It’s Time to Build a Cohesive Data Strategy

George Yuhasz is helping businesses to innovate in another direction: building a rich Data Strategy by holistically and concurrently bringing together analytics, Data Governance, and Information Architecture.

As U.S. Director, Business Intelligence and Data Services, he launched the Data Strategy at the global food services enterprise. It’s been a job of optimizing people and processes, distributing the costs and risks that go along with the initiative, and of gaining incremental value along the way to a long-term and high-yield Data Strategy.

At his recent DATAVERSITY® Enterprise Data World Conference presentation How to Effectively Launch Concurrent Analytics, Data Governance, and Information Architecture Initiatives, he shared some of the strategies behind the work that may be useful to other data leaders:

Get Rid of the Silos: The activities of analytics, Data Governance, and Information Architecture may be siloed across the enterprise. That’s generally seen as a problem but Yuhasz emphasizes the opportunities that lie here. That is, there are pain points at the departmental or organizational level because of silos, and identifying and helping to clear those silos is a path to building alliances.

Yuhasz acknowledges that there is likely competition for territory among disparate parties, which can be a detriment to actually getting things done. However, that competition also can be motivating and good if delivering results is part of the organizational culture. In that case, “focus on scrimmaging and raising the game,” he advised.

Expand the Focus: What technology and practices does the organization spend most of its time on? In healthcare, for instance, attention may be focused on informatics around patient health and outcomes, and getting positive results there by any means necessary. “Things like getting reusable data assets, good definitions, and governance get left by the wayside,” he says.

In this case, a data leader in a healthcare organization can widen the picture to include Data Governance and/or Information Architecture into the equation, aiming at those targets from an informatics perspective. For example, the powerful data leveraged to improve patient outcomes might be reused to improve other aspects of the organization, such as streamlining up-billing. That marries informatics with Data Governance. “The same formulaic approach can fit into other data pockets within the community, so organize around that and plan to be a partner. Come at them with an offer to help improve their operations and refine their requirements,” Yuhasz says. “As a data leader, be a partner.”

Check Out the Lineup: Assessing in-house talent is a smart move – and not limiting the talent you’re looking for to just the very technical side of the organization. That’s important, of course, but equally critical, if not more so, is to look at the business analysts.

“Who gives requirements on what you execute? It’s the business,” he says. Business executives don’t want to provide IT with functional requirements but to talk to someone who knows what they need and can put it together for them.

In the retail sector, for instance, the business may want to understand the performance of all its stores and relate it to their financial plan. Or in banking, it may want to do a risk analysis to learn whether it has enough capital to withstand another financial crisis. The accurate data needed for such Analytics may be difficult to collect if it hasn’t been cleaned and categorized via Data Governance. When data is consistent and there’s also an understanding of where it exists and how it flows across processes within the Information Architecture, business analysts are more empowered to provide the business the answers they need.

Get Broad Sponsorships and Stakeholders: This doesn’t happen all at once, so be ready for a steady pursuit. Maybe there’s a need to come up with a more efficient way to run forecasting models, profitability scores, and so on. That may require analysis of a lot of atomic operational data – everything from production line to supply chain information – that is owned by and spread out among other functions. Perhaps those department leads will need a little help in understanding why their data should be accessible and aligned across the ecosystem within an enterprise Data Governance framework. Technology partners need to become invested, too – the Information Architecture must be able to accommodate storage, and direct the flow of data as well as its security.

“You need the right connections between technology capabilities and producers and consumers to give you clarity, context and priority,” Yuhasz says.

Assemble the Board: You don’t have to overwhelm everyone with the concept of creating a Data Governance board. Call it by any name you like.

It can be facilitated by the primary and secondary stakeholders a data leader has cultivated, and it can take on the work of finding and addressing overlapping pain points to spread risks as well as benefits, and also allocate resources to those with problems to resolve them. In a connected space, “you have the platform to review project portfolios and to work on data issues that come up in the lines of business and their portfolios,” he says.

Having a platform is a good way to overcome the problem that perpetually perplexes most organizations. “They have data issues, but they don’t have a fix,” according to Yuhasz. With this platform, issues like Data Quality can be raised and addressed, and data flow issues, if not resolved in the near-term, will at least be captured, he points out.

The individuals on the board also act as a review board for data issues that arise in capital projects with the help of line-of-business representatives – data champs that can act as stewards, to curate the data and build and sustain the decisions made about data definitions and use of enterprise assets that will reside in and be accessed by a data enterprise hub.

There should be a charter in place that provides a list of functions for the working group – for instance, to provide oversight, to govern and prioritize Data Strategy, objectives, and investments.

The minutes of stakeholder meetings should be published too, along with notification of the data assets available for use. “Market the heck out of it,” he says.

For project portfolio, capability and reference architecture frameworks, “a guiding principle is to do things once and do them the right way,” Yuhasz says, but data leaders have to be practical about bringing all the pieces together. “Sometimes only part of it can be done the right way and the rest put on hold to a later date. You may need the space for that.”

A final word: Amid the complexity, simplify as much as possible. The focus should be to ensure that “data adheres to good IT and business practices.”

Want to learn more about DATAVERSITY’s upcoming events? Check out our current lineup of online and face-to-face conferences here.