Impact = Funding > Impact = Profit

There’s a meme in the public goods funding ecosystem that looks like this:
It’s an argument about the conditions under which we should reward someone for the work they’ve done to contribute back to a community through retroactive funding for public goods (RetroPGF).
For the uninitiated, RetroPGF is about retroactively funding the creation of public goods. Retroactive public goods funding is a funding model where individuals and organizations contribute to a public good after it has been created or completed. This approach allows for the funding of public goods that have already been proven to be successful or effective, rather than relying on uncertain projections of future success. RetroPGF is proposed to be useful for a variety of purposes, including supporting scientific research, funding public art projects, or financing social programs.
The key question that lies at the heart of getting RetroPGF to work is always, “how much should we retroactively fund a project that had huge impact?” Let’s say that a research program found a cure for a particular type of cancer. This is clearly meritorious of some large amount of retroactive funding as thanks for the outcome they achieved. But by what criteria should we reward them?
As seen in the image above, Impact = Profit offers one such rule for awarding retroactive funds. The idea is easy to explain: the amount of funding you should get back ought to make your total profit proportional to the impact you’ve had.
What I’m going to argue is that this is not a suitable criteria for retroactive funding. The problem with Impact = Profit is that profit can be gamed. There are ways to artificially reduce reported profits while keeping the takehome amount stable, this allows nefarious actors to cheat retrofunders out of their money. However, it’s not all bad news, there’s at least one option for salvaging the meme.

What is profit?

I’ve noticed in conversations about this topic that sometimes there isn’t a shared ontology (terms and their meanings), which can make it difficult to discuss the issues. Let’s briefly review some terms.
Revenue is what you earn from selling something. Let’s say one day I sell 100 slices of pizza for $4 each, that means my revenue was $400.
Profit is defined as the total revenue of your business minus expenses.
There are actually several definitions of profit, including gross profit, operating profit, and net profit.
Gross profit is what I make per unit sold. So, if the pizza takes $3 of ingredients to make, then my gross profit is $1 per pizza sold. On 100 pizza slices my gross profit is $100.
Operating profit is what fraction of that I keep after paying for things I know I’m going to have to pay for no matter how much pizza I sell. If I know I need to pay $25 per day in rent, then my operating profit for that day is only $75. Operating profit can also include salaries, remember this as it will become important.
Finally, I’m going to have to pay taxes on that operating profit. If I pay 20% in taxes, then I lose another $15. That leaves me with only $60 of the revenue I’ve earneed as the amount I get to take home as the pizza shop owner. This $60 is my net profit.
Revenue: $400
Gross profit: $100
Operating profit: $75
Net profit: $60
Now that we have the terms defined, we can discuss its impact on retrofunding.

Magic Public Goods Pizza

To think about this from a public goods perspective we need to somehow make our pizza a public good. To achieve this, let’s imagine that the pizza I sell also keeps people healthy by reducing the rate of transmission of viruses and bacteria (don’t ask me how, this is magic pizza). This would make my pizza partially a public good because if everyone eats it instead of normal pizza then everyone will be healthier (not only directly, which would make it a private benefit, but indirectly, because when those around you eat it they are less likely to transmit it to you).
But there’s a problem: my competitor is selling normal pizza for $3 and everyone is preferring to buy from her just because it’s cheaper. This is (a contrived example of) where public goods funding can come in. If I know that I can get funding for selling magic pizza instead of regular pizza then I can lower my prices to be competitive. Maybe I know that the retrofunding authority is willing to pay $1.5 per magic pizza I sell, this enables me to lower my prices to be even cheaper than the normal pizza, so now people can buy it and we all get the collective benefits of being healthier.
That’s the role that retrofunding can play in funding a public good. So, what’s the problem?
In this story, the payment that the retrofunding authority was willing to send was based purely on the impact the pizza seller had (in this case, measured by number of magic pizzas sold). But Impact = Profit says that instead we should try to figure out the impact that a company had and then pay them extra until their profits reach the right level.
To achieve this, instead of paying per pizza, the retrofunding organization would say something like, “this was great work, you’ve improved community health worth $150, and so we’d like to give you an extra payment to make your net profit bigger.”
If I’ve improved community health worth $150 then they should give me an additional $90 to boost my profits from $60 up to $150.

The Problem

But now I want you to notice something. What would happen if my costs per pizza were to increase from $3 to $3.5? Now this is what my income looks like:
Revenue: $400
Gross profit: $50
Operating profit: $25
Net profit: $20
Now my costs have gone up, my net profit has gone way down, and so this time the retrofunding authority has to pay me $130 to make sure that my profit = impact.
This isn’t such a big deal? Right? I mean, sure the retrofunding authority has to pay more than before but it’s for public health so it’s all worth it, right?
Unfortunately, there’s something shady going on that I haven’t told you about: I also own the company that supplies me with the ingredients for the pizza. So, it’s actually me that’s benefitting from the higher costs of ingredients, and I know I can charge this additional amount because the retrofunding from Impact = Profit will make up for it.
Maybe you think this isn’t too big of a problem, you could just check who owns the ingredients company, right? Well, not so fast. There are actually dozens of ways I can make money by decreasing my apparent profit margin, here are a couple ideas:
as an administrator of the company, I can pay myself a large salary, increasing our costs and decreasing our profit
I can hire my friends or family as employees and overpay them
I can buy the building I work out of and charge myself a high rent
I can make a backroom deal with an ingredients supplier (that’s not owned by me) to agree to pay them more than the market rate in exchange for them sending me kickbacks.
Tricky, huh?
It turns out this is a well known problem in contracting, it’s called , and it often results in over charging because as costs rise profits rise, allowing the team to pay themselves more while increasing their takehome!
If you’d like to play with some scenarios, here’s a visualization of the problem, showing that when a company has impact and revenue, and they sneakily pay their revenue to themselves, they can significantly increase their takehome while causing retrofunders to have to pay much more.
impact was worth
company’s revenue
Takehome pay as companies fake their costs
You can see that this gives companies a strong incentive to fake their costs. This wouldn’t be a problem if it was obvious when they were cheating, but it seems that it’s really hard and expensive to detect cheating (if you don’t believe me, ask the IRS why they’re having to hire so many additional analysts).
Furthermore, even if the business is not cheating the system, Impact = Profit also strips them of any reason for them to be efficient about their costs. If their suppliers want to charge them double it’s no worries, the company will just shrug their shoulders because the bill is going to be paid by the community anyway.
This is not the outcome we want, we’d much prefer that companies are incentivised to be good stewards of the funding they receive by deploying it efficiently.

An Alternative.

But there’s hope! There’s at least one promising alternative to Impact = Profit: Impact = Funding.
With Impact = Funding this incentive is fixed. The retrofunding authority would determine how much they’re willing to pay for a unit of impact, e.g. for lives saved or CO₂ sequestered or a project delivered, and then companies can find creative ways to achieve it. Not only does this still give them a reason to maximize impact, it keeps them with an incentive to minimize costs and to run efficiently.
This is exactly what we did in the story of the pizza shop owner: by just paying a certain amount for the sale of each pizza, we reward the owner for something we can measure that we think produces valuable outcomes.
Now, how does a retrofunding authority figure out what to measure and what to fund? Ha, if you know the answer to that please let me know.

Ways I Can Be Wrong

No argument is complete without a section that offers surface area for invalidation. Here are several ways I could change my mind:

Profit (and project sustainability) should be taken into account when evaluating funding

Perhaps the most compelling argument on behalf of Impact = Profit is that it captures an intuitive notion that projects that are struggling should get better access to more funding than projects that are already sustainably funded.
For example, let’s imagine there are two projects:
Uniswap, which has a functioning revenue model and is providing obvious impact
A security researcher whose impact is also evident but it’s less obvious how to monetize it
The prima facie elegance of the Profit = Impact framing is that it takes into account the fact that the security researcher doesn’t have an obvious way to monetize their work. This implies that under Impact = Profit, the researcher’s lack of earnings would show up in the evaluation and would result in them getting more rewards, whereas Uniswap’s earnings would ensure they get smaller rewards.
This makes intuitive sense to many people, there’s clearly a moment in a project’s lifecycle where a grant (even a small one) can decide the fate of the project going forward.
However, as we’ve already discussed, this is complicated to implement because of the difficulty of inferring profit accurately. I.e. if we gave bigger rewards to bigger profit deficits then Uniswap would have reason to reduce their profits, taking them from the community instead, thereby nullifying the difference between them and the researcher.
The role of retroactive funding is to reward a project for their existing success, not their projected success — the latter is exactly the job of prospective funding and investment. So, at least for retroactive funding, you don’t want Impact = Profit.

It’s Actually Easy to Infer Profit Accurately

The easiest, knock down way for me to be wrong is if it is somehow easy to infer the true profit of a project. This I highly doubt, however, if you can show that I’m wrong about this then the entire preceding argument evaporates.
Some things you might show:
You can figure out that employees are being paid “fairly”
You can figure out who owns the various different suppliers of goods
You can detect collusion between players (e.g. a backroom profit sharing deal)
And you can do all this without imposing unnecessary burden on legitimate business operators
My sense is that none of these are easy to do, but if you can show that they are (I think you need all of them, but even one of them would surprise me) then I would substantially change my mind.

Impact = Profit has Great Desirable Second Order Effects

Instead of showing that the cost of Impact = Profit is low (as in “It’s Actually Easy to Infer Profit Accurately”), show that the upside is large. If Impact = Profit substantially increases certain desirable outcomes over and above Impact = Funding, then I’d be willing to revisit this and think more deeply about the tradeoffs. Some things I value that you could show Impact = Profit produces more of:
More transparency
Long term sustainability for projects
Survivability of good projects through key funding moments
More projects go open source
Equivalent outcome at a lower cost to funding authorities

Impact = Funding makes it much harder to figure out a fair impact metric

Under Impact = Funding we have to find a way to fund projects for their impact independent of their profit.
We want to do this because we’d like to avoid overrewarding projects for doing things that are considered good, but are already profitable. This makes it much harder to figure out how much of a reward to offer.


Take for example the logging company that plans on planting a whole bunch of trees. Sure, those trees will sequester a bunch of CO₂, which might be an impact we’re willing to pay for, but the company was already going to do that as part of their business operations, so should they really get the additional reward? This question is confounded by the fact that it’s unlikely the company will pay for the CO₂ they emit when logging the trees, so they’re getting the upside but not the downside of their behaviors.
This is definitely a problem, but it’s a problem that’s general to retrofunding, not to Impact = Funding. Even under the Impact = Profit regime you’d have to figure out if you want to count this logging company’s sequestration as a type of beneficial impact or not, adding in the necessity of knowing their profit doesn’t do anything to reduce the difficulty of that burden.
If you can show that it does somehow mitigate the double counting problem then that would be quite interesting.

Optimism only metaphorically means Impact = Profit

One possibility that has been pointed out to me is that perhaps Optimism has used the Impact = Profit meme to essentially mean Impact = Funding all along.
I don’t think that’s the case. If we look at the image:
It’s clear that somehow in order for Optimism to fill up the yellow bar they had to know how much profit the project already achieved — leading to all those problems with inferring profit that we’ve already covered.
However, if it really is as simple as a misunderstanding of their meme then it should be a simple change to make since it was more of a communication than a mechanism problem.
Want to print your doc?
This is not the way.
Try clicking the ⋯ next to your doc name or using a keyboard shortcut (
) instead.