Better Data Governance
Data Governance: Harder, Better, Faster, Stronger — Part III
Considering the challenges mentioned previously (cf part I and part II) and IT budgets growing slowly, we can ask ourselves how organizations can:
· Allocate their resources in the most efficient way to support their existing businesses,
· Investigate new data opportunities before being disrupted,
· Comply with regulations efficiently and economically
Classic data management approach cannot just scale. Considering the volume data, it is like crossing the universe, it is just too big, it will go to slow, it will be to error prone, it will cost too much, and it will be probably useless. Not all the answer can be brought by technology, it must be mixed of cultural changes supported by technology as an enabler. Collaboration and automation are the fundamental for unleashing data governance.
As it is not that difficult for machines to do better and quicker than human, automation is the only option for scaling and facing the variety, the volume and the velocity, boosting productivity, supporting data governance principles, detecting new opportunities. With automation business self-service enablement increases through driven and sustainable actions. Collaboration is also critical as data is no longer the concern of few people in the organization. It must become a shared responsibility for generating trust and acceptance. Automation does not mean that there is no longer anything to do, it means that data stewards, data scientists, data analysts can focus on more value-added activities and collaborate on making data democracy real and sustainable.
The physiological need of trust
We have seen earlier that the notion of trust is essential for creating the conditions for efficiency and generating value. And it is obvious that we usually trust:
· what we can understand, when knowledge is shared and not esoteric,
· what relies on a robust process and when we have elements for checking the credibility,
· when roles and rules are clearly communicated,
· when communication is done frequently, and vulnerabilities ae not hidden,
· when excellence is recognized, and feedback is considered
In a recent Harvard Business Research journal, Paul J. Zak, Harvard researcher, Founding Director of the Center for Neuroeconomics Studies and Professor of Economics, Psychology and Management at Claremont Graduate University, and author of “The Trust Factor: The Science of Creating High Performing Companies,” shared that there is a direct correlation between the amount of oxytocin a person’s brain produces and the level of trust they feel in any given situation. The higher the oxytocin, the higher the empathy. The higher the empathy, the deeper the connection. Yes, you could suppose that this link is far-fetched but for creating a culture of trust in data we need to work on increasing end user’s oxytocin. And Zak summarizes height strategies and some of them can be applied to data governance:
1. Share information broadly: If you’ve seen the film “The Big Short”, about the 2007 housing market crash and sub primes crisis. it opens with a fake Mark Twain quotation mark “It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.” Zak mentioned that “only 40% of employees report that they are well-informed about their company’s goals, strategies, and tactics [..] which leads to chronic stress (a fear-based response), wish and exhibits the release of oxytocin and undermines teamwork.” From a data governance perspective, it is about rising data awareness, training data users about the data ecosystem of the organization, the major internal sources of information, as well internal data policies and data privacy requirements. Data users must know and understand why they have or don’t have access to certain data.
2. Recognize excellence and Intentionally build relationships: Some data users can be good at preparing data and building nice dashboards and reports. Through this process they can also discover inconsistencies and data quality issues and become real data governance guards and advocates. Recognizing immediately and publicly their contribution, or even through gamification activities. Comments, ratings, challenges, contests, surveys integrated in the data platform are often a great source of motivation and productivity. Users get to practice giving and receiving feedback in a way that is meaningful and timely.
3. Facilitate whole-person growth: Training has a well-known effect on engagement and retention of employees. So, let’s unleash the data knowledge. There are so many ways today for facilitating self-training through virtual learning and MOOCs. Gamification of data governance can also help here with users leveling up in their data management awareness through a clear learning path and getting access to more data preparation capabilities.
4. Give people greater control over how they work with data and enable job crafting: We can suppose that giving more autonomy to data users could be tough for organizations, which try to standardize their processes and their software for managing and processing data. However, once data users have access to data they should be able to act as “citizen data scientists” and build solutions that meet ever-changing needs. At the same time, through the creation of innovation labs or crowd sourced projects, new experimentation and approaches for manipulating and controlling data can be identified. They will bring their specific business knowledge and insights to bear on defining analytics needs.
The driving idea is that enabling business users to define their own information requirements, answer their own questions, and create their own tools will energize business processes and generate trust by engaging these users’ innovation and business knowledge. This is foundation for data democracy.
Democracy in Data Governance
Data democracy has already started by bringing analytics closer to decision makers and business. This was partially addressed by self-service BI and got recently extended to data preparation and manipulation. The next step is Data Governance for giving access to business definitions, the underlying technical metadata, and to answer to data user when they ask the following questions:
Looking at what has been done for BI and data preparation, infusing democracy into data governance requires a new generation of solutions made for business users combining three critical themes: Simplicity, Quality, and Collaboration.
1. Simplicity: It could sound a bit obvious but proposing simple and easy to use solutions is the most critical. It means, the ability to support zero-coding features, for example, click through or drag and drop design processes. Far is the time when users had to master SQL code for doing data management. Simplicity means also the ability to make easy as possible the integration with third-party solutions and business applications. Moreover, data governance tools should be adaptable whatever is the size of the organization, the maturity, the industry and the volume of information being governed.
2. Quality: Quality is indissociable from the notion of governance as data governance aims at creating quality data systems that users can trust. The focus is here on the ability to support “Just in Time” requirements, like the famous 5 zeros from the Toyota Production System:
· 0 delay: Data must be easy to find and immediately available. The last thing you need is people running around trying to figure out which data assets can be trusted.
· 0 stock: Data redundancy should be constantly checked, and retention policies defined and apply. The less rogue or shadow data sets are maintained the more trustworthy is the data ecosystem.
· 0 paper: All the meaningful information about data should be centralized within data catalogs and business glossaries. Comments and feedback from users as well as the ability to raise alert, change, or data request is critical in terms of traceability.
· 0 default: Getting the best data quality is the top priority when on boarding new data sets or preparing data, to minimize risk.
· 0 weakening: Thanks to a regular, rigorous maintenance and review of business definitions and data quality controls, data governance processes are aligned with business expectations and priorities.
3. Collaboration: Equally important, Data Governance products you need to create the conditions for a supportive environment and to put the effort into fostering collaboration and creating enablement avenues. You don’t want users to feel stuck and alone when they hit a roadblock. If they do, the adoption of data governance will suffer. Creating a supportive environment is part of the cultural adaptation that needs to happen. Supporting the ability to record business user feedback and expectations through discussions over data assets, comments, or ratings toward the quality of certain tables or metrics is a must have. Data governance must become fun and gamification principles could help in fostering the adoption and the change of behavior.
These three themes are essential. However, as for BI and data preparation, the democratization of data governance must be supported by one fundamental capability: “Automation”. The automation of data management activities is the key for masking the underlying complexity and density of data environments, and for allowing to surface and prioritize the best actions to be taken.