About This Episode
In episode 14 of Data & AI Heads, host Ian Allison welcomes Malcolm Hawker, Chief Data Officer at Prophecy and author of the Data Hero Playbook, for a frank conversation about why the data and analytics function consistently falls short of its potential. Malcolm draws on more than 25 years in the industry, including a prior role as Chief Product Officer and time as a Gartner analyst, to argue that the root cause of short CDO tenures and low business adoption is not a lack of data culture or data literacy among stakeholders. It is a failure by data leaders to articulate, quantify, and deliver genuine business value while defaulting to an external locus of control that blames the business instead of examining the product being delivered.
The conversation moves from mindset to operating model, exploring how misaligned incentives push data practitioners toward technology adoption over problem-solving, how the pendulum between data centralization and decentralization reflects unresolved value gaps, and how generative AI introduces a fundamentally probabilistic worldview that clashes with the deterministic rules-based frameworks most data teams have built their careers around. Malcolm argues that adaptable, context-aware data governance, likely delivered through agentic AI workflows, is both the destination and the only scalable path to get there. Ian and Malcolm also wrestle honestly with what that future means for data roles, citizen development, and whether AI accelerates dysfunction or enables transformation depending on whether the underlying operating principles are sound.
Whether you are a CDO trying to extend your tenure, a data practitioner navigating the shift toward business partnership, or an executive wondering why your data investments have not paid off, this episode offers specific, actionable perspective grounded in product management thinking, behavioral psychology, and real-world governance experience.
Key Takeaways
- The average CDO tenure of 18 to 24 months is not primarily caused by lack of data culture or data literacy; Malcolm argues those are symptoms of a deeper failure to quantify and communicate business value to stakeholders.
- Data leaders who blame poor adoption on stakeholder illiteracy or cultural resistance are exhibiting what psychologists call an external locus of control, which prevents the self-reflection needed to improve the product being delivered.
- Treating the data function like a business running a profit and loss requires understanding what data products actually cost, what customers are willing to use, and what value the delta between cost and benefit represents.
- Incentive structures that reward data practitioners for deploying the latest technology rather than solving business problems are a structural root cause of misalignment, and changing those incentives is a prerequisite for sustainable value creation.
- Generative AI operates probabilistically, producing answers that are contextual and conditional rather than binary, which directly conflicts with the deterministic, rules-based frameworks most data governance and quality programs are built on.
- Adaptable data governance, where rules and definitions are tailored to the consuming context rather than forced into a single enterprise-wide standard, is the right model, and Malcolm contends AI is the only way to deliver it at scale.
- The recurring pendulum between data centralization and decentralization is itself evidence that the data function is not delivering enough value to hold the trust of the business, making federated or domain-driven models a symptom rather than a solution.
- Hiring product managers into data teams and running user research practices such as usability labs and focus groups with business stakeholders are concrete steps toward building data products people actually want to use.
What We Cover
Frequently Asked Questions
Why do CDOs have such short tenures, and what is the real root cause?
Research and Gartner survey data suggest the average CDO tenure is roughly 18 to 24 months. Malcolm Hawker argues that the reasons CDOs themselves cite, such as lack of data culture or poor data literacy, are symptoms rather than causes. The underlying issue is an inability to quantify and articulate business value. When stakeholders do not use data products, data leaders too often blame the audience rather than examining whether the product itself meets real business needs.
What is the growth mindset Malcolm Hawker describes, and why does it matter for data leaders?
Malcolm draws on Carol Dweck's research from her book Mindset: The Secret to Success to describe a growth mindset as one that prioritizes learning, adaptability, and customer success over self-preservation or technology deployment. For data leaders, this means responding to low adoption or short tenure by asking what needs to change in the product and the approach rather than attributing failure to the behaviors of business stakeholders.
How should data and analytics teams think about their stakeholders as customers?
Malcolm argues that data leaders should treat business stakeholders explicitly as customers, the same way a product team treats end users. If a product has poor reviews, low adoption, or usability complaints, the product needs to change. He references running UX labs and focus groups when he led a product function, and challenges data teams to ask when they last conducted structured feedback sessions with the business to understand unmet needs.
Will generative AI help or hurt data and analytics teams that are already struggling?
Malcolm's view is that at current course and speed, AI risks automating and accelerating existing dysfunction rather than correcting it. If a data team's operating principles are misaligned, focused on technology deployment rather than customer outcomes, AI will amplify those misaligned behaviors faster. The prerequisite for AI to be a positive force is fixing the underlying operating model and mindset first.
What is adaptable data governance, and why does Malcolm say AI is necessary to achieve it?
Adaptable data governance, a concept Malcolm attributes to Gartner, means governing and managing data according to the specific needs of each consuming context rather than enforcing a single enterprise-wide rule. A finance definition of a customer is legitimately different from a marketing definition. Malcolm argues that managing the exponential combinations of contextual definitions and hierarchies at enterprise scale is beyond what any human team of data stewards can handle, making AI-driven governance processes a practical necessity rather than an option.
How do misaligned incentives in data teams contribute to poor business outcomes?
Malcolm agrees with the observation that data practitioners have historically been rewarded for learning and deploying new technologies because each new tool on a resume commands a higher salary at the next job. This creates a structural incentive to implement the latest platform rather than solve the most valuable business problem. Malcolm's book addresses this directly and argues that tying team incentives to actual business performance metrics is one concrete way to break the cycle.
Does data democratization and citizen development reduce the need for centralized data teams?
Malcolm frames this as a product question rather than a binary policy choice. Some users want self-service access and the ability to mix their own data, while others need pre-packaged, governed outputs. A mature data function should support both segments. The recurring swing between centralization and decentralization, which Malcolm says he has observed at least seven times across his career, reflects an ongoing failure to deliver enough value centrally, not evidence that decentralization is the permanent answer.
About the Guest
Malcolm Hawker
Malcolm Hawker is a data and analytics executive with more than 25 years of industry experience spanning practitioner, consulting, analyst, and leadership roles. Before moving into data leadership, he served as a Chief Product Officer, where he built and managed product teams and ran structured UX research programs. He later became a Gartner analyst covering data and analytics, advising large enterprises on strategy and operating model design. At the time of this episode, Malcolm is the Chief Data Officer at Prophecy and hosts his own data podcast. He is also the author of the Data Hero Playbook, a book that examines why data and analytics functions struggle to deliver value and offers a growth-mindset-based framework for data leaders to drive meaningful business transformation.
Connect on LinkedIn →More Episodes