DATA GOVERNANCE
As technology is only one part of business requirements solutions, the understanding and embracement of data governance is essential to deliver the promised value of information management initiatives. There is, unfortunately, no universal model of data governance that would work for everyone so the first challenge is to develop the unique model that will be successful for a given organization.
Data management
After evolving in the past decade and becoming a priority for enterprises, data management initiatives range from initiatives of information management infrastructure to business-driven initiatives significantly dependent on the quality of underlying data.
Some of the common initiatives include:
- compliance
- legacy data migrations
- mergers and acquisitions
- operational BI
- master data management
- enterprise data warehousing and traditional BI
- information-as-a-service (IaaS)
- metadata management
- other functional business initiatives.
Being suggestive catalysts of data governance, does not mean that in the continuously expanding complexity of data management surroundings, they are not thwarted by equally powerful inhibitors of adoption, such as:
- lack of business involvement
- ROI calculation's ambiguity and the absence of the business case
- cross-enterprise complexity
- not agreeing to disagree
- unrealistic scoping.
Many successful data governance programs have taken many years to deliver required results and mature as with good resources, prioritization and mostly important - patience - best practices lead to exceptional value.
Long time investment in data quality and data governance allows various organizations to realize extraordinary success as direct and indirect calculated benefits exceed significantly after beginning of the information programs in enterprises. Surprisingly, in some companies data governance seems to be threatening to people as it revealed that data problems are on their behalf, not because of the automated processes. So that is why appropriate education should be provided as it is a critical step in the road to success.
Data architecture and data governance
At the center of it there is the manager - the 'chief data steward'. His responsibility is to drive the whole process and to ensure that data stewards know how to focus and prioritize their efforts.
The first step of the data governance program is always to analyze, define and baseline quality levels of the data so that ongoing major performance indicators are monitored going forward. The business stewards are also responsible for strengthening the business case by identifying extra opportunities for improvement. Sometimes the largest risk anticipated by a company is prioritization.
When looking to enhance of functionality for a database, sometimes the lack of data ownership and governance helps to move the decision towards creating a new customer database instead of increasing the existing one. Critical milestones are, among others, socialization and education of the governance and stewardships concepts throughout the enterprise and up through the senior management.
Data ownership is a representation of an important concept within data governance program. When the company has determined stewardship, every column and table with the data has its responsible owner, depending of the type of data or metadata. Usually, the owner of an individual row in a table, may be the company's executive from the highest level, whereas the owner from the level of data structure might be middle management in a business unit. Multiple managers and owners usually cooperate to create a shared definition and one manager accepts final decision.
Company culture elements include the ability to take long-term perspectives and future investment in more collaborative and less political environments than it would be without paying attention to them, as it happens in many companies of different sizes.
In some companies the concept of data stewards and custodians is adopted. Data stewards are usually business stakeholders responsible for data definitions, quality and access authorization, whereas custodians are IT roles that are responsible for data management in order to comply with business requirements.
Additionally, from a methodology point of view, a project certification 'building permit' approach may be agreed to adopt and eforced. Such approach impelments a single standard for process change management for all data modelling. Data governance is managed through a tightly gated process in which 'data management' represents an official gate to be passed by any architect, designer or data modeller across the enterprise to move to the next steps within the data management project.
Questions to ask
When designing the data governance methods appropriate for a given company, some vital questions need the answers:
- evengelism - is there executive sponsorship to get started?
- scope - how will the scope be reduced to face only the most important data?
- ownership - who will be the owners of the most important data in the company?
- program management - what should drive the program - business or IT?
- structure - is the organization centralized, decentralized or hybrid?
- staff - is data stewardship full-time or part-time job?
- funds - where does it come from?
- business drivers - can they help build a business case for governance?
- inhibitors - what are the primary ones and can they be migrated?
- timeliness - is everything really ready?
Data governance roles and responsibilities
- Data governance roles and responsibilities:
- executive sponsor - identified early, active participant, CXO level, evengelist, final escalation point ; if necessary, it can be a commitee
- program driver - not biased, stewards' coordinator, communicator of all decisions, drives data quality auditing and metrics, primary escalation point
- business administrators - IT savvy, strong communicators, educators across the business
- IT administrators - business savvy, strong communicators, educators across IT.
To design data governance basing on company culture, one needs to face from the start that this culture will not change to support data governance, therefore it is better to work with understanding of this culture and leveraging its strenghts.
The rules listed below might help here:
- Start small and grow
- Determine executive sponsors as well as data administrators early - they are the evangelists.
- Do not forget the business case.
- Connect participation to employees' performance goals.
- Prioritize with hub-and-spoke approach.