The 20% Who Reply to Recruiter Emails Are the Ones Nobody Else Wants

Source: Lenny's Podcast | Published: 2026-08-09T12:30:32Z

Cursor's head of recruiting Adam Ward points out that the 20% who respond to cold outreach are precisely the least sought-after candidates on the market — top talent never responds under those circumstances.

Adam Ward calls this the "death funnel." He first heard the term in 1998, during an internship at a company called Trilogy, where he went on campus recruiting tours across the country with his first manager, Jeff Daniel. His manager coined the phrase on that trip. Adam spent the next twenty years proving why it holds.

He's now Head of Talent at Cursor. Before that, he scaled Pinterest from 200 to 2,000 people, led a 65-person global recruiting team at Facebook that grew the technical organization from 1,000 to 10,000, and founded an independent recruiting firm called Growth by Design — which Cursor acquired outright. A recruiting firm being acquired by a tech company is rare in itself.

A 20% Response Rate Selects for the Bottom 20%

The logic works like this: you send out 100 recruiting emails, 20 people respond, 20% conversion. Most companies consider that normal.

But those 20 people aren't the best 20% in the market. They're just the ones who happened to be having a slow day when your email arrived — scrolling through their inbox, with time to spare, nothing urgent on their plate. The truly exceptional candidates are not the ones who respond under those circumstances.

You then filter those 20 people down sequentially, and ultimately hire whoever's left. Adam calls this remainder hiring — you're hiring the person who wasn't eliminated, not the best person.

The process is designed like a sales funnel, but the problem is that sales involves a relatively stable product. Hiring involves two irrational humans — the hiring manager and the candidate. Run a person through a transactional process and you get a transactional outcome. Recruiting as a profession is maybe forty or fifty years old; sales has centuries of history behind it. We habitually apply familiar frameworks to new problems, and in hiring, that habit creates systemic distortion.

Recruit Every Role Like You're Hiring a CEO

The alternative isn't complicated, but the execution is completely different: identify who's best first, then go after those people — instead of waiting to filter whoever finds you.

This thinking has always existed. Most companies just reserve it for C-level hires. Adam's argument is that every role deserves the same treatment.

Step one is definition. Before approaching any candidate, write down what "excellent" actually means — not "I'll know it when I see it," because that doesn't work, but specifically: what skills matter, what experience matters, in what order. That document drives everything that follows: how you find people, what questions you ask, why this particular offer is built for this particular person, and what you're ultimately evaluating.

Step two is identification. Find the top 50 people in the world for this role. That's not posturing — for a senior position, a deep enough network can genuinely surface a list of that size.

Step three is pursuit.

Ask the Wrong Question, and You'll Never Find the Best People

The most common mistake in sourcing is asking: "Who's the best product engineer you know?"

That's a bad question. It's too broad. There's nothing specific enough for the other person to latch onto.

Better questions sound like: "Of the product engineers you've worked with, who impressed you most in how they collaborated with designers?" Or: "Who had the strongest ability you've ever seen at turning a framework into a shipped product?"

Specific questions trigger specific memories. When you ask the same precise question to six or seven trusted people, the same names surface repeatedly — "Lenny mentioned him, Sally mentioned him too, let's add him to the target list." That cross-validation gives you a signal you can actually trust.

The questions must map directly to the traits you defined in step one. Only if you've thought clearly in advance about what matters can you ask valuable questions — and get valuable answers.

Stay in Contact, Even When They Say Not Now

Once you have the top 50 list, the real work begins.

This process isn't about asking people whether they want to interview. When someone says "the timing isn't right, maybe in the fall," that's not a signal to end the conversation — it's an opening to stay connected differently: "Want to come by the office for coffee?" "Want to meet someone on the team who's working in the same space?"

Adam says they're willing to stay in active pursuit of a candidate for weeks, months, sometimes over a year. Seeds germinate on different timelines. But you have to plant them first.

He describes the rhythm this way: plant broadly and simultaneously — don't stand in a single furrow waiting. Instead of sending 100 emails and hoping for 20 replies, plant a seed with each of the 50 people you actually want, and let them ripen at their own pace. And when someone finally becomes ready, the team moves immediately — the calendar fills that same day, leaving no room for second thoughts to take hold.

Get Them in the Building First

Early at Airbnb, Joe Gebbia had an internal strategy: "Get them in the elevator."

The office was upstairs. Getting there required an elevator ride, and that time alone was enough for candidates to feel something. Once they were in the elevator, the deal was half done.

Adam calls this home games — whenever possible, hold meetings on your own turf. He describes Cursor's office as something like "the Hogwarts library": warm, quiet, full of blackboards. They also opened Cafe Cursor in San Francisco, and the concept has since spread — spaces with a similar feel appearing in cities around the world.

When a candidate walks in and sees people working with focused intensity, when they absorb the atmosphere, sometimes they get a gut feeling of "this feels like home." That feeling can't be conveyed through a job description or a phone call. It requires the candidate to be physically present.

A Turkish Guitar That Influenced Millions of Dollars in Decisions

They were pursuing a candidate in New York. Through several deep conversations, they'd learned that he attended a performing arts high school, had trained in classical violin, and had a particular passion for stringed instruments.

The day they invited him to dinner to present the formal offer, Adam's team made a detour to a specialty instrument shop in New York first and found a rare instrument — something between a Turkish guitar and a violin — and presented it as a gift over dinner.

It cost a few hundred dollars. Against the size of the offer, that number was irrelevant. But the signal it sent was: we took the time to genuinely know you as a person. We remembered.

Adam calls this logic caring is free. In an industry he describes as transactional by default, most people are caught up in the funnel's numbers game — send 100 emails, screen 20, interview 10 — and once you're inside that game, the person disappears and only the numbers remain. When you shift your focus to "we are looking for a human being, not filling a slot," the cost of doing something unexpected is usually low. It just requires thinking one step ahead.

Side-by-Side Work Signals What Interviews Can't

Research consistently shows that work samples are the highest-validity predictor of whether someone will succeed at a company. Most companies still rely on conventional one-on-one interviews.

Cursor is known for extended on-site work sessions — candidates work alongside the team on a real project for several hours or an entire day. They once tried removing this step, on the grounds that it was costly and burdensome for candidates. Their confidence in hiring decisions dropped sharply, and they brought it back.

The reason is that real work surfaces signals far beyond technical ability: how someone collaborates, how they respond to ambiguity, the quality of their interactions with teammates — almost none of which is visible in a formal interview. And for candidates, after spending a full day in the actual environment, their own read on the job becomes far more reliable. You let them choose to join. And people who choose tend to stay.

The Most Dangerous Window Is After the Offer Is Accepted

Adam has noticed a new trend: offer rescission rates are climbing.

A candidate accepts, and then in the two or three weeks before their start date, a better opportunity appears, or they begin to waver. During this window, most companies do nothing — the recruiting team has already moved on to the next candidate.

Cursor deliberately designs a pre-onboarding phase: invite new hires to a group dinner before they start, introduce them to each other, build a sense of community before day one. Ship a Cursor laptop to their home, so they feel like a team member before they've even walked in. Stay in regular contact to reduce uncertainty.

Going through the entire process, getting to yes, and then losing someone at the final moment is among the most draining outcomes in recruiting. Investment during that window is the necessary cost of protecting everything that came before it.

Recruiting Teams Shouldn't Own Hiring Targets

A view that makes many companies uncomfortable: recruiting KPIs should not belong to the recruiting team.

Because hiring decisions are always made by the hiring manager, not the recruiter. So why is the hiring target the recruiter's responsibility?

Adam defines the recruiting team's role as a confidence engine, not a decision engine. Their job is to give hiring managers strong enough signal that they can make decisions with conviction. No one understands the pain of a bad hire better than the hiring manager, and no one understands the value of a great one better either. When you outsource the verb "hire" to the noun "HR department," the people who are actually accountable lose their urgency.

Two Competing Needs Shouldn't Fall on One Person

For early-stage companies, Adam sees a recurring mistake: stacking two contradictory needs onto a single hire.

You urgently need to fill 10–15 roles, and you also want to build a scalable recruiting infrastructure. In practice, these compete. Someone with the execution speed to fill seats fast may not have the inclination to build systems. Someone focused on building infrastructure might not close a single role in the first three months — and by then, you might already be losing sleep.

His advice is to decouple: hire one person to build the system, and bring in outside help to handle near-term demand. Keep the two tracks separate, rather than betting that one person can do both well simultaneously.

He also describes a trait of the best recruiters: they perform best at 90–110% capacity. Too little to do and they coast. Overloaded and they break. Right at the edge, they're most stable. Find people who become more energized, not more anxious, when they're operating just past their limits.

Make the Whole Team Grow Exponentially, Not a Patchwork of Stars

The most commonly misunderstood concept: talent density is not about assembling a collection of 10x engineers and putting them in the same room.

If you take a group of individually outstanding people and place them in a dysfunctional team, they won't be 10x — they'll be rate-limited. Adam believes that truly great managers are exceptional puzzle-builders. Their work is assembling people with different strengths into a whole that reinforces itself, producing exponential team growth rather than linear addition.

This perspective also explains how he builds his own team. People who could have gone elsewhere as Head of Talent choose to work as individual contributors under him instead. His selection criteria come down to three things: exceptional ability, a drive to prove themselves, and a willingness to put the team first. People who meet all three simultaneously are rare.

He tells his team: we should all be selfish enough to want to work with people who are genuinely great.