Bringing Data Governance to an Agile environment requires a focus on thin-slice, small step implementation, said Sarah Rasmussen, Senior Manager, Data Management, at CUNA Mutual, speaking at DATAVERSITY© Enterprise Data World 2017 Conference during her presentation titled “Delivering a Data Governance Program the Agile Way.” “It’s about being incremental, and the key is to communicate out to leadership and your stakeholders that this is a long-term process,” she said. “As these Agile teams are moving very fast, we need to show some of those key wins, but also stress that this is going to be a long haul.” CUNA Mutual is an insurance company that has approximately 5,000 credit unions and their members, where Rasmussen managed the integration of the Agile Data Governance program.
Four Phases: Initiate, Plan, Build, Grow
“The initiate phase is intended to get leadership buy-in from the Senior Vice President up to the CEO,” she said. The steps in this phase include assessing the organization’s maturity, aligning to corporate strategy, and aligning on the value of data.
Initiate: Assess Maturity
“CUNA Mutual has been very siloed in nature,” she said. “We needed to really shake up what we were doing within CUNA Mutual and get buy-in at the enterprise level,” and understanding the current state was an important first step to organizational unity. They brought in Deloitte to do a high-level maturity assessment, interviewing people throughout the company to learn, “The current state of where we were as an organization as it pertains to data maturity.” Based on the assessment, Deloitte recommended that they focus on “Master Data Management (MDM), Metadata Management, Data Quality, and then Data Governance.”
The company started by focusing their initial efforts on bringing metadata into an enterprise repository, working primarily on building a data dictionary, she said. They had defined some roles but it was considered a failed attempt. “They had really just focused on the data dictionary and not all the other types of metadata,” she said. “It’s good to acknowledge failed efforts, because you definitely learn and grow from them.”
One of Rasmussen’s first tasks was to bring in First San Francisco Partners.
“They were able to meet with all of the senior executives and do interviews to assess the current state, understanding, alignment, [and also to] have some difficult conversations that would have been hard for me to have in a new role,” she said. “These were conversations that helped to promote the idea of investing in a Data Governance function long-term.”
First San Francisco Partners also created and administered a survey that went out to 550 business and IT representatives.
Rasmussen recommends bringing in external parties to gain an outsider’s perspective and benefit from lessons learned by other organizations, “That was really key,” she said. “I can’t stress to you enough how important this is”
Initiate: Align on Corporate Strategy
Senior leaders wondered how they could use the Agile Data Governance program to help meet the company’s goals. “How do we apply what we can do within Data Governance and management to support that strategy?” The next steps involved “looking at our Data Strategy and trying to find those wins,” she said. “We have an awesome amount of information that we haven’t really used to its full potential.” To maximize the value of that information, they decided to lessen the size of their data footprint and to streamline people and processes. “Putting Agile Data Governance and Data Management in place really made sense to people,” they just didn’t understand what that entailed, she said.
Initiate: Align on Value
Rasmussen showed a slide illustrating alignment on value in an Agile context. “The Agile way is looking at how do we deliver value and how are we going to measure that?” Thinking of data as the product for a new line of business, “makes data available, trusted, timely and effective,” she said.
“By rolling this out, we’re optimizing all of the processes that people are doing on a day-to-day basis, as well as really helping to optimize the technology footprint and have data in a little bit more controlled fashion, as well as reducing the amount of data stores we have.”
Initiate: Define What it is
Defining and contrasting the concepts of Data Governance and Data Management was useful, “Not only with presenting to the executives, but also in orientation and training opportunities for people across the organization.” Data Governance is “the establishment and enforcement of our policies and standards,” she said. Data Management is “applying those rules of the road that were created with Data Governance.”
Plan: Vision and Goals
Rasmussen said the Plan Phase focused on how to build operating models, roles and responsibilities, guiding principles, a communication strategy, and a plan to educate those in the organization: “Setting that foundation before we start to actually tackle the activities needed to roll out this function.”
They also wanted to leverage the playbook that First San Francisco Partners had used to develop “guiding principles and policies specific to CUNA Mutual,” develop “a charter for the Data Governance organization and Data Governance Council,” and develop operating models for Metadata Management, Master Data, and Data Quality Management, she said.
The planning process also highlighted the importance of “not just focusing on a tool, or doing training, or finding a repository for all of our artifacts. We really needed to roll this out in a thin-slice manner,” she said. “We needed to focus across the board,” using an incremental, collaborative process to bring people, processes and IT along.
Build: Hypotheses
The Build Phase focused on increasing maturity in Agile Data Governance and management, putting a team together, and identifying a roadmap for rollout, she said.
“If we got our maturity level up to where First San Francisco and Deloitte recommended, we’re going to start to really see some successes in alignment with the corporate strategy and start thinking about data as that asset.”
Build: Data Governance and Data Management Functions
“Our first hypothesis is: what is the basic thing that we want to do to get this moving? In this case, it was creating the Data Governance bodies,” she said. “I wanted to focus across the board to grow this and make sure we were staying in balance both on Data Governance and Data Management.” They planned to bring together some areas that were siloed and “get the tools and technologies for Metadata Management and Data Quality at an enterprise level,” she said.
Build: Minimal Viable Foundation Hypothesis
Rasmussen shared a slide outlining the process they used for establishing a minimal viable foundation for successful Data Governance. It included a communication plan, orientation planning, a strategy map for Agile Data Governance, Organizational Change Management planning, and management and operating models.
Build: Data Governance and Data Management Scrum Team
The process of identifying stakeholders “was one of the most valuable exercises in getting this going,” because “if you look across this group it’s pretty much everyone.” As a result of changes going on in the organization, legal, compliance, risk, and CISO departments had been doing a lot of work that overlapped, she said. “They had questions and didn’t know who to go to.” After the structure was in place:
“All of a sudden they had a means to bring those concerns, questions to the forefront and also apply them to what we were doing and really test our operating model and plans. That was pretty awesome,” she said. “Having my boss come into a room and see the work that we are doing and the value provided, he can then go out and be that champion for the organization to say, ‘hey we’re making strides in these areas and we’re meeting our goals.’”
Build: Sprint Project Team
Rasmussen said they then took two weeks off-site with a group consisting of the project team, and some stakeholders and subject matter experts. This time was spent going over high-level plans, breaking them up into stories or product backlog items. “How are we going to move this forward? What are the roles? Do we have the right people in the room?” she remarked.
“We looked at each of them and we decided who would be best to be working on those. We saw dependencies and then we sequenced them as best we could, and scored them to make sure our goals where attainable.”
Then they were broken into sprints, and goals were set for the first few sprints, she said.
“This is probably a little bit different flavor in rolling out an Agile Data Governance and Data Management function,” she remarked. Having people trained in scrum was an advantage she said:
“Because once we got operational in nature they are working with the project teams. They know and understand what is Agile and what people are driving towards and how are they doing things quicker, faster, better? How can they work and embed themselves in those teams to start thinking about data a little bit more than those teams had been in the past?”
Grow: Data as an Asset
During the Grow Phase, the processes that have been put in place with other teams are spread further throughout the organization. “We can start to have project teams think about data as an asset” and identify how they can look at their work in a different manner, she said:
“Especially as our organization is very much investing in data and analytics, as I’m sure a lot of your organizations are doing right now. Most scrum teams are thinking about a product. What are they delivering? It was really just over a year and we put the resources in, we had the plan, and we had people training other people, and that was a huge success for our organization. It was a model that I wanted to use in getting this whole idea of Data Management and Data Governance out into the ether.”
Check out Enterprise Data World at www.enterprisedataworld.com
Here is the video of the Enterprise Data World 2017 Presentation:
Photo Credit: astephan/Shutterstock.com