How to build a successful cloud data architecture data lake
Tip

Developing an enterprise data strategy: 10 steps to take

Consultants detail 10 to-do items for data management teams looking to create a data strategy to help their organization use data more effectively in business operations.

With data and analytics increasingly driving business decision-making in organizations, data management is no longer an isolated technical function. As a result, developing an enterprise data strategy to ensure that business operations have the information they need is becoming a top priority for data management teams.

That puts new pressure on data architects, data governance managers, data modelers and other data management professionals. In the past, many complained they were "all stuck in the basement and no one cares about us," said Donna Burbank, managing director of consulting firm Global Data Strategy Ltd. "Be careful what you ask for, because we're definitely in the spotlight now."

In a Dataversity webinar, Burbank said creating a business-focused data strategy can seem like a daunting task -- and is usually a significant undertaking. But it's an essential element to support data-driven decision-making, strategic planning and digital transformation initiatives, according to Burbank and other data management consultants.

Here are 10 steps they said chief data officers (CDOs) and data management teams should take to develop a data strategy and get it approved and funded by business executives.

1. Align data management efforts with business strategies

Business strategies should be "infused with data and analytics thinking," Gartner analysts Mike Rollings and Frank Buytendijk wrote in an October 2019 report on crafting an enterprise strategy for both data management and analytics. Well-aligned data resources and analytics systems "can fundamentally shape the business and reinvent how it creates and delivers value" to an organization, they added.

Burbank echoed the importance of mapping business goals and drivers to data management programs. "You can manage data all day long," she said, "but what's the business value?" She also recommended translating data management objectives into business benefits. For example, you could say you're working to improve data quality and the use of data internally, "instead of just saying you're doing data governance, which sounds boring."

Key stages of the data strategy development process
The four phases of enterprise data strategy development

2. Assess your data management capabilities

To help shape the details of an enterprise data strategy, you need to collect baseline information on your organization's efforts to manage data and control its use, said Bill Jenkins, managing partner at Agile Insurance Analytics, a professional services firm that focuses on insurers. Such assessments should examine staffing and skills, processes, technologies and organizational culture, Jenkins wrote in a May 2019 post on the website of consultancy EWSolutions.

Other key capabilities to assess include data governance, data literacy and change management, Rollings and Buytendijk said. They added that as you work to identify deficits and gaps, you need to look beyond your organization's data management and analytics teams and the IT department to get a view of existing data assets and competencies across the enterprise.

3. Line up executive sponsors and business advocates

As with any technology-related project, it's crucial to have people from the business on your side when you propose a data strategy and seek approval for new investments in data management. But the need for business support is even greater because of the scope of the work involved and the potential costs of fully implementing an enterprise strategy, Burbank said in the webinar.

"Having somebody else sell it for you is going to be much more impactful than you selling it yourself," she said. "[Corporate executives] expect the data folks to talk about data." Ideally, the early backers will also become internal advocates with other departments and business units when you're working to get the data strategy adopted throughout the organization, she added.

4. Build a realistic business case for the data strategy

Burbank said the business case for an enterprise data strategy can focus on a combination of things: increased revenue, new business opportunities, lower costs and reduced corporate risk. Before getting started, make sure you know your organization's priorities so you can tailor your business case and data management plans to match, she added.

However, you need to be realistic -- and honest -- about expected costs and ROI projections, Burbank cautioned. You may also have to start small, with some proof-of-concept projects to demonstrate the value of new data management initiatives. "We've done these business cases, and it could be absolutely true that we're going to save $2 million if you give us $1 million today. It's still a hard sell," she said.

5. Create a detailed and achievable project roadmap

Alongside the business case, a roadmap that details the proposed deployment schedule for data projects is also a must. Burbank said to be realistic on the roadmap, too, and avoid promising things you can't deliver. She recommended including some attainable "quick wins" to show the data strategy's potential benefits, with a focus on projects that will provide value to multiple business units.

In their report, Rollings and Buytendijk emphasized the need to link a data and analytics strategy to an operating model that enables the strategy to be executed as planned. The operating model should include a roadmap for implementing a unified data architecture as an underlying foundation, in addition to plotting out the proposed initiatives, they said.

6. Strike a balance between planning and implementation

Burbank pointed to the need to put a solid data architecture in place before moving forward on a data strategy. "Some people say, 'Oh, we don't have time to do that architecture -- we're way too busy.' But you're not going to get anything done without an architecture," she said.

On the other hand, you need to balance data architecture design with getting things done for the business. "If you spend six months just documenting a data model, that's probably too long," Burbank said. She suggested doing the required architectural work in smaller chunks if need be to meet the timelines for data initiatives.

7. Document data-related metrics to help measure success

In addition to quantifying an enterprise data strategy's business value, you should track metrics and key performance indicators for your organization's data assets, Burbank said. That typically includes data quality metrics, such as measurements of the completeness, accuracy and timeliness of data sets and the number of errors found and fixed.

It's also important, though, to link metrics on data quality improvement to potential business benefits, she added. For example, you could calculate the cost savings and increased response rates on marketing campaigns that could result from eliminating duplicate customer records and fixing incorrect name and address data. "That's often what gets the buy-in," Burbank said.

Commonly used data quality metrics
Metrics used by organizations to measure data quality levels

8. Market and sell the data strategy in your organization

Evangelism and outreach are crucial both in getting approval for a data strategy and then getting it adopted by departments and business units across the organization. "Communicate, communicate, communicate," Burbank said. "Part of your job is also marketing."

Tools that can be used to help sell the data strategy internally include newsletters, webinars and lunch-and-learn sessions, as well as formal training programs. Burbank also recommended creating a short PowerPoint or other visual presentation for business executives, instead of just dropping a thick binder that details the plan on their desk. "You have to make it compelling," she said. "No one wants to read a 100-page document," at least as a starting point.

A mission statement that succinctly lays out a vision for a data-driven organization is another good way for CDOs and data management teams to promote a data strategy, Rollings and Buytendijk said. But one that's "nothing more than slogans stuffed with buzzwords and clichés" won't help you much, they noted. Instead, the mission statement should outline your intentions, include relevant information on the potential business benefits and "be inspiring and authentic."

You can have the best technology in the world, but it has to be a people-driven thing.
Donna Burbankmanaging director, Global Data Strategy Ltd.

9. Develop a plan to manage cultural changes

Corporate culture can be a big roadblock in getting an enterprise data strategy adopted, Burbank said. "You can have the best technology in the world, but it has to be a people-driven thing." Plan how to manage cultural issues upfront -- a lot of organizations have even set up formal change management programs as part of their data strategy, she said.

There will likely be an ongoing need to manage organizational change, partly because the data strategy itself will change. "Data continues to evolve along with the business's strategy and objectives," Jenkins wrote in his post. "This means the data strategy should be seen as a 'living document' and that it is adjusted with the times and the needs of the organization."

10. Start with the strategy, not technology

You can recommend purchases of specific tools for managing data governance, data quality, data integration and other initiatives as part of a data strategy, "but that shouldn't be the start," Burbank said. "You should do all the other things and then pick the technology that meets all the requirements. Don't say, 'I want to buy product X' and then try to justify it."

Next Steps

6 essential big data best practices for businesses

What a big data strategy is and how to build one

Dig Deeper on Data integration

Business Analytics
SearchAWS
Content Management
SearchOracle
SearchSAP
Close