Building an Outlier Firm: What Makes Amplify Different

I’ll be the first to admit that I’m not known for many things, but if there's one thing that's gone right in my 17+ years in venture capital, it's building what might actually be an exceptional team at Amplify. Since 2012, I’ve been hiring people who keep becoming partners — which is as good a mark of success as any in this business. And so far, three out of our first four funds are top 5% in their vintage. 

I’ve been thinking about what makes certain venture firms work while others don't. I believe the answer might be simpler than anyone wants to admit: outlier firms are built by outlier investors looking for outlier founders.

The Outlier Business (Or: Why Most VCs Are Doing It Wrong)

Every early stage venture capitalist will tell you that our business is an outlier business (shocking, I know). The entire job is outlier identification. We've seen time and time again that a tiny percentage of startups succeed at the scale required to make venture economics work. They even wrote a book about the so-called Power Law. Fund managers are well aware that outliers drive their returns.

But here's where it gets interesting (and where most VCs completely miss the point): if you accept that this is an outlier business, then how exactly do you find outliers?

I would posit that outliers cluster together. This isn't just true in venture—it's true everywhere, and the evidence is overwhelming once you start looking for it. You see it in sports, you see it in music, you see it everywhere: outliers cluster. Why? Because people who are exceptional like to be around other people who are exceptional. 

Take Monet and Renoir, who met as struggling art students in 1860s Paris and painted side by side - their friendship helped birth Impressionism. Or take Spielberg and Lucas - two film nerds who met in 1967 when teenage Spielberg caught Lucas's student film at a festival. They didn't actually become tight until 1971 at Francis Ford Coppola's house, watching Spielberg's TV movie "Duel." Three ambitious kids who'd go on to basically invent the modern blockbuster and reshape how Hollywood works. It's like that old saying about how you become the average of the five people you spend the most time with, except these people were already exceptional and just needed to find their tribe.

So if you believe that venture capital is about outliers, and you believe that outliers cluster, there's really only one way to hire and construct your firm: you hire outliers, and you let them tell you where to invest.

The Committee Problem (Or: How to Kill an Outlier in 10 Easy Steps)

Here's where things get interesting, and by interesting I mean "completely maddening." Most VC firms understand (at least intellectually) that you need to hire outliers. But then — and this is where the whole thing falls apart — they structure themselves in ways that completely dilute the outlier effect.

They do this through some sort of voting structure:

• Deal voting, where all the partners vote on whether a deal gets done, or
• An investment committee (a subset of partners) who vote on deals

What both of these approaches mean is that the outlier needs to convince other people to see what they see. But here's the thing: early-stage deals often look, well, completely ridiculous.

At the risk of alienating some of our most beloved founders, I won’t name names, but you can take a gander at our portfolio and take a guess (and you probably won’t be far off) – some of them were completely crazy when we backed them. Our best deals are almost never universally loved at the time of our first investment. And that’s the point. The point isn’t that everyone loves it. It’s that someone sees it — even if nobody else does.

Picture this scenario (and if you've worked at a traditional VC firm, you've lived this): The outlier investor thinks they see something that everyone else doesn't see. The first time around, they attack the process with the zeal and vigor you'd expect from an outlier. They prepare detailed memos, create compelling presentations, maybe even bring in the founders to pitch directly to the committee.

But then they're told "no." Maybe the committee thinks the market is too small, or the founders are too young, or the idea is too crazy, or any of the hundred other reasons that committees use to kill deals that don't fit their mental models.

Maybe the first “no” isn’t completely demoralizing. Maybe the outlier investor tries again, and again. But as the “nos” continue to roll in, at some point, that outlier stops running after those deals with the same intensity. Over time, they start sourcing deals that they think will get approved, not the deals that excite them most. Group voting structures rob outliers of their agency and completely defeat the purpose of hiring them in the first place!

It's like hiring Jimi Hendrix and then telling him he can only play songs that the record label (i.e. the committee) approves. You'll get perfectly acceptable music, but you'll never get "Purple Haze." A manager who properly understands the outlier effect must accept that she will not always understand the deals they want to do. And that’s OK.

The Math Problem (Or: Why Being Right Doesn't Matter If You Miss the Big One)

The thing with outliers is that you can be wrong a lot – investing in a lot of companies that don’t end up working out – and it's perfectly fine as long as you're right once in a while. But you absolutely cannot afford to be wrong about the fundamental law itself. You cannot miss the actual outliers.

Here's the terrifying math: the investment committee can be right 9 times out of 10. They can correctly identify deals that would have lost money, founders who weren't ready, markets that weren't real. They can pat themselves on the back for their discipline and risk management. But if they're wrong on that 10th deal—if they miss the company that would have been the outlier returning the entire fund—they haven't just lost money on that particular investment. More perniciously, they've sent a clear message to their outlier investors: only bother bringing us deals that we'll like, even if your instincts about that "crazy" company turn out to be completely correct.

In binary classification this is referred to as the false negative. And in venture capital, false negatives will kill you faster than false positives ever will.

The Amplify Approach (Or: How to Actually Let Outliers Do Their Thing)

We've tried to structure Amplify to operate in a way that gives our outlier investors the autonomy to make their own decisions. We do not have deal-by-deal voting, nor do we have an investment committee. The decision to make an investment rests in the hands of a GP, and a GP alone. Whatever they see — and the rest of us may not always understand it — they can back.

This carries its own organizational overhead (imagine the fun of explaining to LPs why you don't have "proper oversight"), but I believe it's the only way to run an early stage venture fund if you really want to find outliers.

Amplify’s history is full of moments where one of our investors saw something the rest of us didn’t. They pounded the table, we found a way to say yes. Not every deal worked out. But when they did work — when they really worked — they became some of the most exciting and biggest companies in our portfolio. Companies like dbt Labs, Chainguard, Temporal, and Runway – all incredible today, but entirely non-obvious when Amplify led their respective seed rounds.

Hiring Outliers (Or: How to Spot Them in the Wild)

Hiring outliers is probably a topic that deserves its own post (and knowing my propensity for tangents, it probably will get one). We tend to hire much slower than other firms. We get to know people over longer periods of time. But you can really tell if someone is an outlier based on who they spend their time with.

Years ago, I met someone who ended up becoming one of our first junior investors. But before she got the job, she invited me and my partners to a dinner she organized. Everyone there was sharp — including Andrej Karpathy, then an under-the-radar AI researcher who’d go on to lead Tesla’s Autopilot and advise OpenAI.

It's like that phenomenon where you meet someone new and immediately want to know who their friends are, because extraordinary people somehow always seem to know other extraordinary people. It's not coincidence—it's selection. You know you’ve found an outlier when they’re surrounded by other outliers (and often they don’t even realize it).

All of the investors we’ve hired who ended up becoming partners here, it was obvious at the time that they were outliers. You could tell by their orbit. They were just always around great people.

The Bottom Line (Or: Why This Actually Matters)

We’re trying to build Amplify as a place where outliers come to do their best work. . We do our best to hire outliers and then—and this is the crucial part—we get out of their way. We don't make them convince an entire partnership of what they see. We give them a canvas where they can do what makes them special.

In an industry obsessed with pattern recognition and risk management, we've bet everything on pattern breaking and intelligent risk-taking. It's worked out pretty well so far, though I acknowledge that saying this out loud probably means we're about to have our worst fund ever (that's just how these things work).

But if venture capital really is about finding outliers, and outliers really do cluster together, then maybe—just maybe—the best way to find them is to build a firm full of outliers and let them loose on the world.

Seems pretty obvious when you put it that way, doesn't it?

# antipartnervotingclub