Often folks hit us up because they are thinking of starting an incubator and want advice.

Typically their motivation is either that (a) they have a list of specific things they want built that no one is building,  or (b) they think an ecosystem needs more new projects generally to absorb more talent and deploy more funding effectively.

Here are six questions we often ask prospective teams, to help them figure out what to do. If you're incubator-curious, we recommend asking them of yourself. This is a companion to "What even is an incubator?" which lays out our framework for deciding what kind of incubator to build. This piece is the messier, more opinionated list of everything else we tend to say.

(Note: This is AIM’s ‘house view’ from incubating 60+ charities across Global Health and Development, Animal Welfare, and “Meta” (including many bangers, we think!). We've only recently launched a MEL function to more robustly test our hypotheses about how to incubate field-leading charities, so expect us to learn that some of what follows is wrong)

((Also: We are generally excited about more, strong, organisations doing incubation with a variety of theses and specialisations and are happy to share resources and learnings to new projects))

1. Are you sure the answer is starting new orgs?

Starting new orgs is a powerful tool for getting things built, but it’s not the only one. Is it the right tool for the specific gaps you’re seeing or problems you’re trying to solve? As an incubator, AIM needs to constantly ask ourselves this, because when you’re holding an incubation-hammer, everything starts to look like a charity-nail.

Consider a wider toolkit:

• Incubating new orgs: New orgs can move fast because they don’t depend on big, slow, risk-averse incumbents. They can attract entrepreneurially-minded talent, who are eager to build things as founders that they’d be unwilling to build as part of an existing organization. You can instill important values from the outset, instead of pursuing an often-futile path of trying to change existing institutions. But they have to build credibility and systems from zero (costly when credibility or fast scale is critical for success). They require a bunch of foundational work (e.g. figuring out your legal entity, gaining the ability to legally hire staff, banking/accounting, branding/website). And they require a type of talent that can build from 0 to 1, which is exceedingly rare, as a large part of the talent pool wants or needs more stability than founding can offer, and/or doesn’t want the unique kind of stress that comes with being the person the buck stops with.
• Incubating independent projects: If the thing is too small to justify org setup costs, or if going well means it no longer exists in a year, just fund a person, possibly under a fiscal sponsor, and keep it lean. Also ask whether it deserves your scarcest input: excellent founder talent. Smaller projects may be better used as an interim for future founders, or for strong people not suited to something as demanding as founding an organization.
• Removing blockers for an existing org to take it on: Existing organizations have already built the basic organizational infrastructure, they come with existing brand, network, and track record that may help access more/better resources, and have invaluable learnings and mentorship to share. The risk is that the new project is beholden to the existing org’s decision-speed, risk-tolerance and competing priorities (our COO experienced this first hand trying to build a grocery delivery business at Uber Eats – competing with scrappy startups while stapled to a behemoth is painful).
  • The how: Figure out why existing actors aren’t doing it, and plug the gap. This could mean persuading them it’s a priority, helping them secure funding, helping them hire the team, or helping connect them to the right expert advice. Often leadership bandwidth is the bottleneck, in which case you competently doing things for them (like hiring) can make the difference.
• Just launch the org yourself. If you’re entrepreneurial, and you’ve spotted a specific gap or problem in the ecosystem, have you considered just taking the idea that’s most promising and best suited for you, and just launching this? Oftentimes people who are well-cued-in enough to spot a problem are well-suited to fix it. We are generally bottlenecked by people who actually want to ‘do the thing’, so why not just go at it yourself? We cover this in a section below, but we have seen that having founding experience yourself makes you much better at helping other people found.

2. What do you mean by "incubator"?

People usually have something very specific in mind, e.g. "AIM's program, but for GCRs", or "take five org ideas we want founded, hire a solo founder for each, and mentor them". But new orgs are the outcome. The activities that get you there can vary enormously, and the picture in your head is one option among many. Rather than defaulting to it, ask: what blocks these orgs from existing?  

For AIM's classic program, the roughly ranked list is: (1) a quality charity idea, (2) an excellent co-founder, (3) permission to do it, (4) training and mentorship, (5) funding, (6) support with operational hurdles. Yours will differ with your target founders and context. Experienced founders may not need permission; funding may be a non-issue while choosing the right project is the bottleneck.

Once you have your list, three design choices follow: How deep will the support you provide be? How much of it is the same for everyone? And are you pairing generalist founders with generalist ideas, or specialists with specialists? We've written those up, with the mistakes we made along the way, in "What even is an incubator?". The answers can land you anywhere from an eight-week program for talented generalists to removing the single blocker in the way of an experienced founder.

To reiterate: Figure out what blockers you’re trying to remove from first principles and then design the incubator that removes them.

3. What are you going to do about talent?

We hypothesise that founder quality is the most predictive factor of charity success. Attracting and vetting them is extremely hard, so it should probably be your core focus. AIM has a full recruitment team dedicated to finding and extensively assessing thousands of applications per year. For many “incubator-curious” folks that come to us, it seems like the talent strategy is an afterthought.

So: Who is your target audience? How are you going to find them? Where do they hang out in-person and online?  How are you going to reach them with your message? Given that the best founders often have many good career options, how many others are vying for their attention and what can you offer that they cannot? How certain are you that your program will allow them to have more impact than they would have had in other roles? What do they currently think and feel, what do you need them to think and feel to found with you, and how are you going to close the gap between the two? And how confident are you in your ability to accurately identify them in your vetting systems? Surface traits are often misleading!

Understandably, a lot of organizations who want to incubate, follow the path of least resistance: (a) Target founder talent near the ‘bottom of the funnel’, i.e. people who are already interested in founding with a strong track-record of doing so before, (b) Copy the outreach tactics that have worked for other incubators, reaching out through the same channels to the same audience.

We worry this is negative sum: the new incubators end up redistributing a finite pool of founders across a wider range of incubators, and this is worse than zero-sum because we believe excellent founders are multiplicative when paired and working together in tightly knit-communities.

This means that if you're starting or funding an incubator, we believe “growing the pie” of impact-oriented founder talent should be a requirement.

4. What are you going to tell that talent to do?

Sometimes we see somewhat of the reverse of the above: we speak to people who are very oriented towards finding great founders, but don’t have a very particular view on what, exactly, these people will actually be doing. We definitely think you can build incubators “founder first” rather than “idea first”, with founders doing their own idea exploration (in fact, several organisations like Halcyon and Catalyze do this). However, even the best founders tend to need help channelling their energy towards interventions that can make the most of their talents.

In our case, we feel our research process and team has been crucial to our portfolio’s success. We run an extensive 6-month research process that evaluates ideas on their potential impact and develops them to the point where they are ready to be picked up by a founder. We share all of our methodology and training here. We think a dedicated research team has several advantages over founder-led ideation. First, evaluating evidence well is a specific, challenging skill, and not one that particularly overlaps with the skills that make someone a great founder. Second, researchers accumulate context across a field that allows them to spot patterns and failure modes that would be invisible to a newcomer. Third, and maybe most importantly, they don’t get personally attached to ideas and are biased towards not recommending rather than recommending: they are happy to toss 99 ideas they’ve worked on in the bin to find 1 that hits our bar.

It’s worth noting that we fully expect the organisation a founder actually builds to diverge from the research report they were handed. As the saying goes, no plan survives first contact with the enemy. Within months, founders will know more about the problem and implementation context than the person who wrote the report, and will identify ways things should be done differently to what was recommended or that they need to pivot entirely. But they’re far more likely to arrive at an excellent intervention if they’re pivoting from a well-chosen starting point.

We think how far towards the idea-first end of the spectrum you want to sit largely depends on how well-evidenced your space is. Where the evidence base is thinner, the marginal value of more desk research falls quickly and you should get to real world testing sooner, which is often best done by the founder. For example, in animal welfare, the evidence base is thinner than in global health. As such, for our upcoming dedicated animal welfare round, we’re experimenting with a hybrid model: rather than recommending specific interventions, we’ll give founders prioritized "frontier problems”, and support them through 3-month exploration prints in which they develop their theory of change themselves.

5. Do you have domain expertise and entrepreneurial talent on your team?

Do you have founding experience in your team? We believe it makes you a much better program designer, teacher and coach: you have insights into the common challenges founders face, the emotional journey of founding is like, and how to navigate the law of equal and opposite advice^[1].

Do you have domain experience in your team? AIM started with global health direct delivery because our co-founders had founded exactly that kind of charity. When we've branched out to other kinds of charity we've done better than you'd naively expect, but likely worse than if we’d had internal talent with domain experience. Incubating factory farming charities? Have people who know the industry intimately^[2]. Incubating policy charities? Have people who have experience driving policy change from the outside. Building direct delivery global health charities? Have people who have worked close to programs on the ground rather than worked from a distance. This seems obvious, but is worth emphasising, as the EA community often underrates accumulated learnings and overrates figuring things out from first principles.

Both gaps show up in the same failure mode: people who haven't launched programs from 0-1 or who haven't worked in the field, tend to think their charity ideas are fully fleshed out and shovel-ready when they're nowhere near. The crucial considerations that determine whether it's actually a good idea (often tractability, externalities, limiting factors) get left undebated. If you're going to hand an idea to a founder and ask them to spend years of their life on it, we think the bar should be higher than what we typically see.

6. What are your takes on how to beat the market on entrepreneurship?

Most startups fail. Most charities have negligible impact. It’s not good enough to merely incubate new organizations – you need to do better than the reference class.. to ‘beat the market’. To do this, you need takes on what is typically done wrong or overlooked, and how you will exploit that alpha. (Being effectiveness-minded is not enough).

For example, here are some of AIM’s takes:

• A lot of the best talent gets overlooked. We are willing to solicit thousands of applications per program and sift through them for diamonds in the rough. We weigh credentials and work experience less heavily than others, and look harder for overlooked indicators of being scrappy and agentic. We end up giving a chance to many founders that wouldn't get one otherwise. It's a massive hassle, and targeting very talent-dense groups (e.g. only prior founders) is an understandable choice for efficiency reasons, but please at least recruit from outside the effective-impact orbit so you're #growing-the-talent-pie.
• Structured co-founder matching beats most alternatives. The odds that your college drinking buddy is the best available co-founder are low, yet that's roughly how most pairings happen. We assemble a talent-dense cohort of ~15 and give people structure for testing fit, which we think produces far better pairings than the market default. To quote one of our alumni: “Structured cofounder matching is vastly better than any other method to find a cofounder. It's better than Y Combinator's cofounder matching platform, better than OnDeck's founder program in SF, better than personal networking. It's goated.”
• Help founders choose good interventions. Identifying the most promising org to found is a different skill from founding one, and a random shower thought or passion-project is unlikely to beat the outcome of a rigorous research process considering hundreds of ideas. This needn't mean recommending a set of very specific interventions (AIM’s historical model), for example you could use your research team to support founders on their search.

If you've read this far you may have noticed that these six questions collapse into two. Do you have a clarity on what's actually blocking the orgs you want to exist, and are you sure you're the right team to remove those blockers? The first question is what our framework post is for. The second question only you can answer.

We'd rather see more incubators than fewer. If you're working on one, get in touch. We'd be happy to share what we've learned, and to learn from your experience so we can improve our thinking.

Related materials we’ve previously published:

• Learnings from our Founding to Give (For Profit) incubation rounds
• Similar advice post from our founder, Joey Savoie - 3 years ago
• Interview between our CEO Samantha Kagel and Coefficient Giving
• Webpage describing how our program actually runs 

1. ^{^}

    Stole this concept from Scott Alexander, who points out that for any piece of advice that someone needs to hear, there is often someone else who needs to hear the exact opposite. For example (quoting): “You need to stop being so hard on yourself, remember you are your own worst critic” versus “Stop making excuses for yourself, you will never be able to change until you admit you’ve hit bottom.” Or “You need to be less selfish and more considerate of the needs of others” versus “You can’t live for others all the time, you need to remember you deserve to be happy as well.”

2. ^{^}

    The same logic applies to the founders themselves. Some charity ideas need a founder with a network, credential or identity that can't be built in a program. Our framework post covers how to incubate those well.