Stop asking “how good is this” to decide between donation opportunities I recommend

Zach Stein-Perlman27 Mar 2026 15:00 UTC
32 points
6 comments1 min readLW link

Prospective donors often ask how cost-effective my team’s various recommendations are — they want to donate to the most cost-effective opportunity. But the answer is often roughly:

If we do our job right, then all donations we recommend to you are equally valuable on the margin, and that value is our bar. If an opportunity is substantially above our bar, we will make sure to fill it with or without you; you’re not counterfactual. (Alternative framing: we will just recommend more funding for that opportunity until its marginal value falls to our bar.) Your donations funge[1] with donations to our other recommendations. (That’s fine; our bar is high.[2])

This is a simplification:

  • Sometimes there are legal limits like “max $7K per donor” and this means we can’t direct as much money to the opportunity as we’d like, so marginal donations are above our bar. And sometimes the identity of the donor provides various costs or benefits, so certain donors have an advantage. For this reason, many donors should set aside some money for capped opportunities or opportunities where they’re a particularly good donor.

  • Sometimes a great opportunity is sufficiently urgent/​sensitive/​weird that we’re uncertain we can fill it, or at least without costly delay or some nonmonetary cost. Such opportunities are above our bar even on the margin.

  • Sometimes tax considerations mean that certain donors have an advantage for certain opportunities.

But mostly when people ask “how good is this opportunity on the margin” to decide between opportunities we recommend, that’s the wrong question.

We sometimes talk about the effectiveness of the first dollar or the average dollar to convey our excitement about an opportunity. But this shouldn’t affect donors’ decisions — e.g. you shouldn’t wait for opportunities we’re especially excited about; it all funges. (Total value matters to grantmakers; grantmakers generate value by recommending donation opportunities with total value much greater than their cost. But it’s not relevant for marginal donors.)

Alternative framings are in this footnote.[3]

Thanks to Eric Neyman for suggestions.

This post is part of my sequence inspired by my prioritization research and donation advising work.

  1. ^

    “Funge with” (or “funge against”) means “substitute with” or “displace or be displaced by.”

  2. ^

    Our current bar is roughly $700M per 1% future-improvement for US direct political donations and $3B per 1% future-improvement for 501(c)(3) opportunities.

    But for political donations, your first ~$50K/​year is substantially more valuable, since it can go to opportunities where individual donations are capped and the constraint is the number of donors. For example, we might recommend: donate $7K to Alex Bores, then save $5K for unknown amazing future stuff, then donate $7K to Scott Wiener, then save a little more, then donate to the next best thing…

  3. ^

    This post says: we have a political donation bar (and other bars for other kinds of money), and constraints like max $7K per donor or urgent can result in opportunities above the bar. Another framing is:

    There are many different classes or sets of constraints—such as political, max $7K per donor* or political (small donors not advantaged) or nonprofit, urgent or nonprofit (not urgent)—and (if we do our job right) all opportunities within each class are equally good on the margin, but different classes have different values. Donors should donate to the best class available to them.

    *Actually for handling small-donor opportunities I guess we need separate classes for each step in a donor’s budget — we need a “a donor’s first $7K” class and a “a donor’s dollars number 7,001-12,000″ class (assuming per the previous footnote that the second-best thing to do is save $5K, and value is constant across those 5K dollars) and then “a donor’s dollars number 12,001-19,000” and so forth. Obviously it’s deranged to deal with budget size using discrete classes like this. Perhaps we could use classes like political, max $7K per donor but add a subscript to indicate the minimum budget for which a donor should donate to the opportunity...

    Another framing is:

    Think about what donors you’re displacing; be excited to displace donors in a great class. One thing that comes up frequently is that large political donors should be excited to displace small political donors; if we’re like “this is great so we could get small donors to fill it if necessary but if larger donors filled it then the small donors would be free to donate to stuff where they have an advantage,” large donors should be excited to donate.

Zach Stein-Perlman27 Mar 2026 15:00 UTC
32 points
6 comments1 min readLW link