I _know_ I am not representative, but I might as well start with personal anecdata. I have bought a metric fuck ton of kindles of various stripes. I've bought them for me, for relatives and friends as gifts, for family members. I'm going to _only_ think about the ones I have bought for myself, and ignore the original kindle which was bought for me and which sits, registered but not currently in use, on the second floor, as a display item. Because. Basically, I've got a Voyage which I use almost daily. My son has a Touch which he almost never uses. Every other kindle e-reader I have bought for myself or a family member (with the exception of one which bricked and was replaced, so we're going to treat those two kindles as a single one; and one which I gave away so we're going to treat that as a gift) has been donated to a library and, as of a few months ago anyway, was still in active circulation.
So: 3 year device lifetime, ha ha ha ha ha. I more or less reflexively buy a new one whenever there is a new one and then I donate the previous one, where it is then used for a really long time by other people.
Back in April, this was produced:
Several things to think about here. That was in the spring of last year, when people were really all excited about Amazon stock, and part of what that piece was written in service of was ... people really all excited about Amazon stock. Also, keep in mind that every number that went into the final how-many-are-in-use-now was ALSO an estimate. Because Amazon doesn't release much data.
"The annual sales for Kindle e-reader peaked in 2011 to roughly 13.44 million, followed by a decline in 2012 when the figure totaled an estimated 9.68 million. The sales in 2013 were more or less flat, which suggests that roughly 43.7 million Kindle devices had been cumulatively sold till the end of 2013....Assuming a 3-year replacement cycle, we conclude that there may be approximately 30 million Kindle e-readers currently in use."
So when The Telegraph, this month, produces an article about Waterstone's and kindle ereader sales, including this paragraph:
"Amazon launched the Kindle, which is now in its seventh generation, in 2007. Sales peaked in 2011 at around 13.44m, according to Forbes. That figure fell to 9.7m in 2012, with sales flat the following year. It is estimated that Amazon has sold around 30m Kindles in total."
You should be thinking, hey, those look ... similar! So I think we have found the Telegraph's source. And the Telegraph is sort of assuming that there have been essentially zero net sales of kindle e-readers (that is, every kindle e-reader sale since the Forbes estimate was produced was a replacement sale, recognizing that there could be some hundreds of thousands or even small number of millions that failed to be captured by rounding to a single significant digit).
A Tim Waterstone quote concludes the article this way:
"“The e-books have developed a share of the market, of course they have, but every indication – certainly from America – shows the share is already in decline. The indications are that it will do exactly the same in the UK," he told the Oxford Literary Festival."
Back at the beginning of 2013, the WSJ put out an article on e-reader vs. tablet sales that boded poorly for e-readers. Basically, everyone who wanted one had acquired one in the last year or so, and they didn't wear out so no one felt compelled to get another just yet.
How does that interact with a 3 year hypothetical lifespan? Well, maybe sales are low because people are reading on their phone/tablet. Maybe sales are low because people went back to paper. Maybe sales are low because their vintage 2011 e-reader is Just Fine, thank you very much. Maybe sales aren't low ... because Amazon isn't sharing.
Last June, the then-current IHS iSuppli numbers were leading commentators to predict that e-readers were going to be a niche only for people who read two or more books a week (what I would call a true power reader -- 100 a year), for everyone else a nice to have that can be replaced with a tablet or phone.
McQuivey over at Forrester was speculating they'd eventually be given away for free.
The actual number of power readers out in the population could be determined, but I mean the 100 a year crowd, not the dozen ish a year crowd, which a lot of the research studies. The 20+ a year crowd seems to be about a fifth of the US population, and I'll just walk right out on a limb and suggest that the 100 plus a year crowd is a comparable fraction of that one fifth.
30 million does not seem nuts, under this analysis, if it's a US centric figure; it would be a little high if it were a US ONLY figure (which it almost certainly is not).
Long term, until we get to a point where the screens on phones and tablets don't keep us awake at night, I think it is plausible to believe that something like 5-10% of the population is going to want to have an e-reader around. Even people who don't typically read enough to really justify having a dedicated e-reader might decide to have one because it's just so awesome to travel with one. That makes 30 million as a US centric figure look downright reasonable (note: not annual sales -- sales figured over some number of years that represents the real world lifecycle of the device).
Of course, the day will come when a new screen technology changes everything. But people have been expecting that new screen technology to come along for a while now (Mirasol, anyone? Liquavista? Color e-ink? Speaking of which: http://www.popsci.com/article/technology/will-i-ever-get-color-e-reader). It'll happen ... eventually. This should not be regarded as a prediction for what happens after new screen technology changes everything; this is more along the lines of what the steady-state might be for dedicated e-readers under current screen technology.
ETA: Not about devices, just a little sidelight on Tim Waterstone's theory that e-book content share is dropping (if he meant content -- it's just not clear to me, altho it may be shortly when I grope around for further details):
Publishers weekly reporting on Nielsen data on e-books, paperbacks, hardcover, back in September.
A plateau is predicted at 35%: