Grow vs stay small for startup nonprofits

Small startup nonprofits, such as those from Charity Entrepreneurship, face the tradeoff between staying smaller, and thus by default more cost-effective per unit of money, and grow, and accept diminishing returns but larger impact. The second seems worth doing because it unlocks more impact as a whole.

The first order analysis: Producing impact at scale is worth the reduction in cost-effectiveness.

In the financial markets, getting a larger return for more capital is harder. Berkshire Hathaway is holding $382 billion in cash, because they can’t find good enough returns, say 20%/year, at that size. But as a private individual, one can get those returns via private investments, education, exploring career opportunities, changing jobs, or following niche market trends. But as one’s capital size gets high enough, one can’t deploy that capital without moving markets too much, and one starts to need ritualized financial instruments, like stocks rather than handshake agreements.

Similarly, even though a scrappy nonprofit with a few talented people is more cost-effective, the number of those opportunities is limited. For instance, there might be only so many areas of operation, or only so many talented people willing to do so. Or, even if Charity Entrepreneurship is able to find a growing number of people to lead those small organizations, the rate at which CE itself can grow is limited.

So my best guess is that, in general, the ecosystem of nonprofit startups face the choice between something like:

• There are \(N\) opportunities which offer \(X\) units of impact per $1K, but can take a maximum of \(A\) dollars. Total attainable impact is \(N \cdot X \cdot A \)
• There are \(M\), where \(M>>N\) opportunities which offer \(Y\) units of impact per $1K, and which can take a maximum of \(B\) dollars, \(B >> A\). Total attainable impact is \(M \cdot Y \cdot B\)$. If diminishing returns are reasonable, e.g., \(Y = X/4\), then total attainable impact is \( M \cdot X \cdot B / 4\), which, because M and B are much more than 4x the size of N and A respectively.

A community coordinating on which of the two scenarios to choose would definitely settle on the second one. Yes, each individual intervention is less cost-effective in the second scenario. But by accepting lower cost-effectiveness in return for scale, the second choice unlocks more impact as a whole. There is also not a harsh binary, one can choose to grow interventions first in order of impact, but the overall decision still seems illustrative.

The second order analysis: Cost of acquiring funding in terms of talent

Acquiring further funding takes bandwidth in terms of…

• Sending funding applications
• Understanding funder desires
• Making work legible to funders
• Making funders feel important
• etc.

As a result, fundraising as a whole can be extremely time consuming, with some sources¹ recommending that founders should spend around 50% of their time on it. This changes the analysis above, because unlike in an initial analysis, now funding isn’t something that is automatically unlocked by impact, but something that needs to be actively chased. This is makes it harder to reason about, because:

• An ecosystem like Charity Entrepreneurship’s one may have a limited amount of funding, which might make growth for one of its startups trade-off against growth in another, or against exploring a totally new opportunity.
• In a small startup with limited talent, dedicating large proportions of it to chasing funding can be extremely costly, particularly if one has a high discount rate with respect to the area one is exploring intervening in, or if the fit towards fundraising at higher amounts is much much worse than the fit for object level work.

That changes the conclusion to something like: growth at the expense of diminishing returns in impact is probably a good move is possible. But accounting for product-funder fit considerations, chasing that funding adds another option: scaling at the expense of some diminishing returns would be a good move if there weren’t costs related to that fundraising, but incorporating them it makes it not worth it.

Some ideal solutions to this problem

• Find a funder that is willing to back you as you expand, and take the hit in cost-effectiveness as you scale, because your intervention is still promising enough
• Find an intervention that has increasing returns to scale, so that your funders don’t have to face that tradeoff
• Find an intervention that has self-reinforcing positive feedback loops; something like:
  • Success makes you more legible to funders, which makes you able to raise more funding, which makes you able to deliver greater successes
  • Your intervention returns to you more cash than you spend, because you charge clients. This is posible to do as a nonprofit, but also in this case you can consider spinning up a for-profit. A for-profit also allows you to sell equity with which to motivate investors and employees.

How does ARMOR fall into this spectrum

Because ARMOR’s mission, reducing anti-microbial resistance, is so broad, and it’s current budget so small compared to the mission (“From our launch in September 2023 to March 2025, ARMoR has received $226,477”), my very strong sense is that if it is able to find promising interventions at the $200K level, it should also be able to find interventions at the $2M and $20M level that don’t fall off much in terms of promisingness.

I have no idea about their interaction with the funding ecosystem, and how averse leadership is to fundraising vs doign object-level work. Potentially atimicrobial resistance could also be viable as a for-profit intervention, which might allow fundraising a few million on the strength of an idea or early prototype. Despite that uncertainty, my strong guess is that ARMOR should seek to scale.

Note that at time of this writting, I haven’t yet reviewed their actual outputs, and so this analysis is perhaps too general; when applied to a specific charity one will also have to incorporate factors specific to that charity. More specifically, I’m not saying that ARMOR in particular will have diminishing returns between the $200K and $2M/year level, just that it might be worth expanding even if it is.

Thanks to Mo Putera and ARMOR for commissioning this piece