Why was the project created?

When the International Aid Transparency Initiative (IATI) was created, it developed an Organisation Registration Agencycodelist to disambiguate different company, charity and government identifiers. Other open data standards have also been relying on this list, but over time we’ve found a number of challenges:

  • One standard should not have to maintain the codelist for everyone;
  • It it difficult to identify when to use one identifier list over another;
  • The IATI codelist lacks clear descriptions of each organisation identifier list;

This project aims to meet those challenges, and in the process to promote wider shared practice when it comes to organisation identifiers in open data.

What is an open organisation identifier?

At it’s most basic, an open identifier is one that can be used, and re-used, within open data. That means distribution of the ID should not be restricted by intellectual property or other concerns.

However, we also want to go beyond the basics. We prefer identifier lists which are:

  • Primary – where there is a direct one-for-one relationship between the ID and the legal entity it identifies;
  • Accessible online – so that users can easily lookup the identifier in an accessible online database;
  • Open data enabled – so that users can download the full list of organisations in bulk;

Are you trying to build a big database of all organisations?

No. We don’t store any information on specific organisations. We only maintain a ‘list of lists’ – a codelist that describes all the organisation identifier providers out there.

Why not use Open Corporates / Global Legal Entity Identifier etc?

Platforms like OpenCorporates.com are doing an amazing job bringing together actual organisation identification data for companies from official registries across the world. If you want to look up company identifiers, it’s probably the best place to start. We’ll be included cross-links to Open Corporates and the LEI Registration Agency codes in our codelist.

But, OpenCorporates, or the Legal Entity Identifier, only cover some kinds of organisations. We provide codelist entries for lists covering charities, government agencies and other kinds of organisation.

What about government agencies?

It is easy to find official lists of companies and charities. Most countries have some sort of formal registration system in place. But when it comes to government agencies, we have found far fewer cases of well maintained official lists that provide stable identifiers.

We’re carrying out research to explore general methods for identifying government entities when no official lists are available.

Why do you include non-open identifiers?

Our codelist includes all the organisation identifier lists we have found in use in open datasets from partner standards. Sometimes this will include identifier lists that do not meet our basic criteria for being an open list.

Through meta-data we will mark these as such, but we still include them as, in most cases, a non-open identifier is better than none.

What about [ edge case / common challenge ]?

As it turns out, organisations are pretty hard to define. There are all sorts of cases where something that most people would agree persists as a continuous unique organisation gets issued with multiple or different official identifiers in the same list. Or organisations change over time through mergers and divisions.

Right now – we’re focussed on the starting problem of working out which identifier a list is drawn from. Our data model has space for human readable ‘need to know’ information about particular identifier lists – but working out how to model all the messy complexity of real-world organisations is a challenge for the future.