5 tips for creating a data-driven culture
Enterprises have spent the past decade attempting to build a data-driven culture, whether that means monetizing data or using it to fuel technology initiatives such as AI, analytics and data science.
It’s apparently a tough slog. A 2023 Harvard Business Review (HBR) whitepaper stated, “many companies are having trouble implementing a data strategy that generates business value.” The whitepaper cited a 2022 NewVantage Partners study that found only 26.5% of the Fortune 1000 data executives it polled said they had successfully built a data-driven organization.
That figure was 48.1% in the 2024 version of the annual study by the consulting firm, which is now called Wavestone. But Christina Egea, vice president of enterprise data product at Capital One, which commissioned the HBR whitepaper, pointed to the contrast between the data aspirations of enterprises and the data reality reflected in the Wavestone reports.
“You have this huge chasm: Everyone thinks data is super important for having a great business strategy, but no one thinks they’re doing it well,” she said.
Egea offered five tips for creating a data culture, based on the whitepaper and Capital One’s experience assembling a data ecosystem that aims to make data accessible across the financial services company.
1. Establish messaging around a data-driven culture
Egea cited a lack of business involvement in data strategy as a top challenge in building a data-driven organization. Data specialists can’t piece together a data- driven company on their own, so senior business leaders must support data initiatives.
To that end, organizations need top-down messaging that reinforces the importance of data, Egea noted. The messaging should emphasize that data is central to the enterprise and that employees will be rewarded for meeting customers’ data needs, she added.
“Data strategy absent business strategy is not successful,” Egea said. “We have to be embedding data in the core of how we approach the business.”
2. Build a data ecosystem
Investing in a foundational data ecosystem is another key component of a data strategy. The ecosystem provides a marketplace in which data scientists, analysts, engineers and other employees can use the organization’s data.
Capital One’s data ecosystem starts with a base layer, where data is created and published, Egea said. The company historically created data through batch processing but now increasingly uses real-time streaming applications. Stream processing rapidly analyzes and transforms data, which can be sent to an application or data repository.
Christina EgeaVice president of enterprise data product, Capital One
The next tier is the storage layer. Here, Capital One initially makes data available in a raw form and then transforms that data for more specific use cases, Egea said. Finally, the access layer lets customers tap data in the storage layer or from streaming applications.
Other important ecosystem considerations include data cataloging and metadata, which let users know what data is available and where to find it, and governance policies for managing data access and usage, Egea noted.
3. Treat data as a product
Egea said a divide often exists between business teams that need data and the teams responsible for creating it. Treating data as a product, however, provides a workaround that promotes collaboration between data producers and their internal customers. With the product philosophy, data producers assess the needs of end users and work backward from that understanding, making data available in a way that works for the users, she added.
The HBR whitepaper reiterates this method, noting that a product approach “maintains a strong focus on the internal customer.”
Egea’s team cultivates the backward-working approach to data products at Capital One. She said the team’s focus is on building enterprise-wide data standards and tools, but not determining how data ought to be used and for what purposes.
“A really important aspect of our product mindset is that no one inside the company can define the data intent for every piece of data across the firm,” she said. “Instead, we have to ensure ownership lives within the business areas that are accountable for that data.”
4. Provide self-service tools and training
Top-down communication helps instill a data culture, and so too does a bottom-up approach that provides self-service tools for accessing and using data, Egea noted.
“For everyone in the company to make use of data, which is our expectation, we have to be offering them access to the best tools,” she said.
Data users also need access to training to use those tools effectively, Egea said. In addition, an internal forum where users can learn best practices and share their own methods also fosters employees’ data knowledge, she added.
Self-service can democratize data accessibility, but it faces some obstacles. A 2023 Capital One-Forrester Research study found several factors can hinder organizations pursuing self-service data strategies. Potential obstacles include a lack of user-friendly tools, funding issues and cultural considerations such as insufficient collaboration.
5. Encourage executives to dig into data
Convincing top-tier executives to promote and build a data culture isn’t easy. That’s especially the case when they might be tempted to neglect data management and just push ahead with high-profile AI and machine learning projects.
Egea suggests encouraging corporate leaders to live a day in the life of a data user. Executives who engage with data will soon learn how easy it is to find and use and whether it’s consistent, she noted. Such a data tour can help leadership determine where to begin a data initiative. For example, an organization might have good data but lack useful tools or adequate storage, she said.
“Walking a mile in the team’s shoes to really understand the challenges out there is critically important,” Egea said.
Read More