In this short case study, MetadataWorks CEO Adam Milward looks back at the work MetadataWorks & the IPCC Data Distribution Centre have done together to build climate data infrastructure users can trust, doubling monthly users and improving user engagement by 19 times...

1. Who is the IPCC?

The Intergovernmental Panel on Climate Change (IPCC) is the leading United Nations body for assessing the science of climate change. Established in 1988 by the World Meteorological Organization (WMO) and the UN Environment Programme (UNEP), its mandate is to synthesise thousands of peer-reviewed research papers into authoritative global consensus reports that directly inform international policy — including the Paris Agreement and the Kyoto Protocol.

The IPCC does not conduct original research. Instead, it coordinates three Working Groups — WGI (Physical Science Basis), WGII (Impacts, Adaptation & Vulnerability), and WGIII (Mitigation) — each drawing on contributions from hundreds of institutions worldwide to produce Assessment Reports roughly every six to seven years. The Sixth Assessment Report (AR6) was completed in 2023; AR7 is now underway.

Underpinning every Assessment Report is the IPCC Data Distribution Centre (DDC) — established in 1997 — the public gateway (ipcc-data.org) through which the world accesses the data, scenarios and observational records that underpin IPCC findings. The DDC is overseen by TG-Data (Task Group on Data Support for Climate Change Assessments) and operates as a federated partnership across institutions in the UK, Germany, the USA and Spain.

2. The Data Problem

Trust is the IPCC’s core product

The IPCC’s authority rests entirely on trust. Governments and policymakers act on IPCC findings — committing trillions in investment and shaping legislation that affects billions of people — because they trust that the underlying science is rigorous, transparent and verifiable. That trust depends, in turn, on the data being findable, accessible and traceable: if the data underpinning an Assessment Report cannot be located, verified or reproduced, the credibility of the report itself is undermined.

“Trust is hard to earn and easy to lose.”

The DDC is the trust infrastructure for global climate science. Its mandate is not just to store data — it is to “archive and provide transparency, traceability, and stability” for the data and scenarios that underpin IPCC findings. Software, methods and people change; but the data must endure and users and policy makers around the world, on all ends of the political spectrum, must have good cause to trust it. That is what makes the DDC’s performance so consequential.

Building an infrastructure our users trust

Because of the diverse range of data owners, the site MetadataWorks inherited was a sprawling, hard-to-navigate resource of more than 5,000 pages, with a bounce rate of approximately 93% and fewer than 1,000 unique monthly visitors.

A 93% bounce rate is not merely a UX problem. It signals a failure of trust: researchers and policymakers arriving at the DDC were unable to find what they needed, verify what they found, or understand how the data related to the report conclusions they were trying to scrutinise.The site was difficult to use and as a consequence, even more difficult to trust.

A structural tension at the heart of the problem

Beneath the surface, the challenge ran deeper than a single website. The DDC sits at the intersection of two worlds that are difficult to reconcile:

• Modern data infrastructure demands highly specialised, rapidly evolving skills, agile delivery and operationally intensive support.

• Intergovernmental systems are built around public legitimacy, governmental oversight and multi-year procurement cycles.

• The DDC had never been centrally funded — it relied on in-kind contributions from member governments, creating year-by-year fragility for a service on which global climate policy depends.

This structural fragility is itself a trust problem. A service that may or may not exist next year cannot be the bedrock of long-term scientific traceability. The challenge was simultaneously technical, institutional and existential.

3. MetadataWorks, IPCC & The Technical Solution

Getting Started

MetadataWorks won a competitive BEIS tender in February 2021, a tender no other organisation had been able to fulfil, becoming the first private-sector organisation ever invited to be a formal DDC partner. The contract ran from April 2021, with extensions taking the engagement through to 2026, spanning the completion of AR6 and the opening of the AR7 cycle.

Transition: earning operational trust from day one

The first phase involved migrating the entire DDC website, data holdings, metadata, guidance documents, Python scripts and map visualisations from the incumbent (CEDA) to a secure AWS cloud environment — with full backups, integrity checks, and no service interruption. Continuity and reliability are the first requirements of any trust relationship; MetadataWorks established both immediately.

A federated catalogue built on a shared standard

The centrepiece of MetadataWorks’ delivery was a genuine federated data catalogue that automatically harvests metadata from partner institutions. A hybrid onboarding model means partners can either federate directly or log in to the tool to contribute metadata manually — all against a shared, stakeholder-endorsed schema built on FAIR principles (DCAT and Dublin Core standards).

Building this schema was itself a trust-building exercise. Before a line of code was written, MetadataWorks convened representatives from across the IPCC Working Groups and DDC partners to agree — collectively — on a standard way of describing the data. Six design principles governed the process:

• Good enough, not perfect — ship a working version quickly, then improve through use.

• Based on user needs — every metadata attribute required a user story justifying its inclusion.

• Use of standards — reuse existing standards (DCAT preferred) wherever possible.

• Impact vs effort — prioritise attributes by user value against documentation burden.

• Stakeholder endorsement — schema endorsed by an agreed focus group of representatives.

• Version controlled — changes managed formally through the stakeholder group.

The result was not just a technical schema — it was a shared institutional agreement about what trustworthy climate data looks like. DOI minting via DataCite gave every dataset a permanent, citable identifier. The schema’s version control meant that changes were transparent and traceable. These are the mechanics of trust made operational.

Service, support and sustained reliability

Beyond the catalogue, MetadataWorks ran the full DDC operation: service desk and wiki, author onboarding, bi-annual user surveys, geographic usage reporting to TG-Data, and ongoing coordination with DDC partners in Germany (DKRZ), the USA (CIESIN/SEDAC) and Spain (CSIC). The site maintained 99% uptime throughout, with ISO 9001 and ISO 27001 certification and a Severity-1 response commitment of one hour.

4. Topline Results

The most direct measure of restored trust is engagement: researchers and policymakers who can find, understand and use the data.

19× increase in user engagement

~2,000 unique users / month (up from <1,000)

~40% reduction in bounce rate (down from >90%)

Cumulative impact across the five-year engagement:

~61,000 global users reached

~209 unique countries accessed the catalogue

~2,955 datasets archived (including ~346 final AR6 figure datasets)

~14,251 data files downloaded

~300 researcher questions answered via service desk

In 2025 alone, the catalogue served approximately 25,000 users from over 100 countries, with 4,006 direct downloads from MetadataWorks archives and 111 service desk tickets resolved.

5. Key Learnings

FAIR data is trust made operational

MetadataWorks is immensely proud of its work on the DDC. AR6 was the first IPCC assessment cycle to implement FAIR Guidelines for data. MetadataWorks and IPCC operationalised those principles in practice: a stakeholder-endorsed schema, persistent identifiers via DOI minting, automated metadata harvesting, and a hybrid onboarding model that reduced the burden on contributing institutions. Our catalogue illustrates that FAIR is not a compliance exercise — it is the mechanism by which climate data earns and keeps the trust of the researchers, policymakers and citizens who depend on it.

That said, as the project draws to a close due to a lack of funding, we can’t help but reflect on the challenges as well as the successes.

Technical trust and structural trust are not the same thing

The most important lesson from the DDC engagement is the distinction between earning trust technically and sustaining it structurally. MetadataWorks and the IPCC team built a catalogue that worked — users could find data, trace it back to Assessment Report figures, and verify the science. That is technical trust, and the 19× engagement uplift is evidence that it was earned.

But the DDC’s funding model, which is subject to political and even individual personnel related adjustments, meant that this technical trust was always sitting on an unstable foundation.

The DDC is not a niche service. Its 61,000 users span 209 countries; the data it holds underpins policy decisions worth trillions. The service is hugely globally significant when it comes to arming the world with the knowledge it needs to help combat climate change.

When knowledge infrastructure of this kind is treated as an optional in-kind contribution rather than a funded public good, the risk of setbacks is real, and the consequences extend far beyond the organisations involved.

If it wasn’t for dedicated academics in university data services, who beyond their remit and funding put their own time and effort into making sure that the data is archived, trust in climate science would be at risk.