A DataIQ peer exchange is a confidential, application-only discussion designed for senior data and AI leaders. The events include practical case studies, small-group discussions, and networking opportunities with a pre-qualified group of data and AI peers. Peer exchanges enable the success of data and AI leaders and their teams by providing insights and strategies on any topic that they may be facing difficulties tackling.
The full learnings of this DataIQ peer exchange are available exclusively to DataIQ clients.
A large UK-based insurance group, operating multiple service lines, is rapidly maturing its centralised data capabilities with Databricks, Microsoft Purview, Unity Catalogue, and a unified governance framework. However, many smaller business units operate outside the central team’s direct remit, creating a “long tail” of governance gaps, inconsistent standards, and unmonitored risks. This DataIQ client sought pragmatic peer advice to create sustainable data governance in federated environments that works in day-to-day operations. The discussion focused on maturity assessment, cultural alignment, policy and risk leverage, data literacy, and the practicalities of federated champion networks.
Start with a maturity and culture assessment
Before implementing structures, data leaders must assess organisational data literacy, appetite, and cultural tendencies. This shapes whether “carrots” (value-add) or “sticks” (risk/compliance) will work.
“Every company is different… If you haven’t done a short, sharp maturity assessment… you’ve kind of missed a big key point because you’re already starting on the back foot.”
Mapping appetite also reveals pockets of strong practice that can be scaled, and resistance points where engagement strategies must differ. Additionally, benchmarking against industry peers can set expectations and direction.
Embed accountability at the top before cascading
Effective federated governance begins with senior leadership acceptance of clear roles and measurable objectives for data governance. Data leaders must link these to organisational processes, such as performance scorecards or OKRs, to ensure visibility and consequences.
“We put it into people’s objectives… the C-Suite had it in theirs. Even side-of-desk roles became standardised expectations.”
Use policy and risk frameworks as the backbone in risk-averse organisations
The data leaders defended that group-wide data governance policies linked to the enterprise risk management framework provide a “stick” that works in regulated, risk-averse cultures. This ensures all business units must attest to compliance, regardless of structure.
One peer strengthened this model by adopting a “1½ lines of defence” role: advising the business on risks, controls, and attestations, while leaving second-line assurance intact. This bridged governance and risk functions, helping business units understand expectations and design effective controls. Coupled with mandatory policy attestations and performance measures, it ensured governance obligations were met even without direct operational control.
However, the leaders stress that policies should define “what good looks like” and be supported by business partners to translate them into action.
Align with existing process owners
Where possible, data leaders must link data ownership to established operational roles (e.g., process or risk owners) rather than creating entirely new structures. This leverages existing authority and avoids duplication.
Policies should include clear, measurable controls, such as retention, deletion, and third-party data management, aligned to these ownership structures.
Focus on new developments to drive change
In complex ecosystems with legacy systems, the data leaders advised prioritising embedding governance in new projects, migrations, and platform builds. This avoids retrofitting challenges and progressively builds a governed core.
Simultaneously, automating quality checks and classification where possible was recommended to reduce the manual burden on federated teams.
Tailor engagement: carrot for adoption, stick for investment
Peers found risk and compliance arguments worked best for securing executive buy-in and investment, while operational staff responded better to benefits such as reduced rework or better decision-making.
“The stick worked with the exec to understand why the investment was required… the carrot worked for those on the ground doing the work.”
Treat governance as an evolving journey
Accept that federated governance will change with business priorities, leadership, and technology. Build flexibility into models, and revisit structures when they lose traction. Go beyond “gold-plated” frameworks and focus on pragmatic, lived processes.
The full learnings of this DataIQ peer exchange are available exclusively to DataIQ clients.
