What do Amazon’s book stats really mean? Nobody knows for sure

Note: Media observer Dan Bloom wonders, like a lot of people, just what Amazon’s popular and must-see book stats really mean. His opinion here reflects his own personal hunch about how things operate in the shop that Jeff Bezos runs. Comments welcome below, pro and con.

AmazonThe next time you read in a press release or newspaper that a certain book “has been propelled to the top 100 rankings among paid Kindle titles on Amazon.com,” think again. What does that really mean? And the next time you hear that a certain book “has been propelled to  the top 100 rankings among regular print book titles on Amazon.com,” think again, too. Do the stats  figures mean actual sales, pre-orders or what?

Nobody outside Amazon headquarters seems to know. All Amazon will say on the record is that the famous/infamous book ranking “stats” are from actual sales and they are updated every hour. So one must assume that this is the case—at least until someone who understands the code comes forward and dishes. For now, it’s a true puzzle.

When news broke worldwide a few years ago about a controversial self-published e-book titled “A Peodophile’s Guide to Love and Pleasure,” triggering a ground swell of protest on social media networks from Twitter to Facebook, the vanity press title suddenly found itself being ”propelled” to the top 100 rankings among paid Kindle titles on Amazon, according to news reports. But did those short-lived ranking stats mean actual sales had taken place, or that pre-orders had been temporarily filed, or what?

People want to know. Me, too. So far, we’re all in the dark.

Ask any publisher or book editor in New York or London. Amazon’s book stats are not easy to decode, their true meaning almost zen-like in their inscrutability.

Maybe I can find someone who used to work at Amazon, or who coded for Amazon, who might be willing to talk. My email door is open. Bezos, you can email me, too. Dish. My door is always open.

In the case of the peodophile guide, some news reports were saying that “less than 24 hours before the book was taken down by Amazon, the virtually unknown digital book ranked well north of 157,000″ on the online book ordering site. Suddenly,
the news reports reported, the guide was up in the top 100.

Top 100 of what? Top 100 of sales, all paid for and delivered to happy customers? Or top 100 of pre-ordering interest with such pre-orders routinely never actually purchased?

Did the controversy actually spur sales of the guide, which was being sold sold only on the Kindle platform? Perhaps. Amazon stats don’t lie.

It’s strange, because the author of the book told an Associated Press reporter in Denver that he had only sold one copy of the book.

David Carnoy, who routinely follows the e-book industry as an editor at CNET.com and is the author of a novel titled ”Knife Music”, believes the Amazon stats are about sales. He recently told CNBC reporter Bertha Coombs: “That’s the disturbing part about [the peodophile guide media controversy], that it led to actual sales.”

Carnoy told me in a subsequent email:

“You know why [I believe] it was sales? Because the book cost something like $4. So people were buying the thing. When you have a few hundred thousand curious people looking at the book, it’s not hard to believe that a couple of hundred would buy it, is it? At that price?”

Point well taken.

A midlist novelist in New York whose books are sold on Amazon and who therefore prefers to remain anonymous for this story, told me in an email:

“Like most people, I always believed the Amazon book rankings were based on sales. I don’t think Amazon ever publicly posted its methodology, and I only get an occasional royalty statement that doesn’t reflect the daily gyrations of the statistics. However, I have noticed that when a friend or relative buys just a few copies of one of my books, the ranking jumps enormously. Most authors could be a gigantic tie for last place.”

When I recently asked an American technology beat writer what he knew about the Amazon stats, he replied: “I actually don’t know what the truth is about the sales stats. People in the publishing business here in New York tend to say they are meaningless, but I’ve never seen proof.”

He went on:

“However, the one aspect of Amazon’s book stats that could be debunked with readily available evidence is the way they claim e-book sales now exceed sales of old-fashioned books. If you look at their press releases, they do this in a way that suggests it’s an industry-wide fact. But the claim is very deceptive. It’s based on the number of Kindle sales versus hardcover sales, and only on Amazon. That leaves out paperbacks, which are still the lion’s share of book sales (on Amazon and in the world at large). And, just as important, it leaves out hardcover sales in the rest of the book industry, which, if tallied together with the Amazon hardcover sales, reduces the Kindle number to a tiny percentage of the total. It’s a real scam, and the media tend to buy into it, as part of the ‘paperless society’ trope—now over half a century old, but somehow still kicking.”

And finally, when I spoke with a veteran publisher in Manhattan, who has been in the business for more than 40 years, he told me:

“As a publisher, I know that the stats are not ‘the real thing’ and this awareness on my part is based on following the rankings of our own books and seeing how the rankings coordinated with the following weekly Amazon orders. Since Amazon orders on a regular once a week basis, it’s pretty easy to follow. Basically, the stats thing is a ploy used by many publishers and self-published authors to push their rankings up and try to get a better distribution. Like most things in life it’s all about perception, not reality. Welcome to the world of publishing.”

So now, Jeff Bezos, talk to me?

Dan Bloom is a freelance writer based in Taiwan who never checks his Amazon stats.

7 Comments on What do Amazon’s book stats really mean? Nobody knows for sure

  1. Amazon rankings are based on real sales.

    Since they are updated hourly, a large (but short-lived) surge in popularity — like if you have a book launch where many people buy it on day 1 — can propel your book fairly high up the bestseller list.

    If this helps, I sold about 100 copies of my latest book (The Twiller) on the day it was released (most of those sales were within a few hours), and that jumped it into the Top 200 overall paid sales on Amazon for a short time (it was also in the Top 5 in Humor).

    So, the rankings are real, if you understand what they measure: an hour-by-hour snapshot of sales activity. Of course, you could argue that weekly or monthly sales rankings would be more useful, but this way Amazon gets to show more movement and introduce more books to people through the bestseller lists (instead of seeing the same books week after week).

    At one point Amazon did tout that e-books outsold hardcovers (yes, based on Amazon sales only, although Amazon is the world’s #1 bookseller). Then, over 2 years ago, they announced that they sold more e-books than all print books (paperback and hardcover) combined. They even excluded free e-books and included print books without e-book counterparts.

    http://www.davidderrico.com/e-books-outselling-print-books-on-amazon/

  2. Dan:
    There is a great ‘dissection’ of all the various Amazon stats types in David Gaughran’s “Lets Get Visible” ebook.
    The key one is a book or ebook’s “Best Seller Rank” a number that equates with it’s overall sales unit ranking on Amazon. David’s book actually has a chart that let’s you determine the approx. daily sales for a book based on this ranking.

    Other stats include: Top 100 in various categories and category>sub-category, category>sub>sub. etc. The Top 100 is further broken into Paid, and Free. There are also Top 100 ‘New & Hot’ (best sellers released in last 30 days only), ‘Movers & Shakers’ (biggest upward movements up the charts), and Popularity (measures both sales & Free downloads (which only count as 1/10th of sale) over last 30 days approx.)

  3. DD, great comment post and where you said “So, the rankings are real, if you understand what they measure: an hour-by-hour snapshot of sales activity.” Someone in the mainstream media like at the NYT should follow up on your important insight: the stats merely measure hour by hour sales snapshots, and David, many of those “sales” are merely pre-orders and can be returned for refunds later, so the stats while REAL and also UNREAL. But good point and i hope your insights reach the MSM. Most big reporters are afraid to report this because they will lose their “in” with Bezos. For real. He runs the world now, almost.

  4. Stevh, thanks for the heads up on that. Why hasn’t the MSM or even PW published any of this?

  5. Dan:
    Your welcome! I think its because most of the MSM is clueless when it comes to the in/outs realities of ebook self pubbing. The data was all deduced, as Amazon isn’t really telling anyone, via hard-core self pub authors that crowd-sourced a pool of data points to come up with the approximations for daily sales stats and how Amazon ranks/weights various factors for other stats lists.
    The PW (publishing world..?) would rather not know, most are still hiding in their glass houses trying not to acknowledge the sound of guns on the horizon.

    Additionally, Amazon changes the ‘secret sauce’ algorithms sorta regularly to keep things a’ changing and prevent gaming of the system (not unlike how Google modifies their site ranking methods for search results ever year or so).

  6. again, thanks Stevh, good insights. by PW i meant Publishers Weekly, the leading print book trade mag in NYC.

  7. A friend of mind who is a longtime publisher of ebooks and POD books
    for over 100 writers tells me today in response to the above
    commentary:

    Dear Dan,

    ”Amazon’s rankings are based on sales over time, according to what I know.
    Playing with some numbers…

    According to the latest publisher’s weekly regarding trade paperback
    (print) sales, sales were at $5,800,000 for the week. If Amazon has 30% of
    the market, which some have surmised, that puts their total at $1,740,000.
    Amazon is a 24 hour operation, reporting their rankings on an hourly
    basis. There are 168 hours in a week. So sales at Amazon for trade
    paperback average $72,500 per hour. If the average trade paperback costs
    $12.95, you are looking at 5598 units per hour. In the grand scheme of
    things, that is not a lot of units — and of course — peak hours in the US
    are much higher. But, my point is, given the number of titles vying for
    this piece of the pie, the sale of only a couple units would swing the
    ranking for the title quite a bit. In fact, selling 100 of a title in an
    hour would likely put it near the top of the pile. Let’s say there are
    1,000,000 trade paperback titles that have any sales of the 5,000,000 out
    there. Applying the 80/20 rule, the top 200,000 represent 80% of the unit
    sales or 4478 units in total — (this comes to about half a book a day per
    title). Now, use the 80/20 rule again — top 40,000 titles have 3582 units
    — 8,000 have 2866 — 1600 have 2293 — 320 have 1834 — 64 have 1467
    — 12 have 1174 — 2.4 have 939.

    So, to make the top 100, of 1,000,000 trade paperback titles on Amazon, you
    only need to average sales of 23 sales per hour. Being #1 overall is
    probably closer to 500 per hour.

    Your chance of a high ranking during off-peak hours is greatly increased.

    Your chance of a high category or sub category ranking is even easier to
    achieve given there are fewer titles as you carve out the segments.

    My guess is that Amazon uses some sort of moving average, weighting sales
    for the week, day, quarter day and hour in different proportions. This
    would mean, a book that had all of its sales within one hour and achieved a
    high ranking, would gradually settle back down the list. I have witnessed
    such behavior. I just haven’t figured out the algorithm…

    The Kindle market is even more interesting. We have seen a leveling
    off in sales of eBooks — a strong fade on the Nook platform for sure.
    Because most of the eBook market is fiction, it is not a true “read”
    of the whole market. There is a lot of junk, too. I think, by the
    time you sort the wheat from the chaff, you will find a much smaller
    number of meaningful sales on ebook platforms then expected.”

Leave a comment

Your email address will not be published.

*



wordpress analytics