How a book becomes an audiobook, and what it asks of you

An audiobook gets made in one of three ways: you read it, a professional narrator reads it, or a computer voice reads it. What separates them is how much of your time it takes, what it costs you, and who takes a slice of the income afterwards.

The choice looks technical, but it is really a business decision, and you make it once per book.

What an audiobook is, from the shop's point of view

To the shop, an audiobook is a separate product from the written book. It gets its own page, its own price and its own audience — people who listen while driving or walking and would never have bought the same book in print.

Which means the recording is not a version of your book. It is another book you are publishing, with its own work attached.

The three routes, in plain language

Read it yourself. You need a quiet room, a decent microphone and patience. A two-hundred-page book is several days of recording, plus editing afterwards. It works beautifully for personal non-fiction, memoir and how-to books, where the author's own voice is part of the value. In fiction with a lot of dialogue it is harder than it sounds.

Hire a narrator. On ACX, Amazon's platform for producing audiobooks, there are two routes: either you pay the narrator per finished recorded hour and then keep a higher royalty rate for yourself, or you pay nothing up front and the narrator takes a share of the income over time. Amazon estimates that production takes six to twelve weeks. Worth knowing: signing up for ACX is currently open only to residents of the United States, the United Kingdom, Canada and Ireland.

A computer voice. KDP (Amazon's self-publishing platform) currently has a trial feature that produces an audiobook in a synthetic voice at no cost. You can choose from eighty voices and even give each chapter a different one, and listen to the result before you publish. The royalty rate on this route is lower than on the human one, and it is open by invitation only and only in the American store.

Where language enters the picture

This is the part that catches Israeli authors out. Amazon's computer voices exist in American and British English, in Spanish, in French and in Italian. Hebrew is not on the list.

In practice, if you want an audiobook that sells in the big shops, it will almost always be in a foreign language. And that puts the order of operations back in front of you: a finished, translated text first, and only then a recording.

The reason is simple. A narrator, flesh and blood or synthetic, reads a text out loud. If the text is still going to change, the recording gets thrown away. And there is this too: we have written before about the order between digital, printed and narrated, and the recording is almost always last.

What to check before you start

Three questions, before you pick up the phone to a studio.

The first: is my book even suited to being listened to? A book with tables, charts or references to page numbers does not work in the ear.

The second: do I have a final text? If the book is still going through editing or translation, this is not the moment.

The third: how much am I willing to pay up front? There is no right answer, but the answer decides whether you are on the pay-the-narrator route or the share-the-income route.

If you are at an earlier stage and still weighing up which language to publish in at all, that decision comes before everything else, and we have gathered it in how to choose a language. The rest of the questions that come up at this stage are in the frequently asked questions.