Chapter 1 · Part One: Why Your Best Work Keeps Coming Out Average
A couple of years ago I spit in a tube and mailed it to a company, the way you do.
This is what came back.

Hey Nathan! You have more Neanderthal DNA than 91% of other customers.
I paid money to find that out.
Ninety-first percentile. Put seven hundred people in a room and I’m ahead of six hundred and thirty of them, which also means about sixty people in that room are carrying more of him than I am and have no idea.
In fairness to the guy, I went and looked at what he actually gave me. The report lists the variants. Here’s the measurable legacy of forty thousand years of human history as it applies specifically to me:
- A worse sense of direction.
- More likely to sweat during a workout.
- Less likely to blush easily.
- Two separate genetic variants that make me prefer dark chocolate over milk chocolate.
That’s it. I get lost, I sweat, I don’t blush, and I won’t eat a Hershey bar.

Why he was right
Here’s the part that makes him worth keeping around, and it isn’t the chocolate.
For almost all of the time humans have existed, the familiar thing was the correct answer. Not the safe answer. The correct one.
Think about what an unfamiliar option actually cost back then. A plant nobody in your group had eaten. A route nobody had walked back from. A shelter nobody had wintered in. The upside of being right about any of those was a slightly better week. The downside of being wrong was that you were dead, and so, often enough, was everybody who followed you.
That’s an asymmetry, and it ran in one direction for a very long time. When the penalty for novelty is that severe and that permanent, “do what worked last time” isn’t laziness. It’s the highest-value strategy on the board, and the people who ran it are the reason any of us are here to argue about homepages.
So recognition became a signal, and a genuinely good one. If you recognize it, somebody survived it.
That’s the machinery. It was built for a world where familiarity was real evidence, and it’s still running at full power in a job where familiarity is evidence of nothing except that eleven competitors got there first.
Now, in fairness, and I’d rather say this myself than have you catch me at it: the instinct isn’t actually a Neanderthal trait. It isn’t in that two percent. It’s human, it lives in the other ninety-eight, and it would be sitting inside every one of us if the two percent had never happened.
I just like having somebody to blame. He’s a useful character and I’m keeping him.
What he actually does
The instinct has a modern name and it isn’t a metaphor. It’s one of the most replicated findings in social psychology, and once you know it you’ll see it operating in every meeting you attend for the rest of your career.
Processing fluency. The easier something is for your brain to process, the more you like it, and the more likely you are to believe that it’s true.
Not “the more familiar it feels.” The more likely you are to believe it’s true.
The original demonstration is Robert Zajonc in 1968. Show somebody a shape they’ve never seen and they’re neutral on it. Show them the same shape several times and they report liking it more. They can’t tell you why. Ask them and they’ll say it’s a nicer shape. The exposure did the work; the explanation got invented afterward to cover for it.
The effect doesn’t stop at preference. It leaks into judgment:
- A statement set in an easy-to-read typeface is rated more likely to be true than the identical statement set in a harder one.
- A claim you have encountered before feels more accurate than one you haven’t, whether or not it’s accurate, and repetition strengthens the effect.
- Statements that rhyme are judged more accurate than non-rhyming statements that mean exactly the same thing.
- Companies with easier-to-pronounce names have been shown to be judged as better bets than companies with harder ones.
Nobody in any of those studies was fooled by a bad argument. They weren’t evaluating an argument at all. Their brain noticed that the input went down smoothly, filed the smoothness under “this is correct,” and handed the conclusion upstairs to the part of them that believes it makes decisions.
The largest single driver of fluency is repetition. Which means every time a sentence circulates inside your category, it becomes more true-sounding to everybody inside your category. Including your buyers. Including your CEO. Including you.
So picture the meeting

It’s a Tuesday morning. Somebody presents work.
It looks like the category. It uses the words the category uses. It has the structure every deck in this space has had for six years, which is precisely the structure everybody in the room has seen the most, which is precisely why it feels correct.
And everybody nods.
The nod feels like validation. It feels like proof. Nine smart people, several of whom can’t agree on anything else, independently arrived at yes.
It isn’t proof. It’s him. It’s your Neanderthal at the table, recognizing a shape he has seen before and reporting back that it’s safe, and that report arrives dressed as a feeling of correctness, and it gets there before the thinking does.
He isn’t wrong that the work is familiar. He is answering a completely different question than the one you asked.
So here’s the distinction I’d like you to carry out of this chapter. When work feels right, trust that. That’s judgment, and it usually arrives with a reason you can say out loud. When work feels fine, when it looks like everything else and nobody objects, that’s him. Right is a signal worth trusting. Fine is the red flag.
Here’s the version of this I want you to carry around, because it’s the most useful sentence in the first half of this book:
When a room agrees with your work immediately, that’s information. It’s just not the information you think it is.
Immediate agreement usually means you made something familiar. And familiar is the single property you can’t afford in a market where everybody has the same tools, the same playbooks, the same conference circuit, and now the same models.
Psychologists call the skill you need here metacognition: noticing your own thinking while it’s happening. I call it catching him in the act. You don’t have to stop the nod. You just have to notice it, which is harder than it sounds and gets easier every time you do it.
Sometimes the room is just right
I want to be careful here, because the lazy version of this chapter turns you into somebody who treats every agreement as a red flag and every objection as insight. That person is exhausting to work with and also frequently wrong.
Rooms agree for good reasons all the time. The work is clear. The argument holds. Somebody finally said the thing everybody has been circling for a month. That happens, and when it happens the agreement is real and you should take it.
Three questions separate the two, and you can run them silently in about fifteen seconds.
One. Could they have predicted it? If you had described the assignment to this room a week ago and asked them to guess what would come back, would this be close? Agreement with your own prediction isn’t evaluation. It’s recognition.
Two. Did anybody change their mind about anything? Real agreement among adults usually involves at least one person visibly moving. If nobody moved, nobody was persuaded, which means nothing was actually being decided.
Three. Can they each say why, differently? Ask two people separately to explain what’s good about it. Genuine agreement produces different reasons for the same conclusion. Fluency produces the same reason in different words, and that reason is usually a restatement of the work itself.
A room that agreed fast, predicted the answer, moved nobody, and can only justify itself by paraphrasing the deck hasn’t evaluated anything. It has recognized something.
The Dissent Protocol
If the nod is the problem, the fix isn’t to demand more criticism. Asking a room to be more critical produces performed criticism, which is worse than silence because it looks like rigor and consumes the same amount of time.
The fix is to change the order of operations so the fluency reaction never gets to speak first. Five rules. None cost anything. None require you to be in charge.
1. Written before spoken. Nobody talks for the first four minutes. Everyone writes their reaction privately, on paper or in a document nobody else can see yet. The moment one person says “I like it” out loud, the room’s fluency and that person’s seniority have fused into a single signal, and you won’t be able to pull them apart again for the rest of the meeting.
2. Score, don’t react. Ask for numbers, not opinions. Two of them: how clear is this, one to five, and how different is this from what our three closest competitors would produce, one to five. High clarity paired with low difference is the danger zone, and it’s where most approved work in this industry lives. You want that split visible on the page before anybody has a chance to rationalize it.
3. Make dissent cheap, but don’t fake it. Rotate one named person each week whose job in that meeting is to raise the strongest case against. It lowers the social cost of disagreeing, which is the main reason dissent doesn’t happen on its own. But know what it can’t do. Charlan Nemeth at Berkeley has studied dissent in groups for decades, and her finding is that a genuine dissenter makes the whole group think better, even when the dissenter turns out to be wrong, while an assigned devil’s advocate mostly doesn’t produce the same effect. The rotation isn’t the dissent. It’s cover, so the person who actually disagrees can say so without having to be the difficult one.
4. Ask the competitor question. If our closest competitor saw this tomorrow, what would they think? The useful answers are the bad ones. If the honest answer is “nothing, they wouldn’t notice,” you have your finding. The question works because it moves the room from judging quality to judging distinctiveness, and people conflate those two constantly.
5. Find the one who didn’t nod. This is the most important rule in the chapter and it’s the one I personally failed on the single most expensive mistake of my career, which is Chapter 8.
In almost every room there’s one person who doesn’t nod. They go quiet, or they ask a narrow technical question that seems beside the point, or they raise something easy to dismiss because they lack the full context. Spend the meeting there. Not because dissenters are smarter, but because the nod is the majority reaction and the majority reaction is the one you already know. The only new information in the room is sitting with the person who didn’t produce it.
And notice the specific trap, because it has a shape and it will come for you: the more people are nodding, the easier it is to discount the one who isn’t. The volume of agreement reads as evidence against the holdout. That’s the same mechanism firing twice, once to make the work feel right and once to make the objection feel wrong.
What to do tomorrow
Take the last thing your team approved. Not something in flight, something already shipped, so nobody has to defend it.
Ask three people who were in that approval meeting one question, in writing, separately, without discussing it first: what changed between the first version and the approved version, and why?
If the honest answer is a list of things that got clearer, smoother, shorter, and more consistent with how we normally do it, then you have just watched the process work exactly as designed. You have also found the precise place your distinctiveness went.
That isn’t an argument for burning down your review process. It’s an argument for putting one thing in front of it: a written record of what made the work different in the first place, so that in round three somebody can hold up the list and ask whether any of it survived.
The review won’t ask that on its own. It isn’t built to. It’s built to make things easier to process, and it’s extremely good at its job.
He kept your ancestors alive. He is extremely good at his job too.
He is just terrible at making you interesting.
The Comfort Check
Chapter 1: the review meeting
1. The comfortable assumption. When the room agrees, the work is good.
2. The true objective. Find out whether this work would be noticed by somebody who didn’t make it.
3. The redesigned approach. Written reactions before spoken ones, two scores instead of one opinion, a rotation that gives the real dissenter cover, and the meeting spent on the one person who didn’t nod.