Data is everywhere. With the growing interconnectedness of people, companies and devices, we are now accumulating increasing amounts of data from a growing variety of channels. New data (or combinations of data) enable innovative use cases and assist in optimizing internal processes. As a result, data is set to have a significant effect on tomorrow‘s business and will undoubtedly become a value-adding factor. However, effectively using data thorugh data governance needs to be learned.
Companies using multiple systems for various purposes tend to have complex technology landscapes with highly specialized silo solutions. From operational systems to support “smart processes”, to the data warehouse for enterprise management, to exploring new use cases through advanced analytics: all of these environments incorporate disparate systems, each containing data fragments optimized for their own specific task. Data and data management processes are everywhere in the organization so there is a growing need for a comprehensive view of business objects and data. It is therefore vital that data is subject to some form of overarching control, which should be guided by a data strategy. This is where data governance comes in.
“Data governance refers to the individuals, processes and technology required to manage and protect enterprise data assets. Its goal is to ensure interpretability, correctness, completeness, trustworthiness, data security, accessibility and traceability of enterprise data in an efficient and effective manner.”
Data governance is a widely discussed trend at the moment. After all, it is important to network the data competencies in any company in order to be able to implement new types of application and to fulfill compliance requirements such as GDPR. Best practices are still all too thin on the ground. There is a lack of available knowledge and orientation at the current time.
Data governance is a trending topic mainly focused on BI and data warehousing, and mainly driven by compliance
Only 4 percent of participants regard data strategy and data governance as an inconceivable approach for their business now and in the future. A majority focus their current governance activities on business intelligence and the data warehouse. Best-in-class companies have realized that it is important to cover all data environments and establish a feedback loop from data usage in BI and analytics to drive data improvements. While compliance is the major driver for data governance, it bears the risk of reducing it to a very restrictive procedure.
Data quality is the top challenge when it comes to using data, closely followed by organizational issues. Inadequate data quality remains the foremost challenge users face when using data. This has been shown repeatedly by market research and data-centered projects for many years.
Yet it appears that the reasons for the apparent inability of businesses to cope with this problem in order to achieve continuous improvement are largely organizational. In spite of pressing data quality problems, there is a lack of acceptance and priority for data governance at executive level and in lines of business.
Practitioners and planners have different views on how to approach data governance
Best practices in data governance are still rare. However, there is widespread agreement on one point: technology is not the limiting factor. Businesses currently planning their data governance endeavor tend to focus on administrative tasks: they favor the development of a data catalog as their top-rated measure, followed by roles and processes. Practitioners concentrate more on practical execution, such as data quality monitoring and training. This way of working helps generate business demand. Challenging and developing users in data governance issues is a promising approach as it addresses the most widely identified challenges today and in the future.
The most striking benefit of data governance is the effective use of data based on a unified understanding
Targeted data governance creates strategic and operational added value. It is a mechanism to become more effective and efficient in the use of data. Creating a unified understanding of data can raise this effectiveness to a higher, overall strategic enterprise level and help a company along its path to digitalization.
However, if these actions are targeted at the data warehouse, their achievements will be of limited usefulness. Business value is ultimately generated in core business processes.
Data governance success depends heavily on adequate resources which cannot be provided by a central unit alone
There is clear agreement that data governance can only be achieved by interdisciplinary teams with strong participation from the line of business. At the same time, the number one challenge is a lack of resources. Tasks are primarily delegated to business analysts and key users who already own other line duties. It appears that efforts, for example for data stewardship, are not sufficiently acknowledged by executive and divisional management.
Due to the siloed history most enterprises share, convincing these stakeholders of the need for – and value of – a holistic data strategy is presumably the most difficult challenge they face. Data strategy and governance must be closely aligned with the enterprise and digitalization strategy, and with an overarching view of business processes.
Technology is not a limiting factor for implementing governance
Technology is not a limiting factor for implementing governance. Compared to other challenges, ‘lack of tool support or wrong tools in use’ is of relatively minor significance to our panel. Unsurprisingly, data cleansing (58 percent), data quality management (53 percent) and data integration (52 percent) are identified as the most needed functions to support data governance.
Success factors for data governance
Due to the complexity of the enterprise landscape and the often limited availability of human resources, companies need to find innovative ways to boost efficiency in governance tasks. Awareness is growing that machine learning and AI can play a central role in this area.
Participants identified convincing stakeholders and management support as major challenges. Consequently, management support and the identification of priorities based on corporate strategy are top of the list when it comes to critical success factors for data governance programs. The third priority – making data governance a substantial element of all projects and processes – appears difficult to realize in practice. The predominant focus of data governance initiatives on BI and data warehouse environments demonstrates that there is still a long way to go.
When these prioritized success factors are compared with current measures and established roles, it becomes apparent that data governance has not yet transferred from theory to practice. For example, only 24 percent of participants believe that internal marketing of data governance is important to convince potential stakeholders from lines of business. However, this is a crucial measure to secure management support. Overall, the findings of this study suggest that much of the potential of data governance is yet to be discovered.
Companies find themselves facing great complexity and, in particular, organizational challenges. We would like to conclude with the central success factors we have learned from this study:
- Line of business involvement is essential for data governance establishment. Business users currently do not have the appropriate resources and skills available to master this task and must be empowered.
- Executive support and acceptance in business are essential prerequisites for pursuing data governance. Management can be convinced by obvious need. A good starting point is to monitor and visualize (poor) data quality and create a feedback loop from analytical data usage back into core business processes.
- Data governance should be set up as an overarching enterprise initiative, not limited by boundaries set by specific data environments. Focusing data governance on the data warehouse maintains the data problem instead of solving it.
Infographic of the key findings