Introducing OIM Taxonomy: An Easier, Faster, More Powerful XBRL Experience
XBRL International is working a strategic modernisation of the XBRL standard, and that’s what OIM Taxonomy is all about. In this two-parter, we set out where we are, what we’re aiming for, and where – with your help – we’re going next. This post focuses on high-level goals, and the next will dive into new capabilities of the emerging specification. Both posts are based on internal presentations by Paul Warren, XBRL International’s Technical Director.
XBRL is an established standard, enabling regulators, analysts, and other users around the world to work with high-quality, machine-readable business data. It is the backbone of structured financial reporting in most major markets, and is in widespread use in prudential, tax, sustainability and other reporting domains. As the landscape of data consumption continues to evolve, with new tools, new audiences, and the rapid rise of AI, now is the time to take what works and make it even better.
That modernisation is the purpose of the coming OIM Taxonomy specification. We’re overhauling and simplifying XBRL taxonomies, the structured digital dictionaries that underpin XRBL reports. The objective is to build on the strong foundations XBRL already provides and bring the taxonomy model in line with today’s expectations: easier to work with, more consistent, better performing, and optimised for a world where AI is part of how financial data gets used and understood.
Building on what works
XBRL has always been about attaching rich, meaningful metadata to reported data. The evolving standard and new formats have helped make XBRL reports easier to use. Any report can easily be converted into, for example, xBRL-JSON, something people can pick up and work with intuitively, without needing to read a specification.
The opportunity with OIM Taxonomy is to bring that same accessibility and clarity to the taxonomy layer, so that a much wider range of people and tools can realise the full value of XBRL metadata.
XBRL will be easier to use and more powerful for everyone: developers building new applications, analysts consuming structured data, regulators designing reporting frameworks and the AI systems increasingly being used to read and interpret financial reports.
XBRL’s structured approach already gives AI tools a significant advantage when working with business data, enabling more accurate, reliable, and in-depth analysis in comparison with unstructured formats like PDF. OIM taxonomy is about making that advantage even greater, ensuring that XBRL data is optimised for AI tools and other kinds of automated analysis.
At the same time, XBRL is a working standard, successfully deployed in a vast range of current implementations right around the world, and in use by millions and millions of companies. We’re committed to continuing to support and maintain the current specifications over the long term. There’ll be no forced switch. We expect that many environments will be very happy to stay with the XBRL 2.1 specifications stack into the long term. However, we foresee clear benefits to the upcoming specifications, and there will therefore be clear transition pathways for stakeholders that decide to upgrade.
Six goals for a modern taxonomy model
In the requirements document for the OIM Taxonomy specification, we’ve set out six high-level goals that together define what a modernised taxonomy model should achieve:
- Ease of consumption. The rich metadata in taxonomies should be genuinely accessible to developers, to analysts, and to the tools they build. XBRL taxonomies currently tend to be tricky to consume, and that can stop users making the most of XBRL reports. With the removal of unnecessary barriers the taxonomy layer will become an asset people want to use, not something they work around.
- AI optimisation. LLMs and other AI tools are becoming a standard part of how data is processed and interpreted. Simpler and more consistent taxonomies will enable AI tools to make much better use of the information they contain, unlocking new possibilities for automated analysis and insight.
- Intuitive comprehensibility. Someone with a reasonable technical background should be able to look at a taxonomy file and get a good sense of what it’s expressing, without needing to be a specialist. The new modelling approach cuts taxonomies free from the relatively niche XML format and makes them simpler and more intuitive.
- Improved performance. Today’s taxonomies are often very large and spread across large numbers of files, and details of the specification severely limit the opportunities for XBRL processors to optimise consumption. OIM Taxonomy aims to significantly improve performance when working with syntax, both through an inherently more efficient representation and by enabling processors to selectively load portions of a taxonomy as needed.
- Consistent modelling. XBRL is an extremely flexible standard, but this flexibility comes with a cost: different taxonomies take different approaches to modelling the same reporting scenarios. OIM Taxonomy seeks to provide a single, well-designed approach to modelling any given scenario. This will enable tool developers to produce simpler, higher-level tools that work with many different taxonomies, while shielding users from unnecessary technical details. Reducing the number of modelling choices that need to made will also simplify the task of taxonomy creation.
- Improved declarative constraints. OIM Taxonomy will support a greater depth of structured validation logic. When constraints are expressed declaratively rather than as executable rules, reporting tools can use them to guide users proactively, helping preparers get things right from the start rather than discovering issues at validation time. The calculation relationships in XBRL 2.1 are a good example of this principle in action, and OIM Taxonomy extends it much further.
What comes next?
Our staff and expert volunteers are making significant progress towards a first draft of the specification, and we’re aiming to publish a Public Working Draft (PWD) for consultation as soon as possible.
That means we will soon be needing a whole lot more help from across the digital reporting community. Your repeated review and feedback are essential to get us from a first draft to a tested, improved and implementation-ready final specification.
We need you to help us figure out if the specification provides the right solutions to meet the goals we’ve set out – and what’s missing. We need you to help us test the specification against real world use cases, in diverse contexts. We need you to demonstrate successful software implementations in your products, and we enthusiastically invite you to start planning now for everything XBRL Taxonomy will offer. Let us keep you in the loop here.
Meanwhile, if we’ve whetted your appetite, the next post in this series explores some of the more concrete capabilities taking shape as the specification develops. They include new tools for managing taxonomy extensions, a cleaner way to attach metadata to facts, and expanded support for a broader range of report types beyond traditional financial reporting – just a few examples of the possibilities on the horizon.
