“How do you make Data Governance real where the rubber hits the road?” asked Danny Sandwell, speaking at the Enterprise Data World 2016 Conference. Sandwell, who is Product Marketing Manager for erwin, Inc. said that organizations using data models to aid in the creation of a solid Data Governance Program are better able to meet challenges and take advantage of opportunities as they arise: “Data models are not just for DBMS’s anymore.”
Good Data Governance
Starting off with a brief, high-level overview of Data Governance, Sandwell said, “Data Governance refers to the overall management of the availability, usability, integrity, compliance, and security of the data employed in an enterprise.” He stressed the importance of having a governing body with executive sponsorship, as well as representation of all stakeholders. Well-defined and understood procedures and rules, he said, “will drive good governance within your organization.” Having a plan to measure results and publish those back to stakeholders allows a company to prove to the organization “the value of good Data Governance.” And lastly, creating a facility to develop, implement, and manage the Data Governance Program iteratively over time, he said, “that’s our focus today.”
Goals of a Data Governance program – Trust in Data
A good Data Governance Program, Sandwell said, creates trust in the data, so that end-users see it as a valuable, accessible resource for decision making. The goal is to provide a program that is consistent, high quality, and understandable, making it easy for end users to derive value from the data. This, in turn, fosters transparency and accountability for data assets and their management, which is essential for “creating trust in your enterprise data,” he said.
“Leverage this initiative to make your organization as agile and as efficient as possible as you go through those Data Management processes,” so that you don’t become a problematic cost center within your organization and you will, “make your organization more able to meet challenges and take advantage of the opportunities that can be significantly impacted by the data within your organization.”
The Value of Data Modeling
Data Modeling allows an organization to work out a plan before offering it up to users. It’s accepted that the right way to design relational databases is to take time for modeling, do the analysis, understand the challenges and risks, and work out the “what-if’s” before ever showing that database or offering it up for use.
“Data models and the act of modeling [have] an intrinsic value in an organization. [This] has been proven over and over again, especially in what we see as a traditional data modeling use case: the design and implementation of database systems.”
Data Modeling allows for a rigorous process of low-cost, low-risk analysis before making a commitment to build out. Once built, he says, those same values should continue to apply to the Data Governance Program within your organization. “Data Governance is not something that is once and done. It’s something that has to iterate over time – it’s a living breathing thing.”
The Role of Modeling in Data Governance
Data Modeling also works for Data Governance because it provides a better understanding of the Data Governance process. Sandwell said that we’ve moved beyond columns and rows to a more visual presentation of data because visualization makes a greater impact. When used to build DBMS, modeling breaks down the complexity and lets people really understand “the message that the data is trying to give you,” he said. “So why would it be any different for Metadata, and for the structures and schema that are underneath that?”
Modeling provides the ability to standardize across systems so that understanding is easy for all users because the final product feels familiar. It allows users to relate – to connect things that aren’t connected out of the gate, he said.
“Modeling allows you to extend beyond what is in the technology or platform, because you can start to purpose-build elements of that model to get the value and the result that you want out of that model. Modeling has that built in to it.”
Modeling also provides the ability to compare the model to the real world, identify gaps, anticipate what the next challenge is, analyze, and then sync over time, he said.
“Then it becomes the point of true collaboration within your organization because it’s Metadata, it’s consistent, it’s standardized, it is the source of the truth, but it’s a visual source of truth that allows people who are not necessarily data management professionals understanding some of those technical nuances, to be able to understand in their business role.”
Best Practices
Focus on building in the basics from the beginning, he said. “Data design, definition and standardization – that’s the foundation – if you don’t have that, your governance is going to fail.”
Sandwell said it’s also important to build within the company culture, using what is currently being done as a base to build on, and making compliance as easy as possible. “If it’s not the path of least resistance, then we go back to management by exception,” where “there’s always a good reason not to do it the right way.” By building in easy-to-use processes for all users and standardizing procedures from the beginning, compliance becomes easier than non-compliance.
When you put data into a usable context, you can benefit from collaboration with “people who have never before touched a data model in their life.” Business users can make models their own, and provide feedback that can be used to improve the usability of the real data model, making the development process more transparent and sharing accountability across a wider area. You can leverage participation by others to “fire you faster down the track” and accelerate your time-to-value and time-to-success, he said.
By building good Data Governance polices and procedures into your model, you can anticipate potential security issues, he said. “If your models are secure, your data definitions are secure, then your modeling process is secure, trusted, and transparent.”
It’s also important to keep your models in sync because, “a model that’s out of sync is not valuable at all – in fact, it’s a risk and a problem,” he said.
The Value of a Visual Model
Bringing people together from disparate parts of the world around that visualization process and breaking down complexity is a proven way to bring value to an organization, and it has made organizations “more effective, more efficient, and more agile,” he said.
“Will it do every single thing and fulfill every requirement that you have for Data Governance? No, but it will provide a foundation that you can build on and always go to: a standardized, consistent foundation. Then make sure it’s under control, so that there’s accountability, transparency and trust, and publish that out in terms of a visual architecture, a visual data dictionary. That’s a data dictionary on steroids.”
With a model-driven Data Governance Program you can have “a facility that reflects reality. It’s been augmented and improved to not just look at the physical reality but at all of the other realities” for every user, and “it’s easily accessible for people to look at, visualize and understand,” he said. Model-driven processes provide “high value capabilities that reduce risk in your organization,” he said. “And all of those things relate back to Data Governance.”
Models are a proven way to encourage data-driven decision making. When you can offer reliable, self-service visual Metadata, he said, it gives end users the opportunity to better understand and trust their data more, so that when they go to make a decision, they are comfortable enough “to go to the data instead of going to their gut.”
Sandwell said that models “make Data Governance real within your organization, because in those models, you have IT standards, source Metadata. You have business rules and requirements,” and the “tribal knowledge and insight that needs to be put in and reflected effectively within your organization.” Put all of that into a facility that supports “stakeholder collaboration, self-service Metadata, and data asset discovery” he said. “Where would that take you, and where would that get you in terms of a foundation to springboard a successful Data Governance initiative?”
Here is the video of the Enterprise Data World 2016 Presentation:
Register for the Enterprise Data World 2017 Conference Today (in Atlanta, Georgia)
