How to maximize the returns from your data

Recent advances in artificial intelligence, computing power, and data storage have ushered in a golden age for analytics. Aware of the tremendous opportunity, companies are hiring data leaders and experts at a feverish pace. The mandate for these leaders: Embark on ambitious transformations to unearth the valuable data residing in discrete silos across the organization and build the right plumbing to pipe them into powerful analytics that can deliver game-changing insights to decision makers and drive product and process innovation. And do it fast.

Yet a bevy of challenges are slowing data leaders down. Data remain trapped in disparate applications. Sourcing and cleansing efforts are still largely manual processes requiring loads of time from increasingly scarce data talent. Architectural challenges slow the pace of change and require large up-front investments. And rapidly evolving data-privacy regulations force data leaders to spend as much time conferring with company legal officers as they do tackling technical challenges.

At times, these obstacles might make it feel like data really are the new “oil,” as many a pundit has asserted. They can be difficult to extract, require significant refining, and are expensive to transport and store. But there’s a path to making data more like water—clean, easily accessible, and available everywhere in an organization. More than 125 senior data executives from across industries gathered in New York City for McKinsey’s North American Data Summit to share ideas on how to “hydrate” their data and advance their organizations’ data transformations.

Following are some of the factors discussed that resonated with attendees.

Build a business-linked data strategy

While most organizations have already begun the work to extract value from their data, fewer than half of the leaders at the summit had a true data strategy designed to deliver business results. Data executives need to focus on the nuts-and-bolts elements to make data work, including a robust governance and architecture, but not lose sight of the need to focus on business impact. Projects should demonstrate value early and often through quick wins, even as more structural data initiatives—such as centralizing company data in a data lake—take place in parallel. One financial-services leader explained how his company created separate data repositories for important projects, but with the caveat that the project team map out, in advance, a clear path for the repository data to later become part of the company data lake. Another data executive explained how, rather than immediately try to connect disparate systems or create one master system, which could take years, he placed a “layer” over the existing systems that presented a complete view of a process, such as the lending process at a bank, that takes place via multiple systems. He created alerts to signal when steps in the process occur (such as loan approval, when an email is sent to the applicant notifying him or her of the approval, and so on), so that even if each step was completed by a different system, teams could see the whole process as if the systems are working together.

Data ecosystems are also becoming a more critical part of data strategies. While many attendees already purchase external data, often from multiple sources, only a handful had yet engaged in meaningful partnerships to enrich data and use them to build new products and businesses or to more effectively engage with customers or suppliers. Leaders compared the current environment to the early days of the Internet and suggested that improvements in creating trusted ways of sharing data, along with broad-reaching use cases such as resource management and smart cities, could be the unlocks to more mainstream data collaboration.

Finally, any data strategy needs to account for complying with new data-privacy regulations, such as the European Union’s General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA). Executives at the summit noted that no company is immune to the regulations, no matter how local its customer base. Data leaders recognized the need to develop systems that provide fast and flexible means for serving customers covered under different regulations and acknowledged that the level of complexity involved in doing so requires the help of technology. “If we don’t use technology to help us comply with regulations, we will lose,” said one company’s legal counsel.

Invest in new innovations to help speed data-to-value

Speakers and attendees also discussed a wide array of plug-and-play solutions, available through data service providers, that can speed up data processes.

  • Cloud-based software such as Amazon SageMaker can take on some of the onerous data-labeling effort by utilizing machine learning to automatically label portions of data.
  • Machine learning tools can also automate other processes, such as identifying and solving many data quality issues and anonymizing data.
  • Software such as Hadoop that enables schema on read can reduce time spent up front on determining the relationship among data, allowing business users to pull up data sets in a variety of formats and combinations, which can ultimately lead to surfacing value-creating insights faster.
  • New techniques that help make complex algorithmic models more explainable, such as local interpretable model-agnostic explanations (LIME), are beginning to help companies show how complex models come to their conclusions, which can cut down on the time data leaders need to spend with legal.
  • On the analysis side, there are a plethora of prebuilt algorithms that can serve a number of common use cases, reducing the need—and the associated time—to build custom models. And processing power accessible through the cloud can speed up model training.

Make cultural change and communication a top priority

Leaders noted that technical and talent challenges are often just a part of the problem; they also face internal cultural headwinds as they try to steer their organizations toward a more data-driven way of working. Data leaders shared some tactics they’ve used to shift the organizational mind-set, including hackathons, data-visualization competitions, and providing a platform for business leaders to share their data successes with the entire organization. While these techniques often prove helpful, leaders agreed that they serve only as supplements to the fundamental driver of success: relentlessly aligning data initiatives to business outcomes, which can be achieved only through regular communication and relationship building with business leaders. As one data leader put it: “Data and technology are tools—ultimately, you’re doing a business transformation,” and getting critical business-owner support requires framing the effort in this way. To that end, leaders expressed the importance of communicating joint data and business accountability for the success of data initiatives, which should be measured by key performance indicators (KPIs) that align directly with the desired outcomes of the initiatives.

Many leaders also noted that data initiatives are a team sport—the business and data staff should work hand in hand, with “translators” serving as the connective tissue that holds the team together by bridging the communication gap that can exist between the business and technical experts. One data executive explained that her data teams actively engaged sales teams to show how new data-driven tools could help them hit their numbers. This not only encouraged cross-functional collaboration but also spurred sales teams to more diligently collect and help label data to make them most useful (and more easily wrangled by data teams).

With most organizations well along in their journey toward a more data-driven enterprise, it’s time for data leaders to show the value. By taking advantage of innovative off-the-shelf solutions and focusing on aligning closely with the business, data leaders can deliver the return on investment that leadership is increasingly anxious to see.

Josh Gottlieb is a specialist in McKinsey’s Atlanta office, and Tony Ho is an associate partner in the New York office, where Kumar Kanagasabai and Kayvaun Rowshankish are partners and Allen Weinberg is a senior partner.