The Median Foundation Makes Four Grants a Year

Last week I ran a small Facebook ad for GrantSnag. Three day flight, fixed end date, $53 a day. Over the first two days it spent about $106 and sent 403 tagged visitors to the site.

That works out to at most 27 cents a visitor. For paid social that is genuinely cheap, and I was pleased with myself for roughly a minute.

Then I grouped those visits by which page people landed on and looked at.

One row

Every one of the 403 visitors hit the front page. Grouped by page, the whole campaign came back as a single row.

Not one of them reached the search. Not one reached pricing, or signup, or a blog post. There were 412 pageviews against 403 visitors, and the nine extra were people reloading the same front page. Nobody took a second step.

We had already written the rule that decides what happens next: at the end of a run, if an angle produces no attributable signups, it gets no more money, no matter how good the cost per click looks. That rule fired. The angle is done.

There was one genuinely tempting way out. A new subscriber did sign up during the window. I could have counted her, written "1 conversion" in the log, and kept the campaign alive. She had created her account nineteen days earlier, used the free product for most of that time, and abandoned a checkout on her first day. She was going to subscribe whether or not that ad ran. Counting her would have meant paying $159 to keep buying front page views, on the strength of a customer the ads did not produce.

Cheap traffic that converts at zero is still zero. Twenty seven cent clicks do not fix that, they just make the bill arrive slowly enough to feel survivable.

I have watched nonprofits make the identical mistake

The shape of that error was familiar, and it took me a day to place where I had seen it.

It is the shape of a grant prospect list.

You run a search, it returns 118 foundations, and it feels like a productive morning. The number is large, the names are unfamiliar in an exciting way, and the list goes into a spreadsheet. That count is the same metric as my 27 cent click. It measures the top of the funnel and tells you nothing about whether anything is behind it.

So here is the number underneath it, from our own data.

What a foundation actually does in a year

GrantSnag holds the grant records off every 990-PF, which means we can count not just which foundations exist but how many checks each one wrote. Taking 2023, the most recent year whose filings have substantially all arrived:

Grants made in 2023	Foundations	Share of foundations	Share of all grants
1 to 5	77,199	58.2%	7.0%
6 to 10	17,946	13.5%	6.5%
11 to 30	24,664	18.6%	20.8%
31 to 99	10,673	8.0%	25.1%
100 or more	2,203	1.7%	40.6%

That is 132,685 foundations and 2,134,734 grants. The median foundation made four grants in 2023, and at least a quarter of them made exactly one.

Read the first and last rows together, because that is where the whole argument lives. Nearly six in ten foundations made five grants or fewer, and between them they accounted for 7% of all the grants made that year. Meanwhile 1.7% of foundations made 40.6% of them.

What that does to a list of 118

Most of those 118 foundations have about four seats a year. Several have one. And those seats are not empty, waiting for a well written letter of inquiry. They are mostly occupied by organizations that same foundation funded last year, and the year before, because small family foundations renew far more often than they go looking.

A 118-result list is not 118 chances. It is a handful of real chances buried in a hundred or so foundations that were never going to fund anyone new this year, and the list itself gives you no way to tell which is which. So you work it top to bottom, spend six weeks, and get the outcome I got: a big encouraging number at the front and nothing behind it.

The count felt like progress because counting is easy and qualifying is not.

Pick the metric one step past the one that feels good

That is the whole lesson, and it cost me $106 to relearn.

For our next ad, the measure is no longer clicks. It is second pageviews and checkout sessions opened. Those are harder to move and much smaller numbers, and they are the only ones that have ever predicted revenue here.

For your prospect list, the equivalent question is not "how many funders did I find." It is:

How many of these have made a grant this size, to an organization like mine, in the last two years?

Ask that of a list of 118 and it will not survive. What is left will be shorter than you want and worth more than the whole spreadsheet was.

This is also why peer search works backwards from an organization instead of forwards from a filter. Start from a nonprofit doing work like yours and the funders that come back have already answered the qualifying question, because they already wrote that check to that kind of organization. You are not narrowing a directory down toward fit. You are starting from fit and expanding.

Four grants a year is not much of a target. It is a much better thing to know before you spend six weeks aiming at it.