Thank you to @Jacintha Baas, @Judith Rensing, and the @CE team for your help in editing and improving this post.
Introductory Context
Hi, I’m Trish. This is my first post on the EA forum. I spent the past decade building VC-funded, high-growth startups (from pre-revenue to $10M+, from founding team to 250+ employees; pre-funding through Series B), and supporting 100s of early-stage founders through the communities at On Deck and Unreasonable Group.
Last year, I joined the team at AIM to help build the Founding to Give program, which helps people start high-impact for-profit tech startups.
AIM’s Founding to Give program aims to launch for-profit startups that create impact through two ways:
- Build a startup, exit, and donate to effective charities (similar to earning to give)
- Direct positive impact created through the startup
AIM’s research and Founder’s Pledge data show high expected value (EV) for the first route, but this specific post is about the lesser-explored route of direct impact through for-profit companies.
Note: Applications are open for the next cohort of Founding to Give – deadline is August 1st. You can learn more here, and apply here.
Summary
- For-profit companies can create significant positive direct impact (even though they’re not necessarily optimizing for it), and they require $0 in philanthropic funding to do so.
- We think that for-profit direct impact is under-explored and could unlock a significant, counterfactual amount of impactful work that could be done.
- This post does not address the additional potential positive impact that can be created indirectly through for-profit startups via earning to give or Founder’s Pledge-style donations, which we estimate to also be significant – more on that in a future post.
Overview: The Bull & Bear Case
- The Bull Case – Significant, “Free” Impact: There are examples of for-profit companies that create significant, direct, measurable impact, and they require $0 of philanthropic giving.
- The Bear Case – Irrelevant Impact: There are strong tensions between the incentives of a for-profit and direct, measurable impact that make them incompatible. Solving many problems in global health and development, animal welfare, and beyond is impossible to do via for-profit companies.
- My take: Both of these things can be true, depending on the context. Meaning that some problems can only be solved effectively by nonprofits, and, in some cases, for-profits can create significant direct impact (and this route is currently under-explored).
Reasons often given for why for-profits can’t have significant impact (and their counterexamples)
#1 “You can’t serve two masters”
The idea is that for-profit companies have to optimize for profits, so it’s impossible that they could optimize for impact too.
- All organizations deal with tradeoffs and tension. The idea that it’s impossible to optimize simultaneously across two priorities is overly simplified. Here are two examples:
- Cost-effective charities optimize for keeping costs low while simultaneously increasing their scale (often directly in tension).
- For-profit companies optimize for profit, which means simultaneously optimizing for increasing revenue while decreasing costs (often directly in tension).
#2 “Show me a for-profit that is solving for rural malaria prevention”
The idea that for-profit companies don’t touch the most crucial problems.
- There exist market failures. We do not think that for-profit companies can solve all (or even most) crucial problems. But similarly, the global pool of philanthropic funding is extremely limited and unable to solve all crucial problems on its own.
- We do think that for-profits can contribute to solving some crucial problems.
- And in those cases, they don’t require philanthropic capital to do so.
- Two examples of how for-profits contribute to "crucial problems” in global health and development:
- Vestergaard is a for-profit company that develops and manufactures malaria nets that are used by AMF.
- Zipline is a for-profit company that our research estimates might save ~10K lives/year at scale by increasing accessibility to life-saving medical treatment via drones.
#3 “For profits can’t solve problems for those suffering the most”
The idea that for-profit companies require paying customers; those suffering the most can’t pay, and therefore the two are at fundamentally odds.
- There is often a tension here, which is why we think for-profits are not relevant to solve some types of problems.
- However, we see this idea as an oversimplification of for-profit business models and stakeholders.
- Here are some examples of more nuanced business models and stakeholder relationships that circumvent this tension:
- Marketplaces: platforms that connect two parties, one party pays, and the marketplace makes money by taking a fee, often a set fee or % of the amount paid by the paying party.
- Airbnb is a common example, Fiverr is an example that we estimate to have created significant direct impact through impact doublings for people living in LMICs with limited economic opportunity.
- Supply chains: stakeholders throughout the supply chain use infrastructure and products that can directly impact beneficiaries
- Wasoko and In Ovo are examples of for-profit companies that sell products to producers in the supply chain that have a direct and positive effect on end users: by lowering the cost of goods for communities in Africa, and reducing chick culling.
- Another way of looking at this that circumvents the tension to some extent: we aren’t only interested in helping those who have 0 disposable income.
#4 “But what about the perverse incentives that will inevitably emerge”
The idea that even if the founder is impact-driven and rational, investor pressure or market dynamics eventually win.
- There are examples where there are low perverse incentives, where impact and profit are aligned enough that significant perverse incentives are unlikely to emerge. Most examples above fit here, but additional examples:
- Entrova (a company from Founding to Give’s first cohort) or Andela: income doublings are aligned with company profit
- We think that there are strategies that founders can take to reduce perverse incentives, for example:
- Finding aligned investors
- Keeping majority control of their board
- Deeply modeling and validating their idea, problem space, and market dynamics at the beginning of starting their startup
- Working on ideas in industries that have lower risk of potential harm, e.g. not building in spaces like military tech, tobacco, etc.
- There are cases where a founder cannot find a way to continue where the company is commercially viable and creating significant direct impact, and in these cases we think possible pivots could be:
- Shutting down the company and pivoting to a new career path
- Pivoting the company to a less directly impactful route but more commercially successful route and achieving impact through a donation pledge (e.g. employing an “earning to give” model)
#5 “It would have happened anyway”
The idea that if there’s a commercial opportunity, someone else will build it anyway. Or the idea that if a company already exists in the space, the counterfactual impact of another is low or zero.
- This one is a bit more messy, so I’m going to give a few counter arguments.
- Option #1: It is true in some cases that it would have happened anyway, but to assume this as the rule requires the assumption that markets are perfect which is very unlikely. So, we should look at this as something to assess, but not an immediate blocker. I’ll mention a couple of cases in which it is less likely that something ‘would have happened anyway’:
- A founder has a unique combination of experience/insights that allows to spot a unique insight that many others would not have (ie: strong founder-market or founder-problem fit)
- A founder has a unique network that would allow them to sign first customers where other founders would struggle to gain confidence from early customers
- Specific markets, like emerging markets, which are generally less saturated due to higher risk, lower capital, etc
- Option #2: If you do believe that markets eventually are always perfect (seems very unlikely, but for sake of exploration), then yes, eventually someone else will build it, but the timing matters: if we build it X years sooner, it will create Y more impact / impact sooner. Said another way, there can still be a counterfactual impact if you move up a timeline significantly.
- Option #3: In many types of industries/markets, it is unlikely that one company will “win the market” as a single player. Especially in new, innovative spaces, more companies can expand the market beyond what a single player could have reached (“market making”).
- This is not necessarily to say that the impact is 100% counterfactual, but to say that there can still be significant impact from the second, third, or fourth player.
- For example, GFI’s consumer surveys show that 40–60% of plant-based buyers say they would have bought meat/dairy otherwise, so some are substituting meat, and others are deciding between two veggie burgers.
#6 “DuoLingo is designed to keep you using the app, not to deliver real learning outcomes”
The idea that for-profit companies are not impactful if they are oriented towards retention or other business-related metrics rather than other outcomes.
- My experience building startups makes me assume that for-profit consumer apps are optimizing for engagement and retention, but this does not prevent them from delivering significant, real outcomes.
- One example is Headspace, a mental health and meditation app. We can assume that they optimize for commercial metrics like retention, but studies show that they also improve outcomes like reducing depression.
- Another example is Duolingo. If you believe that Duolingo is implementing a suboptimal solution (ie: isn't fully optimized for speed of learning), and a non-profit could implement a more effective program on a per-person basis, the following can still be true:
- Duolingo is creating significant outcomes in improved language abilities
- Duolingo requires $0 of philanthropic funding.
- Duolingo may create more absolute impact than the nonprofit, if it is able to reach a significantly larger scale.
Some ideas for where for-profits might be able to create significant direct impact
At AIM, the team has done some research and BOTEC case studies to estimate the expected value of impact created by some for-profit companies. These BOTECs require some very subjective inputs (e.g. counterfactuality discount) and estimates around key data (e.g. number of people served).
Based on this research, here’s a short starting list of examples and case studies I’d like to look into further:
- Income doublings
- QALYs/DALYs
- Animal Welfare
- Meat/dairy alternatives
- Agritech practices that improve margins and animal welfare (In Ovo)
- Other examples like: veterinarian services, medicine and feed suppliers, water quality
- Improved educational outcomes
Some ideas for where for-profits might be able to uniquely contribute to the impact ecosystem
As an additional benefit beyond “direct impact” as shown in examples above, here are some ideas on unique advantages that the incentives/model of for-profits could be well-suited for.
- Innovation – There are funders in the for-profit world that have high risk tolerance (e.g. early stage VC firms or certain types of debt financing), and this allows for taking more bets or bets with higher uncertainty on R&D and innovation. This innovation can then become the starting point for new solutions to old problems. Some examples:
- Medicine: LLINs for malaria, deworming meds, chlorination tablets and dispensers, fortified foods, micronutrient powders, vitamin A capsules – all would not have been created or manufactured without for-profit health and pharma companies
- Other examples could include: CO2 capture methods, agritech crop yield innovation, alt protein advancements, etc.
- Infrastructure – nonprofits often lack the scale, funding, or technical ability to build infrastructure that can directly improve accessibility, or be used via nonprofits to deliver an intervention.
- Zipline (reached previously impossible/difficult to access communities quickly)
- M-Pesa (MIT study estimate it took 2% of Kenya’s population out of extreme poverty, used by GiveDirectly)
- LMIC governmental partnerships – some LMIC governments (e.g. China, India) seem more excited / less suspicious of for-profit businesses, so it’s possible that they could have an advantage in areas where government relationships are important to success and where nonprofits have struggled to build relationships with the government.
Next Steps
- If you have interest or experience in this space or are already running a for-profit and are thinking about impact, please reach out.
- Our next steps include further exploration in understanding the types of contexts, problem spaces, and business models in which for-profit companies are likely to generate significant direct impact.
- If you have an example or idea for a for-profit company that could have significant direct impact, let us know.
One question that I’m very interested to answer is: What would be the “GiveWell recommended” bar for for-profit companies?
If you’re interested in starting a high impact for-profit startup – Founding to Give is for you. Application deadline is August 1st. You can learn more here, and apply here.
Great post, Trish – appreciate the thoughtful exploration of direct-impact for-profits.
Some thoughts:
EA lens on operational impact: While there's significant investment in "doing good through operations" (B-Corps, social enterprises, etc.), most aren't applying EA cost-effectiveness bars. A company might proudly reduce carbon emissions at $500/ton when that same money could remove 500x more CO2 through other interventions. There's room for more rigorous impact evaluation within operational models and I am also excited to see EA interest in this!
Subsidizing positive externalities: Your framework assumes profitable ventures, but what about cases where massive positive externalities justify subsidies? If a venture generates $10M in social value but loses $1M annually, that's still a 10:1 social return. Traditional impact investing struggles here (expects returns), and traditional charity struggles too (not set up to run businesses). Could be worth exploring hybrid models where philanthropic capital subsidizes high-externality ventures that can't quite reach profitability (of course would have to consider the extremely high counterfactual potential impact of the dollar to effective charities).
Profit for Good as complementary lever: Beyond operational impact, there's the neglected question of where profits flow. Profit for Good (PFG) companies compete normally but lock 90%+ of profits to charity through ownership structures. This creates interesting dynamics: (1) no operational tradeoffs needed since the business still maximizes profit, (2) "charity choice" effects where consumers preferentially buy from companies funding causes over enriching shareholders, creating competitive advantages, (3) philanthropic multiplier where $1 invested can generate many times that for charity over time. Humanitix (event ticketing) has donated $16.5M AUD while taking market share from Ticketmaster – not through guilt but through better user experience plus the fact that event organizers prefer funding education over enriching monopolists. The model needs mission-aligned capital (VCs want returns), but for the right contexts it should multiply philanthropic funding. Happy to share research on this if helpful. (here's my blog post on this competitive advantage thesis).
In theory, many hospitals and universities in the US should fit into the "positive externalities" model -- they significantly rely on both program service revenue and subsidies from private donors. My understanding is that nonprofit hospitals and universities are often poorly governed, so I'm curious whether this is related to the serving-two-masters problem as opposed to sector-specific pathologies.
ETA: Someone who knows more than I do about microfinance may be able to comment on how experience with microfinance orgs updates toward or against the subsidizing positive externalities in a unprofitable business model being viable more generally.
Re subsidizing positive externalities: I'm not necessarily suggesting that typical instances of external subsidization of businesses with positive externalities are particularly impactful, just that there may be possible cases, especially if we are looking for opportunities from an EA lens.