We only love you when you pay. (The small-donor drought)

"Our small-dollar program isn't working the way it used to."

I have heard that sentence in a lot of countries. Always in more or less the same tone — slightly apologetic, the way you would own up to a bad quarter or a hire that didn't work out, with an implied second half hanging off the end of it. And we're not sure what we're doing wrong. Different countries, different electoral systems, different causes. Organizations that have never heard of each other and would not agree on much if they had.

When you hear an identical diagnosis in a dozen unrelated rooms, on three continents, from people with nothing in common except the problem, it is not a local failure of execution.

Here is the part that keeps everyone calm about it. The topline is fine.

The Fundraising Effectiveness Project reported total dollars up 4.3% in the first quarter of this year. Donor counts down only 0.8%, which is a real improvement on the year before, when the losses were three times that. Every segment recovered — small, midsize, major, supersize. Both dollars and donors grew across all four.

Except one. Micro-donors, the people giving between one and a hundred dollars, are the only group still leaving.

And here is the detail underneath it that I think matters most, because it is the one nobody quotes. Retention among those micro-donors actually improved this quarter. It fell in every other segment and rose in that one. So the people already there are staying a little better than before. The file is still shrinking anyway.

Which means the problem is not that the smallest givers are walking out. It is that nobody new is walking in.

I should be fair about the rest of the number, because it costs me nothing and the honesty is worth more than the point. Some of that growth was borrowed rather than made — donors moving gifts forward ahead of expected changes to the tax treatment. Take that out and the picture is flatter still.

So: giving recovered, the small-dollar pipeline did not, and the toplines look healthy enough that in most organizations nobody has had to explain it to a board. The base thins quietly, the dependence on large gifts grows back, and the independence that small-dollar money bought us drains out of the sector over a decade without ever producing a bad quarter.

Now let me say the thing I actually believe about that money, before I criticize what we did to it.

The shift to small-dollar fundraising was one of the best things that happened to progressive politics in my lifetime. It built a source of power that did not answer to nine people in a room. It let candidates who could never have got a meeting with a bundler run and win. It meant an organization could take a position that cost it a major donor and survive. That was real, it mattered, and I am not nostalgic for the alternative.

What broke was not the model. It was what we did to the model under resource pressure.

Data became the currency. Lists were built through petitions, then bought, then swapped, then rented, and the same names accumulated on more and more of them. Then came the boom — the years when a crisis and a screen and a credit card were enough, and the returns were good enough that everyone learned the wrong lesson from them. The lesson we learned was that the answer to a soft quarter is more volume against the same people.

Then the optimization arrived, and it was genuinely good at its job. Analytics took over the decisions about who to ask, when to ask, and in what words. Then machine learning took over the decisions the analysts had been making. Send-time optimization, ask-string optimization, predictive scoring, suppression modelling. Whole departments — some of them very good, staffed by people I like and respect — built entirely around the question of how to extract slightly more from a list without killing it.

And somewhere in there we dropped the one thing we would never, under any circumstances, drop for a major donor.

We stopped building relationships with small donors. Not because anyone decided small donors didn't deserve one. Because at that scale, a relationship is a resource drain and a send-time test is not.

Consider how strange this is when you say it plainly. We know exactly how to build a fundraising relationship. We are extremely good at it. We do it every single day, patiently, over years, for the person giving fifty thousand dollars — we learn their kids' names, we ask what they care about, we tell them what happened with the money, we call when there is nothing to ask for. We decided the person giving fifty was not worth the same treatment.

And then, when they stopped giving, we called it donor fatigue.

We only love you when you pay.

So here is what I have come to think is the real shape of this. Fifteen years of optimization touched every variable in the equation. Subject lines. Send times. Ask amounts. Match claims. Deadline framings. Segments and micro-segments. The name on the from-line, the length of the first paragraph, the colour of the button, the number of exclamation marks. All of it, endlessly, measured to two decimal places.

The one thing that never changed is who was doing the asking.

It was always the institution. Fourteen years of testing, and in every single test, on both sides of every split, the asker was an organization sending an email to a person who does not know it.

That is the variable. That is the one we left alone. And it happens to be the one the evidence says carries the most weight.

Start with the most boring finding in the whole literature on giving, which is that people give because somebody asked them. Bekkers and Wiepking, reviewing the empirical research on philanthropy, found that between 85 and 86 percent of donation acts follow a solicitation. Giving is overwhelmingly a response, not an initiative. Almost nobody wakes up and decides to fund something.

Then add the part that is harder to design a study around, and which Jonathan Meer managed to design a study around anyway. Using university alumni data, he used the fact that first-year roommates are assigned rather than chosen — a genuine natural experiment sitting inside a fundraising programme — and found that social ties play a strong causal role in both the decision to give and the size of the gift. A request works considerably better when it comes from somebody the person asked has something in common with.

Put those two together and you get something that ought to have reorganized the industry and did not. Giving is a response to being asked. And who asks changes the answer.

Every practitioner already knows this in their hands. Ask any organizer with an active chapter what happens when a chapter leader emails their own list instead of headquarters emailing it. They will tell you the numbers are not close.

So the answer to a thinning small-dollar file is not fewer emails. It is not better emails either, which is the more sophisticated wrong answer and the one most organizations are currently paying for. It is a different asker.

Which is why the small-donor crisis is not a fundraising problem at all. It is an organizing problem wearing a fundraising problem's clothes. You cannot solve it in the fundraising department, because the fix requires something the fundraising department does not have and has never been resourced to build: several hundred people who are willing to ask on your behalf, in their own voice, of people who actually know them.

And that is where it stalls, every time. Not because organizations don't want it. Because the people who would do the asking need materials, and language, and something to send, and a reason, and somebody to answer their questions on a Tuesday night — and no central communications team on earth can produce fifty localized versions of anything. Chapter four made this case generally. It applies here with a sharper edge, because here the bottleneck is measurable in money.

Two smaller things follow, and I want to name them because they are the parts I got wrong myself.

The first is that if we are going to ask people for money, we should be honest about what the money and the asking are for. This petition on its own will not change anything. Your participation will. I spent years selling the opposite of that sentence, as I said yesterday, and I think it did more damage than the fundraising ever did.

The second is that people are not tired of being asked. They are tired of being processed. Nothing looks so good as a home-cooked meal, made by somebody who wanted you at the table, after a long trip of eating nothing but junk food. Chapter two used the fast-food half of that comparison. This is the other half, and it is the more useful one, because it tells you what people are actually hungry for. Not less contact. Contact from somebody who meant it.

That is what a fundraising coach embedded in a group's own workspace is for — not writing the appeal for the volunteer, but helping the volunteer write the appeal only they could send, to the twelve people only they can reach. Multiply that by the number of people in your organization who would say yes if asked properly.

Which is, I notice, the same sentence as the problem. We have never had trouble finding people who would say yes if asked properly. We built an entire industry on the assumption that nobody would do the asking.