How do I obtain demographic data for a given list of people?
July 12, 2023 6:47 AM

I want to obtain demographic data on specific people who currently exist in a non-profit CRM. Disclaimer right up front: We will NOT be using this information to market to anyone, at any time. We will not be selling this information to any other parties. We will not be attaching this data to our CRM; hence we will not know the demographic data of any specific person once we get it. The data will be disaggregated immediately upon receipt.

Current situation: Nonprofit org with a database of several thousand active and non-active participants and addresses. Assume the active participants' addresses are correct; we don't know the accuracy of the non-active participant addresses. We do not have demographic data for the majority of the participants (like 95% majority). We also have birthdates and gender/gender identity, which are accurate, for all active and non-active participants.

Problem to solve: We would like to have some additional demographic data. Ideally, race and ethnicity, income level, and veteran status. We ask these questions when people sign up but the limitation of our CRM is such that we cannot, I repeat, we cannot, require them to answer as a condition of participation. This is non-negotiable with our CRM and believe me, we have tried.

Current ideas: First, a survey sent out to all participants requesting the data. This is ideal because we'd be above-board and honest about why we are asking for this data (mostly for community impact, are we living our mission, some funders require it, etc.). But we also cannot make answering mandatory, because, like, it's an email or snail mail survey. People can and will just ignore it. So the fidelity wouldn't be great. Still, we're probably going to try that first. Second, a data buy. This is where I get skeeved out a little, because privacy, etc. But see the disclaimer above. I just want to know if we're serving the community we say we're serving! Literally that is all! So I think this might be a good solution?

WHAT I WANT TO DO TO GET THE DATA: I want to create a .csv or other Excel document populated with the participants' names, addresses, and birth dates (?) and I want to send it .... somewhere, and get back the demographic info I desire. Is that a thing? Can that be done? Again, if we could simply require this information upon sign up, we would. We know other orgs do this. We know schools do this. We cannot get our CRM overlords to bend on this at all. Can I do this via the US Census? I've played around on their site for a while and I don't think I can. Is there a reputable company out there that will do what I want? I realize that this probably won't be a one-time thing; the data will need to be refreshed at some point (I'm thinking once a quarter to start and then depending on how static it is, we'd go from there), especially since we get new participants all the time.

Tl;dr: I want to potentially buy some demographic data, in an above-board way and as ethically as possible. No marketing, no selling, no sketchy reasons. How do I do this?
posted by anonymous to Work & Money (10 answers total) 1 user marked this as a favorite
There are indeed services that will append some or all of this data but in my experience/opinion the quality/reliability is... variable. I know that Blackbaud used to do some of this but I don't know if they still do. They would be a good place to start, at any rate, and would probably give you leads on other companies doing this if they no longer or don't offer some of the data types you're looking for.

Specifically for income the magic term is "wealth screening".
posted by rhymedirective at 7:09 AM on July 12, 2023


There's probably a company that does this service, but I didn't immediately see one.

Looks like you could get veteran status from the VA for free (see Status > Post), though not as easy as sending off a csv, and there's some kind of application process to get access to the data, which probably wouldn't be a problem with your org being a nonprofit.
posted by gregr at 7:17 AM on July 12, 2023


Can I do this via the US Census?

No, the US Census will not release PII for 72 years, so the most current data that you would be able to get is from the 1950 census.
posted by yuwtze at 7:25 AM on July 12, 2023


I think the term you're looking for is "Data Enrichment."
posted by nkknkk at 8:11 AM on July 12, 2023


A survey that enters participants into a raffle for high-value prizes (ipad, gift cards) is the classic technique here, and it does work. I'd put your budget for data gathering toward good prizes.

I would be hesitant about paying for the data grab because you'll still have to demonstrate whoever you are sharing this with (funders? public report?) that this information is accurate and ethically gathered. From what I've read, those companies often have inaccurate data, and regularly pad their datasets with AI-generated assumptions.
posted by veery at 10:47 AM on July 12, 2023


Append services are pretty damned iffy, as mentioned above, and I suspect that when they do work they skew older/whiter/more male for various reasons. So, caveat appentor who's attempting to get demographic info.

If you want to take this query to the community that thinks about this sort of thing the most, I suggest sharing your question on the PRSPCT-L community at APRA https://prspctl.aprahome.org/community

Maybe I'll see you there. It's one of my hangouts.

What's APRA?
formerly= “American Prospect Research Association”
now = “Association of Professional Researchers for Advancement.”
posted by Glomar response at 11:07 AM on July 12, 2023


Instead of getting specifics on your supporters, you could get at this through market segmentation analysis using postal codes. Comparing these postal codes - which are imperfect on the specific level but broadly useful for getting non-invasive information - to the census-defined demographic segments will be, I think, illuminating enough for your purposes. As a benefit of this, these zipcodes are usually non-identifying information about people so you can feel good about this type of research on an ethical level.

The hard part is finding out who does this for you, since you do have a list of postal codes and that's too much for a one-by-one exercise. This company has individual zipcode lookup , but you may be able to contact them to see if they do analysis of existing lists of zip codes.

Hopefully that's a helpful place to start your search.
posted by urbanlenny at 11:32 AM on July 12, 2023


I’m a digital fundraising consultant. We use Experian Mosaic profites (in Australia, but there’s a U.S. version as well, assuming that’s where you are) to gather this information at a postcode level. It’s not cheap, and you won’t get detailed per-person data, but it’s very useful as an indicator.

Slight tangent: Do you have a data policy, or someone on your team who looks after data governance? Sending names, addresses and birth dates would be out of compliance with the data policies of most charities here, and if that information was to fall into the wrong hands it could severely damage your charity’s reputation.
posted by third word on a random page at 2:27 PM on July 12, 2023


When your clients gave you their data, did they give you permission to send it to an outside vendor like this? I'd just be really, really careful about data security and client trust, here.
posted by Ausamor at 5:12 PM on July 12, 2023


If you're not going to target individuals with this information, why do you need it on everybody? Why isn't a subsample good enough if all you want is summary statistics? Yeah, you'll probably have nonrandom missingness, but such is life.

What is your CRM, and why are you not allowed to ask about this stuff at enrollment?

What is the nature of your nonprofit? There are a lot of kinds of nonprofit where I would be pissed as hell to find out my PII had been sent by you to some data broker, because with your csv file comes the implicit metadata "eirias is, for some reason, personally invested in issues relating to anal cancer / suicide / mental illness / abortion." It's maybe okay if it's puppies but ... it's not all puppies.
posted by eirias at 5:59 AM on July 13, 2023


« Older How to trust my gut in a romantic relationship...   |   Where's the city with the most night to see... Newer »
This thread is closed to new comments.