About a week and a half ago I reported on and expressed doubt about a new estimate from Nielsen Pubtrack which said that 6% fewer ebooks were sold in the US in 2014. At that time I had no evidence to prove that the Pubtrack numbers were off, but since then I have been digging into the background.
Today I can report that the Pubtrack estimate of 223 million ebooks sold in the US in 2014 (as reported by The Bookseller) was not just bad, it was completely erroneous. I have two sources which say that the US ebook market is at least twice as large as Pubtrack thinks it is.
My first source is the pseudonymous Data Guy (the one behind the Author Earnings Report). His calculations suggested that around 513 million ebooks were sold in the US ebook market last year.
And for those of you who want an official figure from a named source, my second source is the Association of American Publishers. They just released their annual estimates of the US book market, and they say that 510 million ebooks were sold in the US last year (that volume also grew by a fraction of a percent).
Yes, the AAP says that the US ebook market is over twice as large as Nielsen had claimed.
It turns out that Nielsen widely over-estimated its ability to track the US ebook market, and I can tell you why.
While Nielsen reports its Bookscan estimates of the print book market based on data from several dozen retailers, the Pubtrack ebook market estimates come from unit sales data supplied by 30 US publishers.
That is not a bad information source, if you know what you're doing, but apparently Nielsen does not. Here's what they told me about Pubtrack when I followed up:
PubTrack Digital is a publisher aggregated eBook sales tool in which publishers are the data suppliers who send their complete and confirmed eBook unit sales to Nielsen every month. This eBook measurement service gathers point-of-sale data (indirectly from 40+ retailers) through a panel of over thirty US publishers - representing about 85% of the nation’s eBook sales.
Yes, Nielsen thinks the 30 publishers it tracks represent 85% of the market, when in reality they represent around half that number.
This comes as no surprise to anyone who has read even one Author Earnings Report, and I was tempted to make that point when I first learned about Nielsen's error last week. But I held the story because I knew that the AAP would be releasing new data this week which would either confirm or disprove Nielsen's figures.
And I am glad I waited, because this shows that Nielsen's understanding of the US ebook market is about as accurate as if they had thrown darts at numbers stuck on the wall.
I reached out to Nielsen for a comment on this story. I will update this story should I receive a response.