Amazon Kindle DX priced at $489, product link

[Dahak]

Originally Posted by sirbruce

http://www.engadget.com/2009/05/06/l...-event-in-nyc/

"We've added 45,000 books in the last three months. Kindle sales are now 35% of books where we have Kindle editions -- that's coincident with the launch of Kindle 2."

That's the first time I've seen something claiming to be an actual quote from a speaker. All of the hits I've found on the subject so far have been people interpreting and reinterpreting the graph info shown during the presentation.

Originally Posted by sirbruce

I won't say they weren't counting free Kindle versions of books; that's something I would do.

Yes, but would you do that and not say that you were padding your 'sales figures' with the give-aways and next-to-free public domain books?

If you want the sales figures to be more realistic, to have any meaning, you should also qualify that by also presenting the values only on recently published books... say, within the last few years. Doing so for books that have been published since the introduction of the Kindle would actually be the best idea - it would strip away most of the freebie and public domain texts and also show how the new format fares with its dramatic discounts versus the hardcover paper version.

Originally Posted by sirbruce

Since you're a doubting Thomas, it doesn't matter the exact phrasing; you're not going to believe it unless you corner Jeff personally, pin him down, and even then might believe he's lying.

Of course it matters what the exact phrasing is. By knowing what he's actually said, you have some bounds upon what he isn't saying.

Originally Posted by sirbruce

That's fine, but only if you don't trust figures for any other electronic device like cell phones, computers, PDAs, etc. that are regularly bandied about. Can these industry estimates be wrong? Sure. Are they likely to be wrong by more than 50%? No.

Sony has sold about 400,000, and there's no data that shows any other ereader has sold any large number of units in the US during this time. Amazon does indeed outpace everyone else's sales.

Take these two paragraphs together: with the numbers you've claimed for Kindle 'sales' versus Sony sales, then Amazon isn't only beating Sony, but grinding them into a fine powder and sprinkling it on their breakfast cereal.

Allowing for your limit to 50% error, then you can claim that Sony would be tied or beating Amazon by a slim margin. This despite the fact that Amazon has been hyping the thing ad-nauseum and Sony, well... hasn't.

(Assuming, of course, that the Sony numbers are any firmer than the Amazon 'numbers'.)

Originally Posted by sirbruce

It was on Oprah, you know.

Oh, please... let's not go there.

Originally Posted by sirbruce

As I suspected when I began writing this response.

When someone flatly refuses to tell you what numbers he's using to base his claims of owning the market, I tend to be doubtful when he breaks a two-year news blackout.

"The Kindle is a hit!" Why? Because Amazon says so!

That's pretty much what most arguments on the subject boil down to.

[Dahak]

Originally Posted by Harmon

It seems to me that logically, it has to be units, since ebooks are priced lower than pbooks, so the percentage for units will always be higher than the percentage for dollars, and Amazon - or any seller - would always use the higher percentage.

I suspect that you're right. And, of course, your statement on the usefulness of the higher percentage as a means of advertising is well taken.

Originally Posted by Harmon

My guess is that it's a measurement of sales among Kindle owners. The reason I think this is the same as above: Amazon will always use the higher percentage, and measuring sales limited to Kindle owners will always give a higher percentage than measuring across all buyers.

That's quite true, but that's certainly not how the 35% figure is being presented.

I could actually buy that figure if they qualify it within the bounds of Kindle account holders.

The problem with admitting that (if only Kindle accounts was the actual case) would be that you could then twist the figure into saying: "Look, even with Kindle versions available, Kindle owners still prefer paper two-to-one!"


Thanks for the reply. Your remarks on the 'puffery-factor' are quite true. Of course they'll skew the numbers in the most favorable light to prove their wonderfulness.

I just like to have enough detail to know what is puffery and what is a valid claim.

The Amazon claims are woefully lacking in detail.


This brings to mind the original Kindle debut. They sold out their initial lot in a week or so. Then they use that as proof of the high demand for the reader.

If their first lot was ten or twenty thousand units, that would have impressed the heck out of me. I don't think Nuvomedia sold ten thousand Rocket eBooks (I still use mine on a daily basis and my serial number is something like 7300) over the life of the unit.

If their initial lot was 500 or a thousand units, that would thoroughly underwhelm me, since there are easily enough gadget-geek first-buyers out there to exhaust that-sized lot. Then amazon can do what it did: claim the quick sellout shows huge pent up demand for their reader.

As with all-things-Kindle, we have nothing to base an argument on.

[wallcraft]

Originally Posted by Dahak

If you want the sales figures to be more realistic, to have any meaning, you should also qualify that by also presenting the values only on recently published books... say, within the last few years. Doing so for books that have been published since the introduction of the Kindle would actually be the best idea - it would strip away most of the freebie and public domain texts and also show how the new format fares with its dramatic discounts versus the hardcover paper version.

The public domain ones are removed unless there is also a related paper version. One of the big sellers on the Kindle have been backlists (of a series or an author), and these may be contributing to the 35% but I don't see this as misleading. This is in part an effect that is often seen when the media type shifts (people re-purchase old favorites), which it has not done in the book industry since forever.

[Dahak]

Originally Posted by wallcraft

The public domain ones are removed unless there is also a related paper version. One of the big sellers on the Kindle have been backlists (of a series or an author), and these may be contributing to the 35% but I don't see this as misleading. This is in part an effect that is often seen when the media type shifts (people re-purchase old favorites), which it has not done in the book industry since forever.

Quite true... and this will artificially inflate any values (using 'numbers' here would imply greater precision than is available) being bandied about.

Avoiding that was one of the reasons I suggested looking at the numbers with new releases. Even with the unfair advantage of having a 60%+ discount on a new hardback, it would be more indicative of on-going sales, not a python-lump of backlist sales as the initial wave of Kindle II users get their hands on their readers.

[sirbruce]

The only thing more I can offer is, as someone who has worked with companies to try to get numbers out of them before, is that some companies really are reluctant to provide numbers that would put them in a positive light. Some of them value secrecy. Amazon in particular has been notoriously private historically about its internal business metrics.

[Harmon]

Originally Posted by sirbruce

The only thing more I can offer is, as someone who has worked with companies to try to get numbers out of them before, is that some companies really are reluctant to provide numbers that would put them in a positive light. Some of them value secrecy. Amazon in particular has been notoriously private historically about its internal business metrics.

There is this to be said - information is power, and by controlling or concealing information, Amazon is denying power to its competitors. Apple does the same thing. This applies to positive information as well as negative information, I think.

[sirbruce]

Originally Posted by Harmon

There is this to be said - information is power, and by controlling or concealing information, Amazon is denying power to its competitors. Apple does the same thing. This applies to positive information as well as negative information, I think.

Yes, but I'm a big believer that transparency is better for the industry as a whole than secrecy. It's a Prisoner's Dilemma sort-of thing... if everyone plays openly, everyone benefits, but if one company doesn't, it ruins the game for everyone, particularly if they're a large company.

[sirbruce]

Dahak,

Actually, I found a link the first ten minutes of the video of the announcement, so you can hear the exact quote yourself:

http://www.youtube.com/watch?v=dPRJArsQjxk

"Where we have Kindle editions, Kindle sales are now 35% of books." According to the chart it went from 13% in February to 35% in May with the launch of the Kindle 2. During the Kindle 1 period it was in the 5% - 10% range.

[Dahak]

Originally Posted by sirbruce

Dahak,

Actually, I found a link the first ten minutes of the video of the announcement, so you can hear the exact quote yourself:

http://www.youtube.com/watch?v=dPRJArsQjxk

"Where we have Kindle editions, Kindle sales are now 35% of books." According to the chart it went from 13% in February to 35% in May with the launch of the Kindle 2. During the Kindle 1 period it was in the 5% - 10% range.

BTW... the name's Joe. I haven't used it in the recent posts, but while I use this handle a lot, I don't live behind it!

Thanks for that.

Good quote. Completely unqualified. And I don't for a minute believe it.

I feel certain that if we twisted Bezos' arm, we'd learn that it was a rather finely-tuned stat. Like I said in another post, I have no particular qualm with that, he's a salesman, after all. I just want enough information to know where the facts end and the spin begins.

Strip out all of the freebies and PD-near-free and the not-unreasonable python-lump of backlist sales, then the number of newly released books - which I feel is a more realistic ongoing figure - would have a much lower percentage.


The first graphic he showed with the numbers of Kindle Editions in the store is a wonderful example of lying with statistics: the time scale across the graph is not linear. To greatly improve the 'acceleration' in edition numbers, the initial point (Nov 07) is much closer to the introduction of the Kindle II (Feb 09) than it should be.

If you take the two right-hand data points and say that the difference between the two is a three month gap, then the Nov 07 data point, properly scaled, should be somewhere off the left edge of the screen.

-Joe

[sirbruce]

It wasn't meant to be time-linear. Anyone can grab the numbers and do such a chart to see what it looks like; I did this for the thread on the Kindle forum tracking the number of books on Kindle and the trend is indeed accelerating.

[Dahak]

Originally Posted by sirbruce

It wasn't meant to be time-linear. Anyone can grab the numbers and do such a chart to see what it looks like; I did this for the thread on the Kindle forum tracking the number of books on Kindle and the trend is indeed accelerating.

Actually, it doesn't matter what it's meant to be time-linear, the impression given to the average reader is that it would be linear. Nothing indicates a break in the scale and I doubt that oversight was unintentional.

And... not to put too fine a point on it: I don't believe you can prove an acceleration of numbers based on three data points.

Clearly, the numbers are increasing, but you have nothing to check the rate of increase on that slide.

You have a figure for month 1, months 15 and month 18.

If the rate of increase was constant over the entire first 15 months, you'd be looking at about 9300 titles/month. But that's assuming a constant increase over the entire period. It's much more likely that the pace has increased in discrete steps as more staff is assigned to conversion duty.

The last three months, Bezos implies a rate of 15000 titles/month.

Unfortunately, without more data points, we have no idea how that pace changed during the rather large first time frame.

I ran across an interesting bit here where he's been recording the Kindle Store's own numbers daily for the last two months and the rate of increase has been virtually constant over that time.

-Joe

[sirbruce]

Again, as I said, we have tracked the numbers provided by Amazon on their website since release in this thread. The numbers are indeed accelerating; here's the chart.

[poohbear_nc]

One disturbing trend I've noticed (esp. in Sci Fi titles) is the publication of short stories - are these padding the numbers and being counted as "books"?

[sirbruce]

Originally Posted by poohbear_nc

One disturbing trend I've noticed (esp. in Sci Fi titles) is the publication of short stories - are these padding the numbers and being counted as "books"?

Sure, as are self-published books if you don't want to count them as "real". On the other hand, millions of books in Amazon's physical book database are not in print; they exist only as placeholders or as venues for selling used copies. And with vanity publishing, one could easily inflate the pbook numbers as well.

I agree it's an imperfect measure, but the fact remains it's accelerating. It is unlikely that "real" books are growing steadily but the "books we shouldn't count" are accelerating at twice that rate. And then who decides what books should be counted and what books should not?

[Dahak]

Originally Posted by sirbruce

Again, as I said, we have tracked the numbers provided by Amazon on their website since release in this thread. The numbers are indeed accelerating; here's the chart.

-SNIP IMAGE-

That data is absolutely wonderful. Thank you for tabulating it for us.

But there seems to be a bit of a discontinuity in the summer of '08.

It's been a while since I've played with numbers like this, but I believe that if you fit the curve from the start to about June or the beginning of July '08 and do a separate curve fit for Sept '08 onward, both curves would a be quite a bit more linear.

Yes, there is some acceleration, but that extrapolated end-point would be lower if you start the curve-fit in Sept '08, instead of Nov '07.

Thanks again for the numbers.

-Joe