There has never been a time when more options were available to stand up a process-driven and/or platform-driven Data Architecture. In recent years, some companies find themselves with an embarrassment of riches, having every tool known to humankind. Using tools à la carte, with a different tool for each solution, works for some organizations. Other organizations have core applications in place; there are data flows, but they’re very operational in nature, and there’s nothing even relating to Data Architecture. Still others are trying to use one tool to accomplish everything, end to end.
One Size Does Not Fit All
“There’s no one right way,” said George Yuhasz, VP and Head of Enterprise Data at NewRez, because the demands and the value that Data Architecture practices bring to an organization are as varied as the number of firms trying to get value from data. Yuhasz was speaking at DATAVERSITY® Data Architecture Online Conference.
The very definition of Data Architecture varies as well, he says, so get clarity among stakeholders to understand the constraints and barriers in which Data Architecture needs to fit. Will the organization prioritize process alone? Or process, platforms, and infrastructure? Or will it be folded into a larger enterprise architecture? Without a clear definition, it’s impossible to determine key success criteria, or to know what success is, both in the short term and long term.
The definition should be simple enough to be understood by a diverse group of stakeholders, and elegant enough to handle sophistication and nuance. Without it, he said, the tendency will be to “drop everything that even relates to the term ‘data’ onto your plate.”
No matter where the organization sits on the spectrum of Data Literacy, performing an honest assessment of the current state of data literacy is necessary for success:
• Organizational Data Literacy: To what degree does the organization knowingly use data to both drive positive outcomes and to build a catalogue of benefits over time?
• Operational and Core Platform Footprint: What is the core platform footprint? Is it primarily one or two sets of platforms or many things on a core?
• Framework for Understanding Data as an Asset: What capacity does your organization have for understanding data as an asset? How much capacity for transition does your company have?
• Technology Roadmap: To what extent does the Data Architecture align with the activities of the technology organization?
The Need for a Tailored Approach
There are common approaches and frameworks to Data Architecture, but they must be aligned with the type of organization, the industry in which it operates, and its specific needs and goals. Healthcare, financial services, and manufacturing companies have differing needs, and a startup will have a different profile from a long-established company. Culture is also a significant consideration as organizations with a top-down management style will need a different strategy than one with a bottom-up management style.
Adapt to the Culture and Influence Stakeholders Over Time
Yuhasz shared practical steps for successful Data Architecture implementation. His first recommendation was to approach the process with a sense of humility. This idea is often interpreted as simply agreeing to everything, yet he stressed that it’s more about respecting others as professionals, giving them the benefit of the doubt and assuming that everyone wants to provide value to each other and to the firm. “Even if you don’t feel like you’re getting it back, sometimes you need to be the one that sets the example.”
Leading with empathy and a willingness to listen provides essential insights into the culture and an understanding of stakeholder perspective. Rather than being seen as an external party trying to be disruptive, humility also allows a manager to adapt to the culture and become “baked into the organization,” he said.
Raise Awareness and Education
The feeling of being at a conference can be so inspiring: “It’s a wonderful sense of feeling like we’re there amongst our people,” he said, but it can be a challenge to curb enthusiasm when it’s time to bring it back to constituents at home who haven’t shared the experience. It’s important to meet them where they are as well as working over time to raise their awareness and education. Focus on mutually important results and goals and work as partners to get there, using the new tools you’ve acquired, rather than just explaining what you’ve learned, he said. “They can see the value in that, and you’re going to win them over.”
Act as a Change Agent Appropriate to the Organization
The key to implementing successful change is to start small, be humble, and build inventory and capability from there. Talking about making “a couple tweaks”’ here and there will provide opportunities for quick, small successes that can build momentum quickly, especially if the process involves listening and shared problem-solving. At the same time, partner with those in your organization who seek to innovate, improve, and transform, and find out how are they getting people excited about opportunities.
Foster Innovation, Minimize Disruption
Successful transformation also requires an awareness of how much potential there is for disrupting people’s lives, Yuhasz said: “This is the empathy part.” Disruptions might be for all the right reasons, but understanding and acknowledging the impact will help engender an understanding of the benefits.
Evangelize the Power of Data to Drive Outcomes
Keep making connections between the outcomes achieved and the role of Data Architecture in deriving value from the activities that led to those outcomes. In a perfect world when the story of the project is shared, he said, others will voluntarily tell the story. “And you’re just agreeing and saying what a great job they did in helping to get the value.” Sharing the credit, sharing the wealth, and building others up pays back huge dividends.
Have alignment on the definition of what a good outcome should be, using the current roadmap. Find a shared understanding of the problem and ensure that everyone is working toward the same goal. Work toward a baseline shared perception of “good” with stakeholders and users, as well as best practices for getting there.
Know When “Best” is Necessary
The concept of “good” is in relationship with “best” and there’s a lot of room in between those two spaces to deliver a lot of value, he said. However, when it comes to areas that are critical to keeping the lights on, that’s where the highest standard is important.
Simplify the Big Picture
This starts with a problem statement, clarifying the value in doing X or Y, and outlining the construct by which they’ll be implemented. Be careful about the amount of detail, though, because “The more you throw on the page, the farther you get from ‘simple.’”
A simple overview should answer the questions: Why do I care? What am I doing? What should I expect to get out of it, “And then absolutely be able to drill down in detail from there.”
Champion Quick Wins
“The death of many a Data Architecture [or] strategic initiative was that they didn’t stack up enough quick wins up front.” Asking stakeholders to wait six months before starting to see any value is a mistake, Yuhasz said. Be able to articulate at which points they will start to see some real value.
Be Specific About Outcomes
Although it’s tempting to make big promises up front, providing quantifiable outcomes in relationship to the amount of disruption makes it easier to get through transitions with fewer people grumbling. Metrics that illustrate the value of a project engender a sense of cooperation and trust: “We’ll be slowing down by 25%, but the end result will make us 50% more trustworthy.”
Understand the User Community
Stewards, governors, quality admins, analysts, report authors, developers, data scientists — they each have a role in curation and governance. Some will want to focus on determining value to the business and others will be more focused on operations. Listen to learn how each of them prefers to work, recognize the constraints they are bound by, and then empower them, he said, but with guardrails. “You must figure out a way that lets your users do what they must do, while both protecting your firm and providing a promotion path for each individual’s insights and discoveries.”
Communication is Key
Clarity about Data Architecture in the ecosystem is paramount, he said, and how it informs your governance and analytical communities. Also important is clarity about where data will be stored, how to access it, and how the firm will get value from the data. Outline places where experimentation with data is allowed. Align platforms and technologies with the abilities of stewards and stakeholders and talk about how iterations will occur.
Let the Culture Be the Guide
“We all like to say it’s about people, process, and technology,” Yuhasz said, but ultimately, the culture decides how well the people, process, and technology will work together. “Cultural readiness, the skills, the platforms, and adoption: Those will yield your outcomes.” Even with a very basic skills platform and very narrow adoption, he said, good outcomes are possible.
Want to learn more about DATAVERSITY’s upcoming events? Check out our current lineup of online and face-to-face conferences here.
Here is the video of the Data Architecture Online presentation: