The story this week is that four people in unrelated fields independently stopped optimising the top of their funnel - and started naming the small set they actually wanted.

💡 Top tip for this week's roundup - read the TL;DR, then jump to the section closest to your current problem. Ward on the funnel and Godin on the smallest viable audience are the two to read together.

Remember, we've built an ever-growing library of our top podcast summaries. Check it out here.

Featured in this round-up:

🕒 Estimated reading time: 9 mins. Time saved: 6+ hours! 🔥

Not your topic this week? Try these five instead:

———————————————————————————

There is one genuine disagreement running under all this, and it is worth putting to your own team.

Naming the small set assumes you can identify them in advance. Ward can - he has spent two decades building a network of people whose taste in talent he trusts, and Cursor has a bar narrow enough that the population is genuinely small. Most companies have neither. For them, the funnel is not laziness; it is what you do when you do not know who the fifty are.

Godin's answer is that this is precisely the delusion. The funnel feels like diligence but it is often avoidance - it is more comfortable to process a hundred applicants than to decide who you actually want and risk being turned down. His framing is uncomfortable on purpose: if people do not understand what you are offering, "that's on you."

The tension to watch: does naming the small set require a network most teams do not have, or is the network simply what you build once you stop hiding behind volume? Ward would say the second. Anyone who has tried to hire into an unglamorous category may reasonably say the first.

———————————————————————————

Recruiting looks like a funnel because that is how it is always drawn, and Ward's argument is that the drawing is the problem.

"Companies think about recruiting like a funnel. I'm going to reach out to 100 people," he explains on Lenny's Podcast. Some fraction respond, and you hire from the survivors. "By definition, that is not the top 20%. That is just the 20 people who you caught on a bad day. The way we all recruit is we kind of weed people out at each stage and we hire the remainder."

The alternative is not a better filter. It is a different starting point. "In your mind, okay, there are 50 people in the world who can do this. Let's go find those 50 people and you relentlessly pursue and you focus your energy on activating those people, not trying to find them out of the funnel of doom."

The word doing the work there is activating. In a funnel model the effort goes into attraction and screening. In Ward's model the pool is known and the entire job is persuasion.

Key takeaways:

The reframe extends to the question you ask your network, and this is the most immediately stealable thing in the episode.

"The mistake that most companies make is they'll say, 'Hey, Lenny, who's the best product engineer you know?' That is the worst question you can ask," Ward says. It sounds like the obvious question. It produces nothing, because "best" is unanswerable and the person you asked will default to whoever is most famous or most recent.

His replacement names a quality: "of all the product engineers you've worked with, who stands out to you as the person who's most collaborative with designers or who can translate a framework into a product better than anyone you've ever seen?" As he puts it, you have to name "these very specific characteristics, or skills, or needs for your team" - because that is a question a human can actually answer.

Cursor institutionalises this. There is a Slack channel called Hiring Ideas, and it is busy. "Someone will come across something impressive. A product, a post, overheard at a barbecue. And they'll drop that person's profile or that ex-post in Hiring Channel. And it becomes a rallying effect." The naming is not an annual exercise run by recruiters. It is ambient, and the whole company does it.

Key takeaways:

If the pool is named rather than filtered, the assessment has to carry more weight - and Cursor learned that by removing it.

Work trials are expensive for the company and a significant ask of candidates, so they ran the experiment of dropping them. Lenny raises that Michael Terrell had described the same experiment elsewhere: "he said that it just totally tanked their signal, and they had to bring it back."

Ward's framing of why is the part worth keeping. "We really struggled to have confidence in our decision. Again, this is a confidence building function." Recruiting is not a sorting problem, it is a confidence problem - and if you strip out the expensive step, you have not saved money, you have bought a worse decision. His conclusion is blunt: "if you are not doing this, you're lacking significant signal."

He is careful not to make it a universal law. Not every role or candidate suits a full trial - "some can do the lunch, some can do the greet" - but the underlying principle holds, that side-by-side real work is hard to beat for signal.

Key takeaways:

The same move shows up in a completely different problem: getting AI adoption to spread through a company rather than pooling in one enthusiast.

On The Skip, three CPOs at AI-native companies describe a consistent failure mode - "you've got that one person in the function really doing it really well, and then you've got the rest of the function really falling behind." The instinct is a training programme aimed at everybody. What they do instead is name people.

The screen is specific and demonstrable: "I ask people to screen share, I ask people to show me what they've built. You can screen for that pretty objectively and pretty quickly." That is Ward's principle in a different setting - not "are you good with AI", which is unanswerable, but a concrete artefact you can look at.

Then the naming pairs off. They look for "really curious, hungry people" and put each one alongside a functional lead. The logic is that a head of sales or marketing understands the outcomes they are driving but has no spare cycles "to be tinkering around with every single new model drop." Neither person alone changes the function. The named pair does.

Key takeaways:

Kipp Bodnar's ad-budget episode on Marketing Against The Grain arrives at the mechanic from the money side.

LinkedIn shows up among the most expensive channels in his 2027 breakdown - and yet his advice is that "LinkedIn should be a key part of every B2B marketer's advertising budget." That looks contradictory until you look at what separates the marketers who find it ruinous from the ones who find it their best channel.

His prescription is targeting narrow deliberately: do not target "your whole audience. But instead, a subset of your audience that really goes and aligns with the message." And on the cost objection, his warning is against being "blinded by the initial costs, like a crazy high CPM" instead of measuring the customers you actually got.

Same platform, opposite verdicts, one variable. A high CPM against a named, narrow segment is cheap. A low CPM against everybody is expensive in the only currency that matters.

Key takeaways:

Seth Godin, on The Diary of a CEO, supplies the psychology underneath all of it - and the reason the funnel persists despite not working.

His cure for the aspiring creator who feels overlooked is a question rather than a tactic: "we can get past the delusion by going for the smallest viable audience. What's the smallest group of people that if I made a difference for them, it would be enough."

He is clear that people resist this. "People hate it when I say this." The reason is that the alternative story is more comfortable: "I'm a genius and I'm really hardworking, but no one in the world understands me." A wide funnel and an unresponsive world keep that story intact. Naming twenty people and asking them directly puts the outcome back on you - and as he says of being misunderstood, "if people don't understand it, that's on you."

That is the honest reading of why the funnel of doom survives. It is not only inefficient. It is comfortable, because a low response rate can always be blamed on the market.

Key takeaways:

———————————————————————————

Four conversations, four disciplines, one shape: the people getting results have stopped optimising the top of the funnel and started naming the specific set they want.

Ward names fifty engineers and spends his energy activating them. The Skip's CPOs name the curious people in each function and pair them with a lead rather than running training at everyone. Bodnar names a narrow segment and accepts a high CPM to reach it. Godin names twenty readers instead of demanding the world pick him.

The practical move this week is to take one process where you currently cast wide - hiring, adoption, marketing, even a stakeholder comms plan - and try to write the actual list. Not the persona. The names. If you can write it, your effort should move from finding to activating. If you cannot, the gap is not that your funnel is too narrow. It is that you have not yet defined the specific quality you are looking for, and no amount of volume will do that work for you.

The uncomfortable half, per Godin, is that the wide funnel is doing emotional work as well as operational work. It keeps the rejection abstract. Naming the fifty means fifty specific people can say no to you - which is exactly why it is a better process, and exactly why most teams will not adopt it.

Three questions worth carrying into your team:

Want more of the same? Try these five:

That’s a wrap.

As always, the journey doesn't end here!

Please share and let us know what you liked or want changing! 🚀👋

Alastair 🍽️.