AI in Action
What to Read Next
Every year in December and January, NewVantage Partners (NVP) conducts a survey of data and technology executives in large companies primarily located in the U.S. Every year, we (the authors) collaborate in analyzing and interpreting the results. And every year, we wonder why the survey results suggest that certain aspects of the data environment aren’t getting better faster, or why they sometimes even become worse.
The executives are usually pretty bullish about technology but quite bearish regarding whether their organizations are becoming more data-driven. They also express concerns about the executive roles — most frequently, the role of chief data officer (CDO) — that are charged with making their company’s data environment better.
Email Updates on AI, Data, & Machine Learning
Get monthly email updates on how artificial intelligence and big data are affecting the development and execution of strategy in organizations.
Please enter a valid email address
Thank you for signing up
The exercise of us writing — and you reading — about these results could become a little depressing. But we are somewhat heartened by the 2021 survey for several reasons: a higher-than-ever participation rate (85 large companies), the absence of major COVID-19-related problems for the surveyed companies, the more-than-typical positivity of respondents regarding technology, and some slight improvements in perspectives on the data executive role. Of course, the challenges of changing human behavior and organizational culture remain substantial. To cite only one survey result that illustrates the issue, 92% of respondents attributed the “principal challenge to becoming data-driven” to “people, business processes, and culture,” with only 8% identifying technology as the culprit. This response has remained constant over several years of surveys.
AI Sees a Surge
For the past several years, the survey has assessed companies’ success with the combined category of big data and AI. This year, an impressively high percentage — 96%, up more than 25% from the previous year — said their organizations had achieved successful outcomes. Almost as many, 92%, reported that the pace of big data/AI investment was accelerating — up 40% from 2020. Overall, 81% are optimistic about the future of these resources in their organizations.
We were pleasantly surprised by these findings, since other surveys and market research organizations had suggested that the production deployment of AI systems was a significant problem. In the NVP survey, the percentage reporting that AI applications are in either widespread or limited production reached 77% — although, admittedly, most report only limited production. This seems to be evidence that the experimental period for AI is coming to a close and that more companies are putting AI to work in their businesses.
The optimistic findings in the NVP survey are consistent with those in the 2020 global Deloitte State of AI in the Enterprise survey, which found that 75% of surveyed IT executives expect organizational transformation to result from AI within three years — up 18 percentage points from the 2018 survey. The limits to production deployment are also seen in a McKinsey global AI survey of companies from 2019; 58% of surveyed executives reported at least one successful deployment in products or processes, but only 30% of those companies used AI across multiple business units and functions.
Challenges for Data Executives
In the NVP survey for 2021, 76% of the respondents were either chief data officers or chief analytics officers, or both; chief data and analytics officer (CDAO) roles are proliferating in many organizations. While these executives are clearly positive and optimistic about the technologies they oversee, they are less enthusiastic about their own roles. Just under half of the respondents, for example, said that the chief data executive is primarily responsible for data in their organizations; another quarter said there is no single point of accountability.
Since data is a diffuse resource that’s difficult to manage, it’s not surprising that the data executives are having challenges with it. We think that’s one reason why more CDOs are also taking on analytics and AI responsibilities and becoming CDAOs. It’s easier to show value to the organization with analytics and AI applications than with data management efforts. The survey provides some support for this, with 70% of respondents indicating that offense-oriented data initiatives (involving marketing, sales, and revenue-generating applications) are more important than defense-oriented ones, such as regulatory and compliance issues and minimizing risk. That is consistent with what many data executives tell us they are focusing on.
The prevalence of CDOs has increased markedly over the years that NVP has conducted the survey, but the most desirable background for newcomers to the job has been highly variable over time. The survey asks whether the profile of a successful CDO is an “external change agent or outsider,” a “line of business executive with responsibility for business results,” or a “company veteran or insider.” Each year, responses to that question swing from one profile to another. This year, the largest group argued for “external change agent or outsider,” but who knows whether this attribute will persist over time. There is simply little clarity about what kind of background is best suited to the chief data executive role.
Perhaps the best indicator of the problematic nature of data leadership is that only a third of the respondents feel that the CDO role is “successful and established.” Just under half of the respondents feel the role is still “nascent and evolving.” Some 18% see it as “struggling with turnover.” It’s somewhat good news that the percentage feeling that the role is successful and established has increased about 5% over the past year, and those feeling that the role is struggling with turnover has decreased by the same percentage. From our viewpoint, we do still see relatively short tenures in the job in many companies, but the CDOs who leave always seem to find other jobs quickly.
Some Companies Have Figured It Out
There are clearly some companies that are making progress on these challenging data issues. We encounter — albeit with a relatively low frequency — those that are establishing more data-driven cultures or that have created long-lasting and effective CDO or CDAO roles. Some have even figured out how to make data a revenue-generating business asset.
Over the next several months, we will be publishing descriptions of these successful companies in MIT Sloan Management Review, along with the overall problems they have been able to address. We hope that their stories will help other organizations accelerate their progress to become data-driven with strong governance structures.