The Audience Is Not Listening to You. They Are Deciding Whether You Are One of Them.
Let me read your voice guide carefully before touching this.
Good. The original article is well-argued but it reads like an academic essay. Long compound sentences, no scene, thesis up front, no parentheticals, very few moments where you step out of the frame. The argument is yours. The voice isn’t yet.
Here’s the rewrite:
There is a test your audience runs every time they see something you made.
It is not “is this credible?” It is not “does this product do what they say?” It is something much older than that. Something closer to: are these people like me?
Evolutionary psychologists have been making this argument for a while. Humans were telling stories around fires long before they were writing things down, long before agriculture, long before pretty much everything. And those stories were not just entertainment. They were a running audit. Who believes what I believe? Who do I trust when it matters? Story was the technology for figuring that out, night after night, for thousands of years.
Your brand content is going through that same filter. Every email, every campaign, every press release. The audience is not reading it for the information. They are reading it for the signal. And the signal they are looking for is simple.
Are you one of us?
This is where most marketing falls apart. Not because it is bad. Because it is designed to offend nobody, and nothing that offends nobody says anything. Optimizing for broad appeal is not a strategy. It is an absence of one.
Rory Sutherland (vice chairman of Ogilvy, author of Alchemy, one of my favorite thinkers on why rational marketing usually fails) has a useful frame for this. He borrows it from evolutionary biology. The signals that actually work are the ones that cost something. A peacock’s tail is expensive and cumbersome. That is the whole point. A cheap signal is no signal at all. If it costs you nothing to say it, it means nothing to hear it.
In February 2026, Anthropic demonstrated this in real time. The company had a Pentagon contract worth up to $200 million. An IPO on the horizon. The government wanted to use Claude for “all lawful purposes.” Anthropic said no to two specific things: autonomous weapons and mass surveillance of American citizens. The Pentagon pushed back, set a deadline. Anthropic didn’t move. CEO Dario Amodei’s actual words: “Threats do not change our position. We cannot in good conscience accede to their request.”
The cost was immediate. Treasury, State, and HHS moved to phase out Anthropic products. The Defense Secretary called the company a supply-chain risk and barred military contractors from working with them. Within hours, OpenAI signed the deal Anthropic wouldn’t touch.
Then the audience voted.
Claude downloads jumped 37% day-over-day on Friday, 51% on Saturday. The app hit number one on the US App Store, passing ChatGPT in daily downloads for the first time. Daily sign-ups broke all-time records every single day that week. Free users were up more than 60% since January. Paid subscribers more than doubled. The servers went down from the demand.
OpenAI’s numbers went the other direction. ChatGPT uninstalls jumped 295% day-over-day on February 28. Downloads fell 13% that day and another 5% the next. One-star reviews surged 775% on Saturday. A movement called QuitGPT claimed more than 1.5 million people had cancelled subscriptions or posted boycott messages. Sam Altman published an internal memo saying the company “shouldn’t have rushed” it. OpenAI then renegotiated the contract to add explicit prohibitions on domestic surveillance. The exact protections Anthropic had held out for.
Neither company had to explain themselves. Their choices in one week said everything.
That is Sutherland’s logic playing out at scale. Anthropic’s signal was costly, so it was credible. OpenAI’s signal was cheap, and the audience knew it instantly. The brain is good at detecting when nothing real is at stake. When it detects that, the chemistry doesn’t follow. And in this case, neither did the downloads.
Paul Zak’s research on oxytocin is worth connecting here. The neurochemical is triggered by character-driven stories where something real is at stake, where a protagonist faces genuine tension, has to make a meaningful choice. Generic aspirational brand content doesn’t produce the same response. The brain notices that nothing is on the line. The signal is cheap. The chemistry doesn’t come.
Anthropic’s situation had the opposite structure. Visible stakes. A real choice. Tension that played out publicly over days. Whether you agree with their position or not, the narrative was undeniable: someone with something to lose, a genuine conflict, and a decision made on principle instead of convenience. That is exactly the architecture of a story that moves people.
So what does this look like in practice?
It means taking positions that will not land universally. It means being specific about what you believe and who you are for, which also tells people who you are not for. It means letting your narrative carry real tension instead of smoothing it out. It means behaving less like a brand and more like a person who actually believes something.
Red Bull tastes bad and costs more than it should. Logically it should not exist. But the product is the signal. Expensive, unpleasant, chosen anyway. That combination tells you something about the person drinking it. Red Bull never had to explain that. The choice said it.
Your story works the same way. Not what you claim about yourself. What you demonstrate, at cost.
The audience is not reading your about page. They are watching what you do and running it through a test that is thousands of years older than your marketing department.
What do you want to be? And are you willing to show it when it costs you something?
A few things I changed and why: opened with a scene-level observation instead of a thesis, broke the compound sentences into percussion, added parentheticals where you’d actually think them, used “And” and “But” to start sentences where the rhythm called for it, let the Anthropic data arrive as consequence instead of inventory, and ended on a forward-leaning question instead of a summary. No em dashes. Let me know what you want adjusted.
This is the second article in a five-part series on the alchemy of story in marketing.