The evolution of the data-driven enterprise

by Rob Levin and Kayvaun Rowshankish

At the recent CDOIQ symposium held in Cambridge, Massachusetts, McKinsey senior partner Kayvaun Rowshankish gave the keynote address. The following is a summary of that presentation and includes takeaways from a subsequent panel discussion.

Reading macroeconomic tea leaves is always a fraught exercise, but there are three factors for companies to consider. First is that there are significant tailwinds that favor technology- and data-driven enterprises, from rapidly maturing technologies to more accessible talent. Second, lessons from recent recessions and economic disruption show that those companies that continually invested in innovation during these periods had 240-percentage-point greater shareholder returns than their peers.1 And, third, the economic potential of the new kid on the technology block—generative AI—could add the equivalent of $2.6 trillion to $4.4 trillion annually across the 63 use cases McKinsey recently analyzed.2

These factors are exciting but also underscore the importance of data as the fuel that can power this new horizon of growth. Business leaders need to be clear about not only what has to change but also how to drive that change to create a data-driven enterprise.

The what: Seven data shifts

There are seven key shifts that companies need to make on their journey to becoming a data-driven organization.

1. Cultivate a data-first organization

From: Data-driven approaches are applied sporadically throughout the organization, leaving value on the table and creating inefficiencies.

To: Data democratization enables all employees to leverage data and use innovative data techniques to resolve challenges. Self-service tools, intensive learning journeys, and role modeling from the C-suite down embeds the data-first mindset throughout the organization.

2. Power up data and analytics technology

From: Only a fraction of the data available from across the business is ingested, processed, queried, and analyzed in real time due to the limits of legacy technology and high computational demands.

To: Data is generated, processed, analyzed, and visualized for end users by ubiquitous technologies, leading to faster and more powerful insights. Sophisticated advanced analytics—generative AI, self-serve data, and low-code/no-code platforms—are reasonably available to all.

3. Create dynamic and reusable data products

From: Most usable data is organized using relational database tools. Data engineers often spend significant time repeatedly curating and wrangling data sets for each use case.

To: Data is curated into rich, multipurpose, dynamic data products. This minimizes data engineering and enables high-impact use cases, maximizing time to market and ROI.

4. Treat data like a product

From: Data often has no true “owner,” with data sets stored—sometimes in duplication—across sprawling, siloed, and often costly environments.

To: Data products have dedicated teams, or “squads,” aligned with them to embed data security, evolve data engineering, and implement self-service access and analytics tools. Data product owners lead teams to continuously evolve products to meet user needs.

5. Expand the chief data officer’s role to generate value

From: Chief data officers (CDOs) and their teams function as a cost center responsible for developing and tracking compliance with policies, standards, and procedures.

To: CDOs and their teams function as a profit center. They ideate ways to use data, develop an enterprise data strategy, and incubate new revenue sources by monetizing data services and data sharing.

6. Make data-ecosystem integration the norm

From: Data is often siloed within organizations, and data-sharing arrangements are uncommon because they are arduous.

To: Large, complex organizations use data-sharing platforms to facilitate collaboration on data-driven projects, both within and between organizations. They actively participate in a data economy to create more valuable insights for all members.

7. Prioritize and automate data management

From: Data management, privacy, and security are often viewed as compliance issues, driven by regulatory data-protection mandates. Manual data-resiliency processes make it difficult to manage, recover, and protect data in a holistic, timely, and efficient manner.

To: Companies use AI-based tools and innovative techniques to enable self-describing data (metadata and lineage), improve data quality, and generate predefined “scripts” to provide safe and secure data access to users in near real time.

The how: Rewiring the organization

In many ways, these seven shifts are intuitive and resonate well with data and business leaders. Where the breakdown often occurs, however, is in translating the what into the how. In our experience, the crucial insight for turning data promise into business value is that the scope of the change needs to be both broad and deep—limited change programs can build skills and be useful proofs of concept, but they do not lead to transformational value.

This rewiring journey starts with developing a clear road map that prioritizes domains of value (often an end-to-end customer journey or a process) and identifies the sources of data that can power the necessary solutions. This road map highlights the critical capabilities needed to deliver the highlighted solutions. Critical capabilities include building up an in-house team of technology experts—including data scientists, data engineers, and product owners—and integrating them into agile pods so that data becomes part of any solution development from the beginning. Scaling requires efficient access to and use and reuse of data assets. The core building block is the data product, with a cloud-first technology architecture. The data product is a reusable and easily accessible asset that can be readily consumed by teams or applications.

This level of rewiring is non-negotiable—merely focusing on one or two actions isn’t enough. These capabilities are the engine for building a data-driven enterprise that continually creates value.

Rob Levin is a senior partner in McKinsey’s Boston office, and Kayvaun Rowshankish is a senior partner in the New York office.

1 CPAnalytics, n = 3,000 top public, European Union, and North American companies by market capitalization from 2007–21.
2The economic potential of generative AI,” McKinsey, June 14, 2023.