My new buddy and client over at iobyte, Dan Lubart, inspired a post last week about Amazon’s new Sunshine promotion because he documented its impact on their bestseller list.
Since then he’s put up two new posts that are only worth reading if you care at all about the effect of price on today’s ebook market. I think that includes most of us.
There is legitimate debate about how bestseller lists should be organized in the ebook age. I pointed out last winter that it really wasn’t tenable for ebook lists to keep score on unit sales only when the price range extended from 99 cents to $19.99. And, in fact, that kind of price variation has already led to clever authors and publishers gaming the system: lowering prices to get on bestseller lists and then raising them to capitalize on the additional discovery that takes place once they’re on. My suggestion was that the grading be on “price times units”.
Those who like that idea quickly see that this is similar to how movies work; they report the box office receipts, not the number of tickets sold. Those who don’t like the idea say that what a bestseller list is supposed to communicate is what books are being read by the most people.
Dan’s two most recent posts illuminate how this question plays out in the real marketplace.
In the first, he appears to have discovered that Barnes & Noble may have arbitrarily decided that an ebook priced at $2.99 or less won’t be placed in the top 125 of their bestsellers. This conclusion is based on convincing circumstantial evidence. What Dan first noticed was that only one book in the Sunshine promotion, and then two, had hit the Nook bestseller list. Watching those books, he saw one, “My Horizontal Life”, fall from rank number 1 to rank number 127 in one day.
That’s beyond a statistical anomaly. That’s damn near an impossibility, unless the rules of the game were changed.
So Dan checked further and found that all the inexpensive ebooks had dropped below rank 125, but that they were dominant from 126 to 200.
Since I’m the guy who pointed out a few months ago that lumping 99 cent ebooks with $19.99 ebooks was mixing apples and broccoli, I think it is fine that Barnes & Noble has taken this additional curatorial step. The New York Times makes it clear that its ebook bestseller list does not “actively track” a variety of titles including those that are self-published. I think it is fair to say that neither has a “transparent” methodology. But putting their thumb on the scale might be delivering a more useful result for the users of their lists.
I found Dan’s next post even more revealing about the nature of the ebook marketplace, particularly at Amazon.
What he showed, through analysis of the price of the titles and their rankings, was that the disruptive effect on the bestseller list of the Sunshine promotion was very brief. It lasted about a week. Seeing this recalled a story my Dad once told me and therein lies the explanation.
In the 1950s, Doubleday tried an experiment of putting books into supermarkets. What they found, repeatedly, was that the books sold well the first week and then sales collapsed.
The explanation for this was very simple. Bookstores see their customers — even most of their best customers — relatively infrequently. But supermarkets see their customers weekly or even more often. So a display of books is quickly seen by just about all who might be interested. They either buy something or they don’t. But after a week, everybody who visits that store has seen it and, unless the choice of books the next time they pass it is very obviously different, they will have no need to shop from it again. (By the way, this is a reason to create automatic title rotation by not assigning one title per pocket, but that’s a different subject than we’re discussing in this post.)
What Dan’s Amazon data would suggest is that the low-priced shopping cohort is a herd that responds quickly. When Amazon announced their Sunshine promotion, the most avid low-price buyers shopped it immediately and made their purchases. That created the spike in low-priced bestsellers which we acknowledged in our prior post.
But in the second week, with the same selection of books in the Sunshine promotion, that effect virtually disappeared. The low-price shoppers had done their purchasing from that selection. The normal buying patterns on the site reasserted themselves on the list. What one can see from Dan’s data is that the highest-priced band of ebooks took a real hit in ranking during the first week of the Sunshine promotion but in the second week the impact was much reduced and the lowest-priced books were apparently taking share only from the next band up. The highest-priced band had totally regained its pre-promotion share of the list.
Dan always reminds me that “ranking” and “sales” are not the same thing. It is possible that the Sunshine promotion elevated the spotlighted inexpensive books without reducing the sales of the books knocked down or off the list. In fact, I am one who believes that the purchasers of low-priced books are really a different group of people, for the most part, than those who buy the higher-priced books.
But since those bestsellers definitely lost the discoverability created by their presence on or high up on the list, it would open up a whole new set of questions if they got the same sales without what most of us assume is the important lift to discovery provided by bestseller ranking.
The ebook world is rapidly shifting and changing. With the pool of ebook consumers continuing to grow quickly, the buying patterns are bound to be temporary. The next batch of ebook customers might be more price-sensitive or less; they might respond to a price promotion as quickly as the Kindle customers did to Sunshine in the future or they might get slower. But what Dan Lubart is making clear is that the impact of price promotion is visible, if you have the right tools to look.