From Strategy to Impact at Scale: PGGM’s Data Mesh Journey

Transcript

1. From Strategy to Impact at Scale: PGGM’s Data Mesh Journey Data Expo - September 2026
2. Your Speakers 2 Frank Gresnigt Remy van der Vlist Director, Datashift Supports PGGM's Collibra implementation & federated governance Data Quality & Governance Officer, PGGM Data Management Office – Investment Management • 20+ years in data governance, including senior roles at NN Investment Partners & Goldman Sachs • ~20 years in financial services: leadership, governance frameworks, data quality, change management • Built & led Data Governance Centres of Excellence and governance operating models • Leads operationalisation of federated data governance within PGGM's Data Mesh • Pragmatic approach: turning governance principles into practical processes • Defines governance roles, structures and frameworks; enables owners & stewards || Data Expo - September 2026
3. PGGM in a nutshell Investment management - Pension management - Management advice Facts & Figures Assets under Management* 271 Billion Participants 5,8 Million * Balance at 30 June 2026 3 || Data Expo - September 2026 Customers
4. Contributing to a good pension for healthcare workers, in a livable world PGGM IM Strategy: • From 2D to 3D investing (adding the sustainability dimension to the risk & return dimensions) for long term value creation • Total Portfolio steering • Lean and efficient way of working by increasing the use of modern technologies.
5. Challenge: Our data couldn't keep up with our ambition Strategy 2030 raised the bar. Our centralised operating model couldn't clear it. One team, every request Delivery couldn't scale No domain accountability Data development and delivery dependent on central IT capacity Speed fell as demand grew across the business. Data Governance & Data quality was everyone's problem, and no one's job. The biggest challenge wasn't our data, it was our operating model. 5 || Data Expo - September 2026
6. A different approach: Data Mesh Business driven & federated governance To support Strategy 2030, PGGM IM deliberately chose Data Mesh as the foundation of its future data operating model. Data Mesh as an enabler for: • Business autonomy & data ownership • Faster value delivery through data (reduced time-to-market) • Federated governance for scale & consistency • Improved scalability and reusability of data through data products • Data as core capability within the business domains • Improved collaboration, breaking down silos 6 || Data Expo - September 2026
7. Data Mesh as a guiding principle — but foundations first Data set governance Data product design process Closing the governance gap at dataset level Shaping our first design process • Robust, workflow-supported processes • Quality, ownership, management • Structured path, once mature enough • From managing datasets to products Implementation of Federated governance operating model Ownership where it belongs: within the business 7 || Data Expo - September 2026
8. Governance Operating Model for Data Mesh We operationalized it layer by layer with – with Collibra as the key enabler Principles Federated ownership and shared standards, providing clear direction for further implementations. Processes Onboarding new data sets and certification defined as business processes. Playbooks Every (sub)process became a guided workflow, steps, owners, approvals. Automation Checks, handoffs and status updates run without manual chasing. Measurement Dashboards show certification coverage, quality and adoption per domain. COLLIBRA — enables every layer through workflow, automation and transparency. Collibra didn’t define the operating model - it made it executable. 8 || Data Expo - September 2026
9. First steps towards Data Mesh Technology was ahead of governance, this is how we sequenced the fix. ALREADY IN PLACE OUR FOCUS: PUT THE HOUSE IN ORDER OUR FOCUS: PUT THE HOUSE IN ORDER NEXT STEP Self-Serve Platform Domain Ownership Federated Governance Data as a Product Technology led the way — teams were already building on the data mesh platform. Domains own and are accountable for the datasets already running on the platform. Roles made real: clear processes, training and ways of working, not just a name on a dataset. Grow governed, certified datasets into discoverable, trustworthy data products. First, we put the house in order - now we’re moving for real data products. 9 || Data Expo - September 2026
10. Process & Playbook: one lifecycle, every role guided Guided workflows walk each role through every step - nobody reads a manual. Identify Understand Trust Use Data Sets developed on the Data Platform are registered in Collibra with clear ownership Describe it and link it to shared business concepts. Validate rules, quality and lineage to earn certification. Publish,Data Sharing Agreements enable compliant and transparent data use across teams. Data Product Owner Data Steward Data Steward Data Consumer supported by DMO with Business SMEs with Data Engineer monitored by DMO Certification is re-validated using interactive dashboards - the lifecycle never ‘finishes’. Roles aren’t a RACI table on a wiki - they’re built into every step. 10 || Data Expo - September 2026
11. Automation: Certification workflows & monitoring Illustrative example: team and dataset names have been blurred, and all figures shown are fictitious. 11 || Data Expo - September 2026
12. Measurement: what changed in numbers and in culture The numbers grow every quarter, the cultural change is the real shift. 91 4 84 datasets certified domains and 36 subdomains onboarded owners & stewards active BEFORE • Dependency on central IT capacity • Ownership defined on paper • Limited domain steering on data initiative prioritization AFTER • A clear vision of the characteristics and value of a Data Product • Ownership operationalized in the business • Data domains empowered to prioritise • Limited trust and visibility in datasets • • Training & data literacy dependent on individual initiatives Increased business autonomy, supported by central rules • Certified datasets • Organisation-wide data capability built Same people, same data, a fundamentally different way of working 12 || Data Expo - September 2026
13. Data Mesh starts long before the first Data Product Our lessons learned … 1 2 Put ownership in the domains on day one It’s an organizational change, built the needed capabilities first 4 5 Make governance the easy path 13 || Data Expo - September 2026 Implement, learn, adjust, repeat. 3 Standards and autonomy must evolve together. 6 It is a journey ;-) The foundation came first. Data Mesh follows.
14. One more thing. WHAT'S NEXT? FROM GOVERNED DATA TO DATA PRODUCTS 1 2 3 EVOLVE SHIFT LEFT SCALE Datasets → Data Products Governance by Design From Governance to Value Turn certified datasets into reusable products designed around real consumer needs. Embed ownership, quality, metadata, privacy and certification into the product development lifecycle. Measure adoption, reuse and business outcomes, not just governance compliance. The foundation is in place. Now we build on it. 14 || Data Expo - September 2026
15. Questions? 15 || Data Expo - September 2026