A spreadsheet landed in Tim Ferriss's inbox in June and he did something most authors would never do. He published it.

The post went up on his blog on June 12 under the title Has AI Already Killed How-To Nonfiction? His framing on X: "let me show you, using my own books as the cadaver on the table, what a fatality looks like."

That is the man who more or less built the template every coach has copied for twenty years. Write the book. Get the stage. Get the podcast. Get the fee.

The numbers he showed

Ferriss opened his own sales numbers and reported his catalog down nearly 80 percent since ChatGPT launched, against a self-help category that fell 26 percent in a single year.

The decline compounded rather than crashed: an early drop of around 5 percent, the kind of fluctuation that is normal in publishing, then 43 percent in 2025 versus 2024, then a 57 percent fall in 2026 against the first quarter of 2025, leaving him on track to sell roughly 80 percent fewer copies than he did in 2022. He dates the start of the trouble to late 2022, when ChatGPT was released. It is a particular problem for him, because he writes how-to books.

His conclusion, posted four days later: "The death of prescriptive nonfiction books, at least as a mass-market information business, is nigh."

The category data behind him

This is where it stops being one author's bad year.

In the first quarter of 2026, adult nonfiction print sales dropped almost 9 percent. Self-help fell 26.3 percent. Only two of sixteen subcategories grew at all: religion, and crafts and hobbies. Overall print units fell in the period at outlets reporting to Circana BookScan.

The half-year figures were not a recovery. US print book sales declined just 0.3 percent across the first half of 2026, with adult fiction up 5.7 percent and adult nonfiction down 5.8 percent. Self-help had the biggest drop of any nonfiction subcategory, falling 24.3 percent.

Fiction grew. The advice aisle emptied.

The part that complicates the story

One book explains a lot of the gap. Publishers Weekly attributed the self-help decline largely to the cooling of Mel Robbins' The Let Them Theory, which was still the biggest nonfiction seller of the first half. It sold more than 566,000 copies in the first half of 2026, down from 1.76 million in the same period of 2025.

And the year before that, the category was booming. Self-help units rose 14.7 percent across full-year 2025, when The Let Them Theory sold more than 2.8 million copies.

There is other noise in the data that has nothing to do with chatbots. Young adult fiction fell 25.6 percent, while graphic novels rose 36.2 percent, helped by improved distribution after the 2025 bankruptcy of Diamond Comic Distributors. Categories move for boring reasons.

Ferriss's argument is that his own curve does not fit the boring reasons. Others point to generational shifts or social media; he sees AI as the key change, and tells authors to build on voice, stories and true fans rather than commoditized information.

What the book is actually still buying

Here is the part that matters if you are a coach with a manuscript in Google Docs.

The book was never really the revenue. It was the positioning instrument. And the positioning market has its own price list.

The National Speakers Bureau's 2026 guide puts entry-level speakers at $1,500 to $5,000, mid-range speakers with proven track records at $5,000 to $15,000, professional speakers with bestselling books and executive backgrounds at $15,000 to $30,000, and celebrity names from $30,000 past $100,000. The entry tier explicitly includes authors with their first book.

What that tells you: a book does not move you to the top of the market. A book that sells does.

The spread underneath is wider than most coaches assume. An AAE Speakers survey of 340 event organizers and 378 speakers found the average planner budget was $22,449, with 47 percent between $10,000 and $50,000, while the average fee reported by speakers was around $14,000 and only 3 percent charged above $30,000. Talkadot's State of the Speaking Industry 2026, built from more than a million audience survey responses across tens of thousands of engagements, puts the market median keynote fee at just $2,500, with the 90th percentile at $10,000.

Two credible datasets, two different markets. The bureau world pays professional rates. The wider world pays local-event money. A book is one of the few things that reliably moves a speaker between those two markets, and that is still true whether it sells 3,000 copies or 300,000.

The counterargument

Not everyone read the spreadsheet as a death certificate.

Writing in Inc., executive coach and author Henna Pryor accepted the data and rejected the verdict. She has a nonfiction book out this November. "The timing, as you can see, is just spectacular," she wrote. Her position: Ferriss is right about the data and "one floor short on what it means."

That reading has support inside Ferriss's own comment section. One author in the same space wrote that they were experiencing exactly the same decline, and had shifted focus to long-form content and in-person retreats.

Which is roughly where Ferriss landed too. Find your 1,000 true fans. Surprise and delight them. Overdeliver again and again. "I'm tying myself to the mast of long-form," he wrote, "but maybe I'm delusional."

The uncomfortable read

If your book is a delivery system for information, a free model now delivers a personalized version of it in four seconds, and the BookScan numbers suggest buyers noticed.

If your book is a delivery system for a point of view, a story nobody else can tell, and a reason for a stranger to trust you with $15,000 of their conference budget, nothing in this data touches it.

Most coaching books were written as the first kind while being sold as the second. The market just stopped paying for the difference.