Building Startups that Last (Part 5 of 5): Farming Talent over Finding Talent

This is Part 5 of a 5-part series on the critical factors that startups need to win.

Every founder says they want to hire the best people. Then they write a job description that describes someone who already works at Google.

You can't afford that person. The frontier AI labs and the biggest tech companies pay their top engineers packages that can reach seven figures. You have a seed round. If your talent strategy is to outbid them for proven stars, you've already lost.

And you shouldn't want to win that game anyway.

Sketch of a row of seedlings growing taller from left to right, with a watering can — Farm, don't hunt

The Star Myth

We assume talent is portable. Hire someone who was great at a big company, and they'll be great at yours.

The research says otherwise. Boris Groysberg at Harvard Business School studied more than a thousand star analysts on Wall Street and tracked what happened when they switched firms. Their performance dropped, and it didn't come back. He wrote it up in Chasing Stars. The stars' success depended far more than anyone realized on what surrounded them: their firm's resources, systems, colleagues and reputation. The only stars who held their performance were the ones who moved with their whole team.

Matthew Bidwell at Wharton found something similar when he compared outside hires with people promoted from within into the same jobs. The outside hires were paid 18 to 20 percent more. They got worse performance reviews for their first two years, and they were more likely to leave.

For a startup, that's a warning. The star from a big company built their track record on top of that company's machine: its data, its infrastructure, a brand that gets customers to take the call. You don't have any of that. So you pay a premium for the person and get them without the context that made them a star.

Hire for Slope, Not Intercept

John Ousterhout, a computer science professor at Stanford, has a line I love: "A little bit of slope makes up for a lot of y-intercept."

The y-intercept is where someone starts: their résumé, their years of experience, the logos they've worked for. The slope is how fast they learn. Most hiring processes measure the intercept, because it's easy to read off a résumé. But at a startup, the job you're hiring for today won't be the job a year from now. Someone with a steep slope will outgrow the person with the impressive starting point, and at a startup that can happen within a year.

Big companies can afford to hire for intercept. They have defined roles, stable processes and time. You have none of those. You need people who grow as fast as the company does.

So what does slope look like in an interview? Claudio Fernández-Aráoz, who spent decades in executive search, argued in Harvard Business Review that potential beats experience, and he named the signs: the right motivation, curiosity, insight, engagement and determination. Here's how I'd translate that for a startup:

They taught themselves something hard without being asked. Curiosity shows up as evidence, not as a claim. What did they learn on their own, and how far did they take it?
They finished something difficult. A side project, a hard degree, a tough job, a comeback. Determination is the best predictor I know of someone getting through a startup's lows.
They ask better questions than they answer. In the interview, the people with slope interrogate the problem. They want to know how things actually work.
They care about the mission, not the title. If the first questions are about level and title, they're optimizing for the intercept too.

Then test the work, not the pedigree. A paid work sample or a real problem from your backlog tells you more in two hours than a résumé does in two pages.

Where the Hidden Gems Are

If you're not competing for the same stars as everyone else, you get to look where they aren't looking.

Start with early-career talent. It's the biggest opening I've seen in years. Big tech has pulled back hard on hiring new graduates. SignalFire's talent report found that new grads made up just 7% of big tech hires in 2024, less than half of 2019 levels. Plenty of sharp, hungry people are hearing that nobody wants to train them. Be the company that does.

Then look next door to the job. Some of the best people come from adjacent roles: the customer support rep who understands the product better than the PMs, the analyst who taught themselves to code, the journalist who knows exactly how your customers think. They bring context that no amount of conventional experience gives you.

And look for people others underestimate: career switchers, people coming back after time away, people from schools and places your competitors don't recruit from. Your competitors screen on pedigree because it's easy. That leaves these people for you.

AI helps here. At elvex, we see it every day: a curious person with good AI tools gets productive much faster than they could even a few years ago, so the gap between someone with slope and someone with experience closes sooner.

Farming Is the Work

Finding raw talent is only half of it. The other half, and the half most startups skip, is farming it: setting people up to get good, fast.

Most of that comes down to management. Gallup found that managers account for at least 70% of the variance in employee engagement. Who someone reports to shapes how they do more than almost anything else you control.

What farming looks like in practice:

Give real ownership early. Earlier than feels comfortable. People with slope grow by being handed something real and a little too big. This is the betting culture from Part 4 applied to people: know your affordable loss, make the bet, and don't punish a smart bet that didn't work.
Shorten the feedback loop. Annual reviews are for companies that can afford to wait a year to find out someone's off track. Give feedback weekly, and make it specific.
Pair them with people who've done it. You can't afford a building full of veterans, but one or two experienced people who love teaching multiply everyone around them. Hire for that trait in your senior people.
Let them move. When someone outgrows their role, move them into a bigger or different one before they go looking elsewhere. LinkedIn's research has repeatedly found that people who make internal moves are more likely to stay.

Why This Is an Advantage

You could read all of this as making the best of a bad situation. I see it the other way around.

People you farm cost less, and Bidwell's research suggests they perform better in the critical first years. They already know your customers, your product and your culture. They stay longer. And because they grew up inside your company, they become the people who carry the culture to everyone who comes after them. Remember Part 3: the founding team's values only scale if someone else holds them too.

People also remember who bet on them first.

None of this means you should never hire experience. Some roles need someone who has done it before, and those are usually the decisions that are hardest to reverse: your first enterprise sales leader, your head of security, the person setting up your finance function. Buy experience where a mistake is expensive and slow to fix. Farm everywhere else.

Building Startups That Last

That's the end of the series. Looking back, the five parts make one argument. The startups that last are the ones that:

Execute instead of waiting for luck (Part 1).
Move fast and stay alive, balancing speed and persistence (Part 2).
Build the founding team's foundation before they need it (Part 3).
Bet well: know when they're stuck on a local peak and how to choose the next mountain (Part 4).
Farm their talent: find people with slope and give them room to grow (Part 5).

You don't need a giant round or a famous name for any of these. You need to do the work before you're forced to.

Thanks for reading all five parts. If one of them changed how you're building, I'd love to hear which.

This is Part 5 of a 5-part series on building startups that last. Read Part 1 here. Read Part 2 here. Read Part 3 here. Read Part 4 here.