Organizations collect, process and store enormous amount of data, A growing number of applications/systems, which support various lines of businesses, have been collecting more and more data through various channels. Mergers and acquisitions have made the situation even more complex and confusing when it comes to the management of data and business processes.
Figure 1 shows an example of the way data assets grow over time.
A lack of data management practices (or data standardization processes) results in challenges that are faced by both IT and business. Often, individual applications or systems maintain and manage their own data. This results in data silos or data hubs (logical or physical). The solutions around data redundancies, metadata management and related data issues are always tactical in nature, whereas data anomalies are fixed by temporary patches or left unprocessed. Thus, issues around data integrity, consistency and accuracy make the data unreliable; typical factors impeding clean and conformed data include lack of standards, typos and duplicates, applications being ported from different platforms/languages, lack of standardized quality processes, historical/outdated data, unknown data and many others.
Businesses need clean and conformed data to make better decisions. That’s why data and information quality has become an important factor in making better business decisions.
Such decisions are critical for the following areas:
- Customer services.
- Product optimization.
- Regulatory compliance.
- Managing operation costs and risks.
Without reliable data, business intelligence generated over a period of time is questionable. Inaccurate financial reports and audit reports will not only receive penalties but will have financial implications as well.
Thus, processes/policies and rules that ensure enterprise-wide data asset management are needed. It becomes difficult to maintain and manage data without policies, strategies and dedicated efforts by the team across business functions. Data governance seeks to solve these problems, through information policies, data rules, guidelines for managing key data elements and assigning roles for accountabilities and responsibilities.
What is Data Governance?
Data governance is the processes or policies which guarantee that important data elements that can be trusted. A framework or set of processes is implemented throughout the enterprise, empowering the right people to take control of data and processes. A data governance program also includes technology, which helps identify and fix data issues, resulting in fewer negative events due to poor data. It’s also about the communication, identifying common language that will bridge the gap between IT and business managers.
In short, data governance is about management of the availability, usability, integrity and security of the data. Some of the key focus areas are data quality, data integration, policies around privacy, compliance and security, the data warehouse and BI, architecture integration and analysis and data access, in terms of archival, retrieval and storage.
A data governance program will have certain drivers, such as:
- Identifying data anomalies and fixing them, particularly with regards to key data assets around important business processes.
- Optimizing business processes and defining data rules.
- Designating the right people responsible for information quality and security.
- Creating policies for handling data, in case of initiative changes.
- Coordinating with key business stakeholders to ensure that information policies support business objectives.
The data governance program and initiatives around information quality management need to involve stakeholders representing a cross-functional team to fulfill the objectives.
Stakeholder Involvement and How to Start Data Governance Initiatives
Issues around data management and information quality can be addressed using data governance initiatives. These initiatives need business and IT support, which means stakeholder involvement across the teams. These initiatives start with bringing people together for mutual understanding and educating them about doing the right things in the context of information quality.
Figure 2 shows how different stakeholders within business and IT see data governance differently
Starting a data governance initiative requires answers to three questions around benefits of the program:
- How will the program increase company revenue?
- How can the program lower costs?
- How can the program reduce the risks and address compliance issues?
It would be wise to start with initiatives where data needs to be fixed to minimize the risks or where business users have voiced complaints. Additional initiatives that can be considered may include a CRM implementation, a new data warehouse, BI initiatives and analysis of complaints management.
Information quality surveys can be launched across the line of businesses, with questions about data related to important business processes. For example, the billing and dispatch departments can be asked about the validity of the addresses of the customers, in terms of format. Does the customer get the communication or shipment in time or it is lost or returned? The inventory department can be asked whether the data always reflects the correct inventory amounts and types. The marketing department can be asked about their confidence in contacting potential customers using the email or phone numbers listed in the system.
Prepare a case study for analysis of each problem statement, followed by a detailed business impact assessment. With this, information quality ROI can be calculated, which will significantly help make the case for whether data governance initiatives will be beneficial.
Generating awareness among the business stake holders about latest trends, risks and competitors’ initiatives regarding information quality may also help in selling the importance of data governance program.
To show ROI and get immediate attention, you need to start with small initiatives and share the results and a detailed impact assessment with key business users or stakeholders; this will help to sell the importance and need for a data governance program.
To fulfill business objectives, a data governance program needs to have a roadmap. This roadmap should clearly reflect the high-level approach and iterative nature of future engagements with business and IT.
Data Governance Roadmap
A data governance roadmap outlines the guidelines for its initiatives. It starts with identifying short-term or long-term business objectives around the initiative, which requires input from business and IT stakeholders to asses business and IT processes. (See Figure 3.)
A data governance approach lists phases and high-level activities, as shown in Figure 4.
A data governance forum receives input and assesses the impact on business processes, with the help of a cross-functional team. The forum offers insight on policies, standards, metadata management and control. Information quality scorecards can also help assess the data governance initiatives, point out success factors and provide an executive management summary, thus attracting funding for more initiatives.
Practical Approach for Information Quality Scorecards
An information quality scorecard is a tool used by the data governance team. It involves aggregating technical metrics with business metrics, thus helping business stakeholders remain aware of problems around key business processes and prioritize resolution. The scorecard helps the data governance forum analyze the impact of initiatives on enterprise-wide information policies, as well as compliance and regulations. (See Figure 5.)
Overview of Data Governance Team Structure
Data governance program initiatives can fulfill objectives only when the right people get involved with defined roles and responsibilities. Key roles are:
- Project Sponsor: These can be C-level executives or business leaders who are driving the program. In the case of a financial services company, a CFO who has faced challenging situations related to risk and compliance due to poor quality data may be involved. A CMO may be involved when customer data is at stake and customer data standards are not met. The initial executive sponsor may be a business manager when the program starts with small initiatives, and senior business leaders may get involved down the road when the program is implemented enterprise wide.
- Business Manager: Key business managers or subject matter experts get involved to provide the context of business process and data. They report on the impact of the data on related business process and provide recommendations about the key data elements. They also provide input on the scenarios that can be validated against the data.
- Data Stewards: These people include programmers, data analysts, data architects and database or system administrators. Their key activities include designing metadata mappings, understanding business processes, defining data rules, data mining, creating data assessment reports, cleansing data issues, and managing and maintaining the related infrastructure.
- Project Manager: The primary role of a project manager is to deliver the finished project. This role involves managing all the resources, communication, coordination, and risks and issues management.
Data governance programs consist of cross-functional teams for data issues management.
- A team of business/IT analysts manage and log data issues, categorizing issues by line of business and coordinating with respective business SMEs for resolution.
- A data governance forum looks into policies and standards, aligning business leaders on managing risks and compliance, approval for strategies and funding for projects that involve data cleansing/transformation or incorporating new systems/applications for the information quality assessment.
- A data team involves business analysts, data profiling analysts and programmers for the extraction of source data and infrastructure support.
Important Considerations around Key Challenges
The data governance program will face key challenges that may impact project timelines in large transformation programs.
- Access to data: Basic groundwork for data assessment starts with the challenges, such as identifying key data elements for assessment and their best source. Extracting data from a source and getting the access to live data is another big hurdle. Transformation programs in large organizations may require access to 100 percent live production data.
- Data assessment: Business SMEs and application SMEs should be aligned on metadata management, defining data validation rules and additional data analysis required to support data cleansing or data transformation during the resolution of issues. Care should be taken that the resolution, tactical or strategic, supports the policies and is signed off by the data governance forum. Tracking and monitoring progress on these issues should be done with care.
- Workshops: Alignment of cross-functional team for data governance workshops is another big challenge. Availability of SMEs is important and can be a major issue. Business managers should be informed of resource requirements for workshops. Carefully planned workshops should have a clear agenda and involve key decision-makers. Clarification of queries and issues should be well-documented and signed off by key business stakeholders.
- Infrastructure (hardware platform and software licenses) management and support: It is necessary to coordinate with the IT team for the availability of the hardware platform, installation of required software licenses and application of latest patches. Vendor support can be another big challenge and must be closely managed. If care is not taken in mitigating these risks, data governance program timelines can be impacted.
- Program management: It is necessary to inform the stakeholders about their roles/expectations upfront. The data governance program needs to align with key business stakeholders and business leaders regarding the roadmap, approach, high-level plan, scope and key dependencies.
Organizations worldwide have been facing tremendous challenges in information management. Information quality can be assured only through data governance initiatives and processes, which provide insight on data issues, resolution, standards and responsibilities. Forming cross-functional team within a data governance forum empowers the right set of people to take control of data assets and make the right decisions in the context of quality data. Clean, confirmed and complete data enables a business to make better decisions in order to achieve their business objectives.
As we move along this journey, we need to assess and audit the current situation by referring to a data governance maturity model. This model does not criticize existing practices but provides guidance. (See Figure 6.)
There is no single formula which will organize the enterprise perfectly or quickly. Enormous efforts across cross-functional areas need to be put in, and those efforts need to be rewarded to keep up momentum in the long run. Success only comes through personal responsibility and careful planning.
Santosh Arvind Sardesai is Senior Technology architect at Infosys. He has 14 years of experience in Consulting and implementation of the projects in the domains of Information Management which covers Data warehousing, Business Analytics, Data Migration, Data Governance and Database administration. ûHe is certified professional on Oracle (OCP 9i DBA), SAS (Base SAS Certified), Teradata (Certified architect) and IBM (Information Analyzer). He can be reached at firstname.lastname@example.org.