete business goals. Many organizations establish a cross-functional data governance executive council or steering committee that includes both business and IT executives representing various business units and functions. Integration Competency Center. Successful data governance initiatives frequently involve the creation of an integration competency center (ICC) or center of excellence to support the data integration technology that sustains data governance. ICCs are an organizational approach designed to increase agility and reduce implementation costs by promoting reuse, the sharing of best practices and resources and establishing common processes and standards for integration. An ICC becomes a shared resource for the entire organization, defining the technical standards and processes around data integration and data governance, and providing a pool of highly skilled technical resources that can support specific project and program implementations. Technology Success Factors A data integration technology infrastructure helps to deliver on the goals of data governance by automating the data integration lifecycle. Data integration is not a linear, one-time occurrence.

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

Data integration is an ongoing, iterative process that constantly seeks to improve upon the key data attributes, such as quality and availability. As such, when considering data integration technology, organizations need to take a holistic approach. There are seven key steps in the data integration lifecycle: 1. Access. All data is accessed, regardless of its source or structure. This includes extracting data out of arcane mainframe systems as well as relational databases, applications, XML, messages and even documents such as spreadsheets. 2. Discover. Data sources (particularly poorly documented or unknown sources) are profiled to understand their content and structure, infer patterns and rules implicit in the data and to flag potential quality issues with the data. 3. Cleanse. Data is cleansed to ensure its quality, accuracy and completeness. This may include addressing errors or omissions, enforcing adherence to data standards, validating values and eliminating duplicate data entries. 4. Integrate. To maintain a consistent view of data across all systems, data is integrated to bring together fragmented information and transformed to reconcile discrepancies in how different systems define and structure various data elements. For example, the marketing and finance systems may not only have different formats for the "customer profitability" data element, they may have completely different business definitions for "customer profitability" that have to be resolved. 5. Deliver. The right data is delivered in the right form, at the right time, to all the applications and users who need it. This can range from delivering a single data element or record to support a real-time business operation to delivering millions of records to enable trend analysis and enterprise reporting. Delivery also includes ensuring that the data is both highly available and secure in its delivery. 6. Develop and Manage. A high-productivity toolset enables data stewards, business analysts, architects and developers to collaborate on the implementation and management of data integration rules and processes spanning Steps 1 through 5 above. It also ensures the reliability, scalability and performance required to run mission-critical enterprise systems. 7. Audit, Monitor, and Report. Data is monitored and reports on the data are prepared. This includes the ongoing measurement of key metrics such as data quality, with an eye towards steady improvement over time. The goal is to track progress on the key data attributes and flag any new issues so that they can be fed back into the data integration lifecycle for resolution and continuing improvement. This step also includes auditing. Maintaining a robust audit trail on the data helps to maintain visibility and control and reduces the cost of future change. Data Integration Lifecycle

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

By taking a lifecycle approach to data integration, the technical capabilities of the data integration platform are brought to bear as part of the ongoing data governance program; with the goal of continual improvement of key data metrics and accountability for them. A Technology Evaluation Checklist is included as a Velocity Sample Deliverable.

Step 3 - Develop Enterprise Reference Models

Enterprise reference models are a critical aspect of a data governance program. In essence, the reference models serve as a common description of what the enterprise does and the information associated with its operation. By way of example, an analogy is Grays Anatomy published in 1858 it was the first comprehensive description of the anatomy of a human body and served as a common language for practitioners. Enterprise reference models are necessary for data governance programs so that all functional groups participating in it can effectively communicate. For the enterprise models to be effective and successful, they must have the following attributes: Holistic -The models must describe the entire enterprise and not just one part. Furthermore, the models must be hierarchical and support several levels of abstraction. The lowest level of the hierarchy must be mutually exclusive and comprehensive (ME&C) which means that each element in the model describes a unique and non-overlapping portion of the enterprise, while the collection of elements describes the entire enterprise. Note: It is critical to resist the urge to model only a portion of the enterprise. For example, if the data governance program focus is on customer data information, it may seem easier and more practical to only model customer related functions and data. The issue is that without the context of a holistic model, the definition of functions and data will inherently be somewhat ambiguous and therefore be an endless source of debate and disagreement. Practical -It is critical to establish the right level of granularity of the enterprise models. If they are too high-level then they will be too conceptual if they are too low-level then the task of creating the enterprise models can become a boiling the ocean problem and consume a huge amount of time and resources. Both of these extremes of too little detail or too much detail are non-practical and are the root cause of failure for many data governance programs. TIP >>> The ICC competencies for each discipline are designed to extend or leverage the existing

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

body-of-knowledge and provide information on how to apply that knowledge in an ICC context. For example, ICC Competencies and Best Practices for Financial Management wont make you a finance expert, but they will help you to work more effectively with the finance department in planning and managing the financial aspects of an ICC, and to use chargeback methods to create appropriate organizational behavior. Stable - Once developed, reference models should not change frequently unless the business itself changes. If the reference models did a good job separating the what from the how, then a business process change should not impact the reference models, but if the organization expands its product or service offerings into new areas, either through a business transformation initiative or a merger/acquisition, then the reference model should change. Example of scenarios that would cause the reference model to change include a retail organization transforming its business by manufacturing some its own products or a credit card company acquiring a business that originates and services securitized car and boat loans. Reference models, once created, serve several critical roles: Define the scope of selected programs and activities. The holistic and ME&C nature of the reference models allows a clear definition of what is in scope and out of scope. A common language and framework to describe and map the current state enterprise architecture. The reference model is particularly useful for identifying overlapping or redundant applications and data. They are particularly useful for identifying opportunities for different functional groups in the enterprise to work together on common solutions. They provide tremendous insight for creating target architectures that reflect sound principles of well-defined but decoupled components. The DG methodology is in turn based on an Information Architecture framework shown in the figure below. Key features of the framework include: A four layer architecture with each layer focusing on a level of abstraction that is relevant for a particular category of stakeholder and the information they need: Layer 4 Enterprise View: Overarching consistent context for executives and information stewards. Layer 3 - Business View: Domain models for business owners and project sponsors. Layer 2 Solutions View: Architecture models for specific systems and solutions. Layer 1 Technology View: Technical models for developers, engineers and operations staff. Layers 3 and 4 are developed from a top down perspective. Layers 1 and 2 are created top-down when doing custom development (i.e., able to control and influence the data models) and bottom-up when doing legacy or package integration (i.e., little ability to control the data model and generally a need to reverse engineer the models using analytical tools). Separate models for representing data at rest (data persisted in a repository and maintained by an application component) and data in motion (data exchanged between application components). Relevant information about the models is maintained in a metadata repository that may be

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

centralized (contains all metadata) or federated (contains some metadata as well as common keys which may be used to link with other repositories to develop a consolidated view as required). Note: Step three in the DG methodology, is focused solely on the three models identified in Level 4: Target Operating Blueprint Business Function/Information Matrix Business Component Model

Further details for layers 3 and 4 are provided in Information Architecture. Best practices for layers 1 and 2 are generally considered to be well-established industry best practices and therefore are only referenced here.

Step 4 - Assign Organizational Roles

Most organizations have well-defined owners for applications systems, but typically do NOT have defined owners for data. In a Data Governance program it is essential that data ownership and data stewardship roles be assigned in addition to (not instead of) application system roles. By comparison, it is useful to first understand how application system ownership is assigned. There is generally a strong alignment between application system owners and the business functional group that is the primary user of the system. It is typical for the business owner to also be the sponsor for investments and significant changes to the system. The IT organization often mimics the business structure from an application development and management perspective. In fact, we could consider the individual from the IT organization that is responsible for an application to be the Application Steward while the business sponsor is the Application Owner. Application systems are generally referred to in hierarchies as well some enterprises have a well-defined formal structure and others are more informal. For example, all sales and marketing systems may be owned by the SVP for Field Operations, under which there is a VP of Marketing that owns all the marketing applications, and a VP of sales that owns all the sales applications. Under the VP of sales there would in turn be a further division of responsibility across lead generation, sales tracking, commission management and customer relationships.

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

Whether or not an enterprise follows this typical pattern for application ownership or not, the suggested data governance best practice is to mimic this pattern. Specifically in regards to: Data Owner Typically a business manager that is accountable for: Defining data quality and governance standards. Championing investments in improving and leveraging data for greater benefit to the enterprise. Approving deviations to defined standards. Data Steward Typically a business analyst or IT analysts that is accountable for: Maintaining a clear and unambiguous definition of data elements. Determining how to measure data quality and conformance to standards. Monitoring the use of data in various computer systems. Maintaining adequate documentation on appropriate usage and notes (i.e., metadata). Documenting the origin and sources of authority on each data element. In addition, an understanding of who fills the following roles applicable to data quality within the organization should be captured. The typical roles within an organization when an enterprise data quality initiative is in place include: Project Managers Manage overall data-quality projects, including scoping, tracking progress, etc. Data Analysts Analyze source and target data. Business Users The ultimate consumers of data. Subject-matter experts (SMEs) with vested interest in the quality of data they receive, who can help clarify business requirements and determine whether the data is correct or not. Business Analysts Represent business interests in the development of a data-quality solution. Translate business requirements into technical specifications (i.e., specify rules for fixing data-quality issues). Data-Quality Developers Perform all required data-quality operations. Quality Assurance Perform quality-control analyses and validate results of data-quality operations. Data Stewards Monitor data compliance. There should also be a hierarchy of data owners and stewards. For example, the SVP for Customer Service may be assigned as the owner of the customer information domain, which in turn may be divided into several information subject areas, which in turn may be divided into information objects and eventually into data entities. Finally, it is critical to maintain a record of data owners and stewards. The recommended approach is to integrate this information and maintain the role assignments in a metadata repository.

Steps 5-10 of the Data Governance Methodology

The remainder of the Data Governance methodology (steps 5 through 10) is based on typical industry best practices for programs and projects. The Project Management Institute (PMI)

2011 Informatica Corporation. All rights reserved. Phoca PDF

Informatica's Velocity Methodology

maintains a robust body of knowledge for best practices in these areas which will therefore not be repeated here. The Velocity Methodology does however add significant benefit to commonly accepted best practices when it comes to using an Information Architecture framework (as it applied in step 3 and in steps 5 through 10). Refer to Information Architecture for detailed information.

Competencies
Enterprise Architecture Information Lifecycle Management Metadata Management Modeling Management

2011 Informatica Corporation. All rights reserved. Phoca PDF

Powered by TCPDF (www.tcpdf.org)