Show Me Your Nonprofit ID

Anyone who has basked in the rarified world of nonprofit data has bashed their heads into a certain problem: how do you know what you are talking about?

When you first come to the industry, you look at big open data sets like the IRS 990s (the tax forms every nonprofit files annually that are often used as a proxy for financial disclosures) and see they they are neatly indexed by EIN (Employer ID number or Tax ID number, the IRS’s universal identifier for all corporations nonprofit and otherwise.) Excellent, you think to yourself. Combining data from multiple sources will be a snap! Specifying which organization should get money from a foundation will be a breeze! Researching organizations to find out if I want to give will be a walk in the park.

No so. Once you get into it, you realize how many incredibly important cases can’t be dealt with by EIN alone. Here are just a few important examples:

1. An organization that has chapters can still have just one EIN. For example, The American Red Cross has more than 600 chapters, like Red Cross of Greater Chicago, for example. Each has it’s own management, programs, even boards. But only one EIN.
2. Conversely, an organization may have filed chapters all over the country, but still have strong combined identity. Boys and Girls Clubs do this — each city has a separate entity (e.g. Boys and Girls Club of Chicago), but they also organize together as Boys and Girls Club of America. Which do you mean when you say “Boys and Girls Club?”
3. An organization can be a fiscal sponsor. Imagine something terrible just happened. Like the Boston Marathon bombing. That very day someone wants to set up a relief organization like #BostonStrong. But it can take a year or more to get set up as a 501(c)3 (the federal designation for nonprofits) and you’ve got hours. Fortunately, the law allows you to pick a “fiscal sponsor”, an already incorporated nonprofit that can take money on your behalf. But the sponsor has their own EIN and if you look it up you won’t find anything about #BostonStrong.
4. EIN lookup doesn’t deal with “canonical” vs. common names. There’s an organization here in Chicago that’s been helping kids since the 19th century. When they were founded, they were called Uhlich Children’s Action Network. Now they are strictly “UCAN”. But you won’t find UCAN in the IRS data. And you won’t find Uhlich Children’s Action Network on any of their materials. How do you map between?
5. As an increasingly common alternative to forming their own nonprofit or even seeking fiscal sponsorship, many groups are choosing for formulate as “programs” within an existing nonprofit. A favorite of mine is Cure Violence, which uses healthcare methods to fight violence. But despite being a global NGO with operations everywhere from Chicago to Honduras, they’ve chosen to remain a program of the University of Illinois Foundation. Looking up the U of I Foundation is going to be less than useless when trying to get information about Cure Violence, let alone its own sub-programs like CeaseFire.

Does all this matter? It absolutely does. If you want to find a local volunteer opportunity on a site like VolunteerMatch, given money to victims of a specific disaster on Public Good, or just find out how an organization is using your money on GuideStar, somewhere in the mix this identity problem comes into play.

So how can we make it better?

Many have tried, so far all have failed. I believe there are a few reasons for this, but two that dominate.

The first reason is organizational. Virtually all attempts to create better IDs have been proprietary. GuideStar has its own ID system. So does Good Done Great. So do many others, but if the ID isn’t universal, it isn’t useful. That’s especially true if, as in these cases, you need to pay a fee to get access to the IDs.

The second reason is technical. None of the existing ID schemes are composable or decomposable. What I mean by that is that Boys and Girls Clubs of Chicago might have ID 545452349 and Boys and Girls Clubs of America might have ID 083283111. There’s no way to tell they are related without looking up some kind of translation layer.

One idea to solve this would be to borrow from one of the biggest and most successful ID systems out there — the Internet’s Domain Name System or DNS (that’s the system that tells a computer how to load a search engine when you type “google.com” into your browser.)

DNS names are hierarchical, map to numbers (IP addresses vs. EINs), distributed, open, and composable/decomposable. Once made into full blown URLs, they can even contain the equivalent of program data, that is a directory path. They have both a registered component (the domain) and a component that can be quickly and easily configured by a domain owner on the fly (the hostname.) It’s worth noting the URL stands for Universal Resource Locator, not Way to Find a Web Site — it was always meant to be more globally applicable.

Yes, charities may finally be getting their own TLDs (Top Level Domain, like .com), but that not what I mean. I’m suggesting an analogue to DNS, not literally using DNS. There are a few reasons for this. One of them is that registration needs to be verified (like an SSL-EV certificate) in order to legitimately map to an EIN. The second is that it must be free and universal (organizations should be auto-enrolled at the domain::EIN level so the system starts out at least as useful as EIN is today.) The last is that this is an organizational mapping, not an information architecture (techies think LDAP rather than HTTP.)

With a few big sponsors acting as registrars and root servers (I’m looking at you, all you folks with proprietary schemes), something like this feels tenable.

Thoughts, those of you who have lived this data nightmare?