Is 5,000 a lot of books?

A quick post on book sales

When I attended an MFA program (a little over a decade ago [insert old man emoji]), I was told that for literary fiction you needed to sell 5,000 books to be considered a success. This didn’t mean you’d be rich of course. Didn’t mean you’d win a prize or get a TV adaptation. It probably didn’t even mean you’d get royalties. But 5,000 in sales was the number where a publisher would likely grant you the most important thing: a second book contract.

Earlier this week, the New York Times published an article on how COVID has impacted book sales. The tl;dr is that overall sales have actually been healthy, but backlist sales have been up while new books—especially debuts or books by lesser known authors—were in awful shape. I’ve heard the same thing from publishers and editors.

However, the big takeaway I saw on twitter was shock at a side note in the article that mentioned how few books sell more than 5,000 copies:

A lot of people, especially non-writers, were baffled at this statistic. Many even cited it as an indictment of publishing or the impossible expectations of agents who only want to rep books that sell well.

As my tweet suggests, the implication that it was COVID that caused so few books to hit the 5,000 mark is suspect. Several editors and authors like John Scalzi replied to my tweet saying this stat didn’t seem surprising. Most books don’t sell a lot of copies most year. But how many books “should” a book sell? What’s a “good” amount?

First let’s note that the NYT quote is very vague. Does “books that publishers released” include everything from celebrity memoirs to sudoku puzzle books and experimental poetry books? Is it only counting the Big 5 (soon 4) or is it counting every small press, mico press, and university press in existence? And is it counting just print books or also ebooks and audiobooks (which can be a significant percent of sales for some books)? Without knowing, it’s impossible to judge this stat.

(Also worth noting this stat is only for 2020 books in 2020, not lifetime sales. Many hardcover releases in 2020 will be getting a paperback release in 2021 and hopefully a whole nother round of sales.)

Still, part of the idea behind this newsletter was “publishing demystification” so let me try to answer the question of “is 5,000 a lot of sales?” as best I can. It’s a hard question to answer. The expectations for sales are entirely dependent on an author’s press, genre, and advance. Let’s focus on fiction. Advances for debut fiction range anywhere from 1,000 dollars (or less) to over a million. The average range for advances for a debut novel at a big press imprint is probably more like 10k to 50k. Maybe 1/10th of that range at small presses.

The basic hope for any author is to “earn out” on a book, which means your sales have covered your advance and you start earning royalties. The exact math is complicated by different royalty rates for hardcover, paperback, and ebooks, but let’s say an author makes roughly 10% of a book’s cover prize. Let’s call it $2 a book for easy math, which means you’d roughly need to sell 5k copies on a 10k advances or 25k copies on a 50k advance.

However! A book can sell much less than this and still be considered a success. One publishing friend told me very roughly selling 1/10th of your advance in copies is pretty good. If you pay someone 500k and they sell 50k books, that’s decent. Not amazing, but decent. By this metric, 5k is a good sales target for your average advance.

The reason this math works is that the publisher makes money before an author earns out. Despite what some of the ebook evangelists think, a publisher does not take anywhere near 90% of the cover prize. The distributor and the retail store take over 50% of the cover prize. But for easy math let’s say a publisher gets 40% of the cover prize, that means 50k in sales at $20 bucks a book = a million dollars. Or for 5k in sales = 100k. Not bad.

This gives you a hint at how publishing can survive with book sales being mostly below 5k. But the bigger reason publishing survives is because, well, some of the 2% selling more are selling a ton. Most books might not sell 5,000 copies, but big hits can sell hundreds of thousands or millions a year. A single Fifty Shades of Grey or Obama memoir can keep a whole house afloat for years. And remember this 2% stat was only for one year, 2020, of sales. The dream for a publisher and an author is getting that backlist title that sells well year after year after year.

What if you wanted to know what the average buzzed about fiction title sells? A few years ago, I did an analysis at Electric Literature of the 100 NYT notable books of 2014:

The BookScan sales of those books literally ranged from 1,000 to 1.5 million, with an average (mean) of just over 75,000 copies sold per book. That 75k number is pretty skewed by the existence of Anthony Doerr’s runaway literary hit, All the Light We Cannot See, which sold over 1.5 millions of copies. (The next highest book was about 270,000.) If we remove the best and worst selling books on the list, we get a mean of 46,550 copies and a median of 25,000 copies.

Countless books are published each year. But only some of them get much attention. Only a fraction are books you are probably aware of. What the above means is that if you go to a bookstore and look at the books on the front table, they’re likely going to be selling more than 5,000 copies. (Although not all will.) The above numbers are also BookScan numbers which only tracks physical book sales in bookstores. Some physical book sales aren’t covered, and ebooks and audiobooks are completely missed. So that median of 25k copies might be more like 40k+ when all is said and done.

In that article, I also did some BookScanning of award winners:

Let’s say you really hit the jackpot and are a finalist for the Pulitzer, what kind of sales would you get? Again, the range is huge. I looked up five years of nominees (from 2011 to 2015) and the range was 5,600 to over 1.5 million (yes, All the Light We Cannot See again). The mean was 250,100 and the median was 72,300. For the National Book Award, the mean was 178,600 and the median was 91,318

For comparison sake, I checked the finalists for science fiction’s prestigious Nebula awards. They ranged from 2,100 to 387,900 with a mean of 35,600 and a median of 12,300. That’s surprisingly less than the major literary awards, despite the frequently heard claim that genre fiction is more popular than literary fiction. (Although keep in mind that science fiction ebooks typically sell better as a percentage of total sales than literary fiction ebooks do.)

So again, the books you’ve heard about are likely selling more than 5k copies.

The last thing I’d say about this is that nothing about this model where a few blockbuster books keep the industry afloat is unique to publishing. The fact is that in any art form or entertainment field, most people aren’t big successes. The overall numbers for movies or music sales might be higher, but the exponential curve—where most sell very little and a few sell a ton—is the same.

Anyway, none of this is in an author’s control. All we can do is write what we are compelled to write and try not to worry about the numbers.