How to Build Curation Businesses

I previously wrote about how higher education is too expensive and that their dominance can be partly explained by regulation. In this piece I will focus on the structural reasons for why institutions like universities remain dominant for so long — mainly because they’re more defensible than we think.

Curation businesses — universities, bootcamps, accelerators — are incredibly powerful.

At first glance, they wouldn’t seem to be as powerful as tech companies, in the sense that they don’t have data moats or tech leverage. And yet, when you look at how long these institutions can dominate without being disrupted, they start to look that much more impressive. The last time a new university cracked the top 10 was over 100 years ago, whereas most tech businesses get disrupted every few decades or so. Indeed, curation businesses last centuries.

Curation businesses create socially acceptable context to build close relationships. For example, saying you met someone at Harvard is more powerful in terms of reinforcing the strength of a relationship than saying you met someone randomly at an event. It’s hard to deny that as a brand gets stronger, pride in the credential grows over time, reinforcing relationship strength even further. Intentional communities, fellowships, bootcamps, and “circles” are all trying to unbundle this age old paradigm, hoping to produce a similar effect on relationship building.

So, how do these “curation businesses” work?

Curation businesses attract the best people in the world to apply for them. Their alumni then go on to be immensely successful, which improves the flywheel and further attracts the best people at an even bigger scale. Because the feedback loop is often so long between their acceptance and their success, it’s often very hard to build a competing brand of similar integrity.

It’s unclear to what extent certain curation businesses actually help the people become more successful beyond the credential that they bring, which only emphasizes the strength of their brand. Their brand serves as a credential for the people who join, which helps them convince other people and organizations of their worth.

Indeed, if you’re building a curation business, the main (if not only) moat you have is brand. Your biggest specific vulnerability is alumni NPS — if it’s strong, the business is unbeatable; if it’s weak, the entire edifice is at risk.

Unlike with consumer products, a curation business’s brand can be a durable moat. Alumni are incentivized to promote the credential widely and speak highly of it, as to strengthen their own status accordingly. Their blind loyalty is rational: if they invested a lot to get that credential, and it becomes diluted, then they wasted their time and money.

Which is why once you get curation businesses off the ground, they are very difficult to disrupt, because everyone has incentive to keep the whole charade going. Like money, a credential is an illusion that everyone who has it or wants it is incentivized to keep perpetuating. If your status in life was partly acquired by going to the right nightclubs, you’re going to want to defend those nightclubs even decades after you attended.

To build curation businesses, you need to offer genuine new value to get the best people to join, since you can’t rely on the proof of credential. After you do, they’ll graduate and then over time be role models and the flywheel will begin. Until then, you need to bootstrap a 10x better value proposition to justify people choosing your curation business over a more established one. It could be aligning with pre-existing brands (like if a venture capital firm were backed by luminaries, e.g. Village Global), or a new value add approach.

What might a new approach for a new university look like? We think about that a lot at On Deck.

Byrne Hobart wrote a piece saying the future of education isn’t Lambda, it’s Y-Combinator. He says the real problem with universities isn’t that they’re just about signaling—though signaling is an important function!—it’s that they provide signaling value very inefficiently.

“No wonder college students spend so much time partying and protesting. They’re bored.”

“If you were reinventing the Ivy League as a signaling-focused product, your stripped-down version might look like this: you invite a small cohort of talented people to move to a city for about three months, you host some social events so they get to know each other, you have them work on projects and you advise them on those, and afterwards you introduce them to a bunch of savvy rich people.”

In other words, you’d invent YC or Village Global or other startup accelerators that provide a more focused version of the elite undergrad experience in about 95% less time, but also for adults.

Why hasn’t there been startup accelerator-like experiences for other categories? Why is this education limited to startups? It shouldn’t be.

Kanye West, of all people, tweeted that there should be a YC for music.


On Deck has been building the accelerator for X infrastructure over the last two years. We started with founders too, but pre-company, and have since added writers, angels, and podcasters, with many more to come.

When you think about it, founders want access to hiring, distribution, fundraising, and expertise. So we’re focused on expanding horizontally into verticals that fit with those categories. Two of the questions we ask ourselves when we’re thinking of launching a new category are:

  1. “Does this make our core founder fellowship stronger?”
  2. “Do we have an unfair advantage to build this fellowship, given our resources?”

I learned this lesson while at Product Hunt. I led our expansion into Product Hunt Books, and it failed. I could sit here and rationalize post-facto, but the truth of the matter is that (1) it didn’t make our core Product Hunt experience stronger, and (2) we didn’t have an unfair advantage to build it. Building “Product Hunt for Books” was like building a whole new startup, without benefitting much from the first startup we built. That’s a recipe for failure.

So, what else is important to get right when building curation businesses? As mentioned, NPS is everything. There are a few ways to have high alumni NPS:

Give someone their start or otherwise be an inflection point. Believe in them before others do and they'll be immensely grateful. If you create the turning point in their career or personal growth, they’ll remember you forever. We see this with startup accelerators — in Airbnb, Dropbox, and others, VCs put in 10x more money and time, and yet, in founder interviews, YC gets comparable love. Why? Because they were first to believe, and in some cases were responsible for them existing. You're always disproportionately grateful to the person & institution who gave you your break, so being able to catalyze significant growth moment is an arbitrage opportunity..

Be legible (with your cadence, credibility, or curriculum) so that people have similar experiences and can bond over them. Legibility provides social context for other people across cohorts (even years apart) to build meaningful relationships.

Universities and startup accelerators are quite legible: Their application processes are very clear, students go through common experiences (orientation, graduation, demo day), and people reinforce status loops by giving back (Alumni helping with interview prep, etc).

Alumni NPS can be affected by class size as well.

In my opinion, the two best options for scaling a curation business are either:

  1. Be so curated such that every person is extremely high quality.
  2. Provide so much value that the best choose to join regardless of the average member quality.

When it comes to class size, the tradeoff is "signal" vs "utility". Signal itself has utility, but there are diminishing returns to it. You want to increase utility as much as possible without sacrificing signal. In other words, admit enough people to provide the value, but not enough to kill the signal.

But there's more to signal than filter — for example, the Thiel Fellowship has a higher filter than YC, but many would choose YC because it has higher utility: it’s more legible to investors, hires, your parents, and has more startups & alumni who can potentially become customers, hires, and/or investors.

Indeed, for utility, quantity has a quality all its own. The scalable curation businesses start with signal and then use signal to build utility. The first YC cohort was super high signal — in retrospect at least — and they used that to justify scaling 200x ("statistically, one of the companies can be the next Airbnb...")

If you’re building a curation business, your goal should be to start with a meaningful signal and then optimize for utility as soon as possible, without compromising signal. This is why YC scaled to 200+ companies and didn't miss a beat, and why Harvard can scale too (to some point).

One reason why this works particularly well with accelerators is that startups are a power law game, and the most important thing is capturing the outliers. Also, given the best unicorn rate among VCs is 5%, it's unclear what signal means. Thus, in my opinion, it’s better to make more good bets than "sure" ones. Indeed, the genius of the YC credential that it's tied to their winners, not their median outcomes. It's less that you'll be impressed with the average company — it's more so that "1 in X startups will be the next Airbnb." Even as X rises, the credential means something because the whole point of the game is to not miss the next Airbnb. Which is why YC invests in 400 companies a year and yet many startups still choose to say “YC-backed” over any other firm in their press materials. Those two facts shouldn’t be true, and yet they are.

So, given this, why doesn’t Harvard try to accept 100K or 1M students? It's partly true that credential quality is tied to its exclusivity or its size, but it's also true that credential quality is tied to the magnitude of its biggest outcomes. Keeping the same quality bar, and if it has the ability to scale resources accordingly, you could argue that Harvard should be 10-100x its size.

Utility pulls you in the direction of expanding size. The more people there are, the more total opportunities there can be. Filter pulls you in the direction of less. The more people there are, the less likely the quality of the average person.

Focus too much on utility (get too big), and you kill the signal. Focus too much on signal itself without bolting on utility, then it becomes like Harvard — people know it has no purpose other than the signal and that itself weakens the signal. It becomes a bubble.

Either go so small that everyone who attends is super impressive (e.g. Thiel Fellowship). Or go so big that you capture most of the big winners and people value the signal out of FOMO. Don't be in the middle.

There's some Laffer-curve like point where if you're perceived as too big, even the FOMO of the winners won't override the feeling that anyone can get in. The trick here is to grow as big as you can while keeping your acceptance rate super low.

Certain communities have similar dynamics as nightclubs: You get the VIPs in the door that'll bring everyone else, and everyone else brings the money (or the data, or the hires) that make the community more valuable for the VIPs. If you do it right, you'll build strong network effects.

Stanford, Harvard, Princeton, peter out at under 7,000 students. In a virtual world, why couldn't they accept more? Partly because, like nightclubs, status is measured not by how many people get in, but by how many don't.

On Deck's goal is to accept 100,000 fellows across dozens of fellowships (at 1/100 of the cost) — we want to keep quality high while constantly expanding access.

We hope to bring back positive-sum spirit into education — because education is inherently a positive sum: when I learn something, I can share with others. But current top universities see it as a zero sum game — the more people that attend, the less valuable the credential becomes. Positive sum credentialing would be peer to peer — where giving one is a credential itself in terms of evaluating talent.

We aim to build products and infrastructure that will make our communities seem small and intimate, while also benefiting from the utility larger networks can bring. Our different fellowships will add utility to each other, making the whole ecosystem stronger. Or at least that’s the plan, stay tuned.


Read of the week: The Real Class War, in response to my tweet about the culture war being a front for the battle between the top 1% versus the top 0.1%.

Listen of the week: I interviewed Elad Gil and Kevin Hartz.

Watch of the week: The Great Awokening with Reihan Salam and others.

Cosign of the week: Stephen West. I found a new favorite philosophy podcast, went to look up where Stephen did his PhD, & instead found this.


Until next time,