The Long Tail

October 10, 2004

Objection #1

"57% of Amazon's book sales are of books not available in stores"? That sounds high.

Response: This is based on the work of Erik Brynjolfsson, an MIT economist, and his team at the Sloan School. They fit a standard Pareto power law distribution to Amazon sales rank and summed the total above 40,000 (the amount of books stocked by the average bookstore). As you'll see in the paper above they did a really interesting job of forensic economics, buying multiple copies of books deep into the tail to see how it changed the sales rank and then using those figures to extract absolute sales figures from the rank numbers.

However, their research assumes two things: first, that the research they based their work on measured sales of the head (bestsellers) accurately; and second, that the Amazon sales rank algorithm is accurate. I'm told that in fact neither are quite right. Most analyses (such as this) tend to undercount sales at the very top of the list, and Amazon's calculations are still imperfect. This may account for some overcounting of the tail.

UPDATE:

I've now spoken to Jeff Bezos (and others) about this. He doesn't have a hard figure for the percentage of sales of products not available offline, but reckons that it's closer to 25-30%. That would put it in line with Netflix's and Rhapsody's figures. This is still a large fraction of their business, and the fastest growing segment. I've updated the version of my article on ChangeThis accordingly.

Posted at 10:07 PM in FAQ | Permalink

Comments

I don't doubt the 57% for a couple of reasons. One being that I go to bookstores fairly often (once a week or so)so if something is available there I tend to buy it there. This accounts for most of my buying of top-selling books. Second, I use Amazon specifically to find things that are rare or out-of-print. For all of my Amazon book purchases I know that none of them were available locally. I don't mind paying shipping for something that I can't just go out and buy, but I don't want to pay shipping for something that is in a store that I will go to eventually anyway.

Posted by: Jay | December 21, 2004 at 11:43 AM

OK, ~25% of Amazon/Netflix content is not stocked in a physical store NOW. How has this been changing and how will it change?

Physical stores rotate content out, on-line stores don't need to. Once it becomes available at Netflix, a movie can be available _forever_ at Netflix.

I see a strong case that, as content leaves physical shelves, on-line sales/rentals can continue to be a viable revenue stream indefinitely.

Books already have second-tier publishers who pick up the rights to publish out-of-print books and we are now seeing the emergence of on-demand publishing for even smaller markets - individual books.

Another comment:
Used book stores are becoming a thing of the past. I can browse more titles more efficiently at Amazon.

Posted by: Anonymous | December 21, 2004 at 11:58 AM

I wish you the best with your book.

One thing I wonder is: how are you going to handle the Google books phenomenon?

Posted by: praktike | December 21, 2004 at 12:10 PM

Well, this looks like a very interesting subject for a book, there is a LOT here.

Thrity percent seems about right, given my own shopping behavior. Netflix, for example, has Kurosawa's Yojimbo (and it's in my queue right now), but I doubt my local video store has it. On the other hand, I still rent plenty of mainstream movies too.

Looking back to 2003 (for a complete year), I see that I ordered 21 books from Amazon (yikes!), with about seven of them not likely to be in an average bookstore, in my opinion.

But the interesting thing is that some of the books I order are WAY outside of the mainstream. For example, I've long been interested in the history of the cult of the UFO and that led me to books like _Looking for Orthon_, which is a biography of George Adamski, one of the first "contactees" and root source of many crazy UFO ideas. The book has an Amazon sales rank of 718,232!

Posted by: rob | December 21, 2004 at 12:28 PM

Well, I run two websites that probably qualify as out on the far end of the tail:

http://www.hafernik.com
http://arabiannightsbooks.com

The first one is a family genealogy site. My parents had spent 25 years collecting the data and I thought it would be fun to make it into a web site. The second is about a hobby of mine, collecting illustrations from the zillions of editions of the Arabian Nights that are out there.

Both of these are about as far out on the tail as anyone would like. But I've noticed something funny about being out on the tail: your traffic is still extremely even. I get 9, 10 or 11 visits to the hafernik.com site every day. It rarely varies and when it does it's usually because *I* was checking on the site.

I don't know if there is math to be done here, but both of my "tail" sites show this behavior.

Posted by: Rob Hafernik | December 21, 2004 at 12:43 PM

-----
He and his team fit a standard Pareto power law distribution to Amazon sales rank and summed the total above 40,000 (the amount of books stocked by the average bookstore).
-----

I don't understand how this translates to "not available in stores," unless you assume that all bookstores carry the same 40,000 books, which is demonstrably not the case.

For example, I just looked up the book I bought (new) in a store yesterday (David Garnett, "Lady Into Fox"). Its Amazon rank is 160,354. Trying for something further down the tail, I checked "The Collected Strange Stories of Robert Aickman" (which I bought at my local science fiction specialty store, and would assume is available at others). Its rank was 1,845,023!

Maybe you mean "not available in crappy chain stores." Which is worth noting, but not at all the same thing.

Posted by: Tim Walters | December 21, 2004 at 02:31 PM

"Not available in crappy chain stores" is a big deal, though. First off, while plenty of things are available at specialty stores, and things differ by stores, people are still dependent on things being available in their particular area, of having such specialty stores near them.

In other words, it's a version of the old Sears Catalog effect. You don't have to live in a large city to have access to esoteric items anymore. It's extending the range of retail past physical space, into aggregating large populations.

He doesn't mean "crappy chain stores," either, at least not in the sense that I grew up with. The modern Borders and Barnes & Noble stores are a huge improvement over the tiny, unspecialized bookstores (often chains) that I saw growing up. They have an impressive breadth of titles-- but aren't going to be able to match something like Amazon. Nor can local specialty shops unconnected to the Internet match Amazon or any specialty retailer that can aggegrate sales.

However, I have a more important caveat. Perhaps there are lots of users drawn to Amazon, Netflix, and other sites just to buy the rare items. Perhaps they buy the big hits at their local bookstores, record dealers, Blockbuster, whatever, for the immediacy, but then get their niche items online. I know a lot of people with this behavior. It would naturally exaggerate the size of the tail, at least as a percentage of the whole industry. There may always be people who prefer brick-and-mortar for purchasing hits. At least online exists to capture the tail, though, and certainly makes it possible.

Posted by: John Thacker | December 21, 2004 at 02:52 PM

>"Not available in crappy chain stores" is a big deal, though.

Oh, definitely. I just think it's important to know what we're really talking about when we throw around these numbers.

>The modern Borders and Barnes & Noble stores are a huge improvement over the tiny, unspecialized bookstores (often chains) that I saw growing up.

That's true--I remember the B. Dalton's of my youth and shudder. But that also means that Borders and Barnes & Noble provided a long-tail effect of their own when they started becoming popular back in the Eighties, as did Beverages & More, Whole Foods and IKEA in their various sectors. It would be interesting to track the prehistory of the "long tail."

Posted by: Tim Walters | December 21, 2004 at 04:55 PM

I'd be interested in similar figures for music.

Posted by: Julian Bond | December 22, 2004 at 02:32 AM

Barnes and Noble online doesn't seem to have nearly as long a "tail"
as Amazon; which is unfortunate for me, given that Amazon gave heavily to the Republicans (please ask Jeff Bezos why!!) and BN gave to the Democrats. (I'd rather these big corporations would stay out of politics altogether, but as long as they using their profits that way, we have to be aware of this when we shop).

Posted by: Rudy | December 22, 2004 at 12:24 PM

I can't wait until the book comes out. I'd love to see what you discover about the mathematics of the long tail. I did a quick survey of a week's worth of messages from a yahoo group I frequent and found that 14% of the posts were from people who posted 5 times or less in the week. I created a pie chart and the sequence resembled a seashell as the numbers went down from the person who posted the most amount of messages (11%) to the one who posted the least. I'd be curious to know what the long tail mathematics are for social networks.

Posted by: Yves | December 30, 2004 at 04:12 PM

I don't know if there is math to be done here, but both of my "tail" sites show this behavior

--------------
http://www.bromastelefono.com/

Posted by: garces | March 01, 2005 at 01:04 PM

Interesting notion about the "Prehistory of the long tail". I think the Wired article talked about limited inventories as being a product of early mass production. If we look back further to the days of custom-made items then I would not be surprised if there are some areas (such as furniture) where the variety available then was greater than what we see today.

Posted by: Phil | March 11, 2005 at 07:56 AM

---------------
Both of these are about as far out on the tail as anyone would like. But I've noticed something funny about being out on the tail: your traffic is still extremely even. I get 9, 10 or 11 visits to the hafernik.com site every day.
---------------

Er... Web crawlers?

Posted by: konbs | March 13, 2005 at 01:30 PM

The link Chris gives to my Amazon sales rank analysis above is still active, but the whole system changed just days after he originally made this post. Amazon completely revamped their sales ranking system in October 2004, right after he wrote the articles, becoming stable right at the beginning of November, when I redid the graph linked above. One effect of Amazon's change was to render all previous rank history meaningless, and all of the academic studies done to that point, including several that referred to my old article, became obsolete as reference material. Whatever conclusions they'd drawn to that point may have been valid, but the new system almost entirely ignores historical sales, as I explained in my December update. For a long time watcher like myself, the orders of magnitudes changes in old "permanent" ranks are what gave away the lack of historical weight in the new system.

I still so quite a bit of Amazon watching in my self publishing blog, and this year, Amazon's international media sales may actually surpass their North American sales. As both a trade author and a self publisher, I can say for sure that you need to stay on the short end of the tail at Amazon to actually make a living at it:-)

Posted by: Morris Rosenthal | July 29, 2005 at 01:31 PM

As with any stat, the definition is paramount. Back in '99, amazon was selling several hundreds of thousands of unique book titles each month and over 80% of those titles did not appear on the sales list the following month. So, if a physical megastore only carries 25-40k unique titles, it isn't carrying 90% of the titles that amazon sells in a month.

Posted by: Miles | October 30, 2005 at 08:16 AM

houyuping 07年09月11日

google排名
 google排名
 wow gold
wow gold
wow power leveling
wow power leveling
wow powerleveling
wow powerleveling
world of warcraft powerleveling
world of warcraft powerleveling
world of warcraft powerleveling
world of warcraft power leveling
powerleveling
powerleveling
wow powerleveling
wow powerleveling
wow power leveling
wow power leveling
power leveling
power leveling
wow power level
wow power level
china tour
china tour
搬家
 搬家公司
 搬家公司
 北京搬家公司
 北京搬家公司
 翻译公司
 翻译公司
 上海翻译公司
 上海翻译公司