Macmillan

Data informs the decision, but an editor's gut picks the outlier hit. CEO Jon Yaged wants all the information so he can make a gut decision, because no algorithm has yet identified the outlier and the biggest books are often outliers. Nearly every New York publisher rejected Harry Potter.

Predictions fail both ways: slam dunks flop and modest books become huge, though much of the middle is called close to right. Macmillan still A/B tests ads and tunes metadata with data. "I haven't seen the algorithm yet that can identify the outlier."

How I Write▶ 16:57

Editors with strong track records get the benefit of the doubt. Macmillan publishes roughly one in 20 to 30 submissions and backs editors whose past picks were hits. When someone questions a buy, Yaged points to the other 100 books that editor bought that were huge.

Such editors both improve a book and recognize the next thing, which takes reading widely and being out in the world. They don't get carte blanche, only the benefit of the doubt when asking for money.

How I Write

Every pitch is reduced to a comp: which book will this behave like? Editors and sales teams pick a comparable title, check its sales in BookScan, and work backward to what that sales level can support as an advance. Agents quote comp sales numbers to push advances higher.

BookScan covers about 85% of print sales, including Amazon, but not ebooks or audio. Valuations vary widely: Macmillan might pay $100,000 for a book another house pays $500,000 for, so one rejection is one house's opinion. "This book we think it's going to behave like that book."

How I Write▶ 5:00

About 99% of books come through agents or existing relationships. Macmillan generally does not accept unsolicited submissions. Books arrive from agents, known authors, or an editor's own idea taken to an agent to find an author.

Macmillan publishes about 1,200 books a year, roughly 5% of an industry that puts out 20,000 to 25,000 titles. Emailing the front door generally doesn't work, and getting an agent is itself very hard.

How I Write▶ 10:40

Retailers and authors take most of each sale, leaving publishers thin margins. On a $30 book more than $15 goes to the retailer, so Macmillan gets under 50 cents of each cover-price dollar. Authors earn 10 to 15% of cover price on a hardback ($3 to $4.50 a copy) on top of the advance.

Printing, freight and marketing come out of what's left, and those costs plus salaries have risen over decades. A few buyers, Amazon and Barnes & Noble as the last traditional chain, control shelves and ask for bigger discounts and promotion help.

How I Write

Books over a year old make about half of Macmillan's business. Backlist titles need less marketing and use cheaper paperback printing, so they carry the company. Buying a backlist in an acquisition is almost like buying an annuity, which drives industry consolidation.

Consolidation chases scale and bargaining power, but overlapping businesses compete for attention, so "1 + 1 equals something less than two". Penguin Random House folded roughly 45 companies into one, yet its share has been steady for about a decade.

How I Write

Paperbacks are the long-tail profit engine once marketing is spent. Commercial fiction usually goes to paperback about a year after hardback, when marketing money is already spent and the book runs on word of mouth. Huge bestsellers can stay on a two-year cycle and surprise hits keep the hardback longer.

Ebooks follow print pricing: a $29 or $30 hardback pairs with a $12.99 to $14.99 ebook, which drops when the paperback arrives. Ebook margins are slightly better because production costs are limited.

How I Write

Start each campaign by naming the reader community and its connectors. Yaged begins every campaign by identifying the community a book is for, finding the evangelists inside it, and equipping them to spread the word. Imprints such as St. Martin's Press and Tor now market this way.

The trusted recommender has shifted from scholarly reviewers to bloggers to Instagram to TikTok, but the mechanism is the same. Tactics change per book; the starting question doesn't. "You have to identify the community you're trying to reach."

How I Write▶ 1:05:47

Authors should write a great book and build their own channels. Asked about an author torn between the novel and an email list and Instagram, Yaged says "you got to do both." Great writing remains the base, and owned reach adds on top.

He says the best way to sell a book hasn't changed in hundreds of years: write a damn good book. He also values authors who let editors give the unvarnished truth about jackets and edits.

How I Write▶ 11:18

Reading behavior would beat opinion, but publishers can't see it. Because a bookseller makes the sale, Macmillan can't tie an ad to a purchase the way an online click can. Early reader groups approximate the signal: a reader finishing 30% in the first day signals a hit.

Yaged wishes he could see when readers put a book down or binge it. He doesn't know whether rivals track better, since sharing would risk antitrust trouble, and guesses Macmillan is in the middle.

How I Write

Macmillan launched Swoon Reads after seeing readers wanting books it didn't publish. Running the children's division, Yaged saw demand the house was missing and started Swoon Reads, which ran a few years with some great successes but didn't work out exactly as hoped.

He thinks New York clustering creates energy but a bubble that misses trends elsewhere, one reason self-publishing grew so big, and says watching self-publishing is a good way to see what comes next. "We don't know what we don't know or we're missing trends."

How I Write▶ 15:44

AI demand forecasting let Macmillan print about 1 million fewer copies. Macmillan estimates it printed about 1 million fewer copies in a year by using AI to forecast demand. Yaged also sees AI drafting tip sheets and refreshing metadata keywords, with humans editing.

Tip sheets, the book summary, author background and specs, are a chore junior editors hate. He notes copyright doesn't protect AI-created work, so a book written mostly by AI can't be owned. "Creativity isn't an algorithm. It's just not."

How I Write▶ 53:09

Private, long-term ownership shields creative bets from quarterly pressure. Owned by Holtzbrinck, under Stefan von Holtzbrinck and his niece Christiane Schuler, Macmillan isn't public, which Yaged calls a gift because creativity doesn't run quarter by quarter. His main job is building a place where people aren't afraid to make the wrong call.

He says copycat covers come from chasing what works, while the books that look different often set the next trend. "If people aren't scared of a wrong call, we're going to win in the long run."

How I Write