- Most companies are eager to unleash data-driven insights but struggle to unlock self-service workflows, according to a Forrester report commissioned by Capital One Software. The analyst firm surveyed 150 data science, analytics and cloud technicians and decision-makers at North American companies in May and June.
- Nearly all data professionals see the potential business value in spreading insights across the enterprise. Nine in 10 respondents said their organization has already adopted or is in the process of implementing a self-service data strategy and 86% believe easing workflows is critical to business success.
- Cloud provides the infrastructure and tools to democratize workflows, but successful self-service strategies depend on organizational realignment as well. “Technology is at most half of the solution,” Patrick Barch, senior director of product management at Capital One Software, said. “Culture is the other half.”
Tools aren’t much help until you learn how to use them, and neither are data. As the speed, agility and capacity of data technologies advances, organizations are struggling to keep pace.
Nearly 3 in 5 respondents acknowledged their company had difficulty initiating a self-service strategy. Just over half had trouble securing funding and roughly the same proportion cited skills deficits and lack of data training as an obstacle to adoption.
“You can roll out the greatest tools, the best-of-breed for each use case, but unless you're changing the environment that your teams work in and making data stewardship part of their day-to-day responsibility, you're not really going to get anywhere,” Barch said.
Capital One was early to the cloud and quick to discover and overcome the difficulties of modernizing enterprise data ecosystems. Drawing on its data engineering expertise, the bank launched a technology unit last year, rolling out Slingshot, an enterprise data management solution for Snowflake Data Cloud users.
The company has also democratized access to ML tools across business units to encourage collaboration and improve fraud detection.
“If you have the intent to share data across business units, self-service needs to be a table stakes requirement,” Barch said.
Implementation requires a clear understanding of stakeholder motivations.
“On the cost side, you've got business teams that are running workloads, and they're always going to optimize for performance and getting their job done quickly, which is in conflict with the finance teams, who want to spend the least amount of money, and the platform teams, who want to make everybody happy,” Barch said.
With cloud as a starting point, Capital One built a central data platform modeled around shared accountability, said Barch. Each business unit has responsibility and accountability for its data operations, a core principle of self-service.
Initially, it wasn’t easy. “Enabling several seemingly contradictory use cases, while maintaining a user-friendly experience is really hard,” said Barch.
Scaling self-service for decentralized environments, such as a financial services company, was also an obstacle for nearly half of organizations, Forrester found, as was maintaining stakeholder trust in the quality of data.
“If you ask ten different people to define data quality, you're going to get twelve different answers,” Barch said. “When you have all of these different personas operating via self-service, they're going to bring their own perspectives and their own biases as to what quality means for them.”