CDAO Role Challenges: Data Defense, Offense and Short Tenure
In a talk at the Data Science Innovators Conference 2023, Kaiser Permanente's Faizan Javed explained why chief data and analytics officers average about 2.5 years in the role. He outlined common obstacles and a playbook built on data defense, data offense and measurable business value.
By The AI Enterprise Magazine Team4 min read
Key takeaways
- Short tenure. Javed said CDAO-type roles average about 2.5 years, versus roughly 10 for CTOs, CIOs and CFOs.
- Defense and offense. Governance, privacy, quality and literacy come first; business-impact applications are the end goal.
- Four perils. Unclear scope, weak reporting lines and resources, low data literacy and limited stakeholder support.
- Know the maturity stage. First-generation leaders build foundations, third-generation teams spend most time on applications.
- Measure and start small. Document results, aim to be a profit center and avoid boiling the ocean.
Chief data and analytics officers face a role with growing influence and a short shelf life, according to Faizan Javed, Chief Data Science Leader at Kaiser Permanente. In a talk at the Data Science Innovators Conference 2023 titled "The Promise and Peril of the Chief Data & Analytics Officer (CDAO) Role," Javed outlined why leaders in these jobs struggle and offered a playbook for getting through the first years.
His advice is not limited to chief data titles, he said. The same challenges show up in many chief and executive roles that own data strategy and leadership.
Where did the chief data officer role come from?
Javed traced the earliest sighting of the role to 2002 at Capital One, where a chief data officer was hired to build out data governance and compliance in financial services. Since then, he said, the role or similar ones have spread across a wide range of companies. The captions garble the adoption statistics he showed, so they are not repeated here.
His explanation for the rise: data volume has exploded, and companies want to apply it to business use cases and customer value. "Data is like oxygen" for any application built on machine learning or analytics, he said.
The promise comes with a warning. According to Javed, the average tenure for people in these and similar roles is around 2.5 years, well short of other executive roles such as CTO, CIO and CFO, which he put closer to 10 years.
What does a CDAO actually own?
Javed split the job into two halves. Data defense covers governance and privacy, data strategy, collection, storage and retention, security, data quality and compliance with rules such as GDPR in Europe and HIPAA in healthcare. It also includes data advocacy and literacy: educating the organization, building partnerships and developing a talent pipeline. The goal, he said, is to facilitate data consumption across the organization, not to hold data for its own sake.
Data offense starts once the foundations are in place. It means choosing the tools and infrastructure, such as cloud environments, machine learning platforms and a clear path to production, and then building reporting, predictive analytics and machine learning applications. Javed said this is where business impact comes from. He offered a lawn analogy: knowing how to mow matters, but a proper mower beats a pair of scissors.
Why do CDAOs struggle?
Javed listed four common problems:
- Unclear scope. "If you don't know what's expected of you, you cannot succeed," he said, stressing the need to know your authority and what you are accountable for.
- Reporting structure and resources. A CDAO who reports to a CTO or CIO may face internal conflict and weaker funding, and data science is resource heavy.
- Low data literacy. Organizations can resist change, and Javed noted that working with people can be harder than working with software.
- Limited stakeholder support. Business units may have internal conflicts that a data champion must help bridge.
How do you gauge data maturity?
Javed described three generations. In the first, the leader builds the foundation: defining strategy, selling the vision, assembling a team and focusing on data defense, with quick wins to show value. In the second, the leader builds on that foundation and moves toward business-impact applications while continuing to improve infrastructure.
By the third generation, most foundation work is done and teams can focus on advanced data science applications, reusable data sets and feature stores. At that point, Javed said, data leaders can spend more than 90% of their time on applications and not on infrastructure.
What is the playbook for a new CDAO?
Javed's recommendations came down to a few moves:
- Set clear goals and expectations for the maturity stage you are entering.
- Aim to be seen as a profit center. A cost-center label, he said, invites cost cutting.
- Measure and document results. "If you can't measure it, you can't improve it," he said, giving gross revenue from an application as an e-commerce example.
- Hire a mix of external talent and in-house upskilling bootcamps, since hiring is hard. Roles he listed include analysts, scientists, engineers, data product managers and architects.
- Shift from defense to offense, with short- and long-term plans that deliver incremental value.
- Start small. "Don't boil the ocean," he said, advising leaders to begin with a few internal partners and use cases.
Customers, internal or external, want the problem solved first, he added, and will hear about the technique afterward.
Where to hear more from data and AI leaders
Leadership and strategy sessions are a staple of AI Connect programming. The AI Enterprise Conference returns Tuesday, August 31, 2027, at Pier Sixty in New York City, bringing enterprise data and AI leaders together. Details are at /events/ai-enterprise-conference/; registration for 2027 events is coming soon.
Frequently asked questions
What does a chief data and analytics officer do?
Javed split the job into data defense (governance, privacy, quality, compliance, literacy and talent) and data offense (tooling, infrastructure and ML and analytics applications).
Why is CDAO tenure so short?
He pointed to unclear scope and expectations, reporting structures that limit funding, low data literacy and limited stakeholder support.
What are the stages of data maturity?
Javed described three generations: building the foundation, building business-impact applications on it, and focusing mostly on advanced applications once foundations are done.
How should a new chief data officer show value?
He advised setting clear goals, measuring and documenting results, aiming to be seen as a profit center and starting with small use cases.
This article is based on remarks by Faizan Javed, Chief Data Science Leader, Kaiser Permanente at Data Science Innovators Conference 2023 on April 12, 2023. More coverage: More AI Connect Events.
Filed under: People, Culture & Leadership, AI Strategy & ROI
