Amazon’s Fire Phone was launched in July, after many teaser videos and an expansive unveiling by chief executive Jeff Bezos, who showed off features like its 3D-effect maps and multiple front-facing cameras.
But since then, how many Amazon Fire Phones have actually been sold? How many are in use?
There were many interesting anomalies in the launch of the Fire Phone. The $200 price tag for a 32GB version on an AT&T contract – the same price range as the iPhone 5S or Samsung GS5 – was a surprise to those who thought that the Kindle and Kindle Fire tablets showed Amazon as a low-end disruptor, selling hardware at cost and profiting on the content.
Amazon gave the Fire Phone pride of place on its US web site, and also had a deal to sell through AT&T, the biggest US carrier - though reviews were lukewarm.
Amazon famously never gives sales figures for any of the devices it sells, preferring to let its financial results do the talking. Analysts can make estimates of how many devices have been sold, based on their information from sales channels and any guidance the company might give. But it’s not as definitive as, say, the smartphone sales that Apple or BlackBerry include in their financial figures. (They are now the only two companies which give specific values for device sales in their financial results; Apple goes farther by giving the revenue from those sales too.)
But we can have a stab at estimating how many Fire Phones are in use, based on data from Chitika, which runs an ad network.
According to a release from Chitika, looking at activity on its ad network in the 20 days after the Fire Phone’s release, the Fire Phone accounted for 0.02% of activity - although a more precise figure, in another graph, shows it as around 0.015%.
OK - but how many phones might that be?
Using data from ComScore, which provides monthly data on US smartphone users, we can get close to a figure.
That figure is rising by between 1m and 2m a month - so two months later, by mid-August, when the Chitika data was collected, we could estimate that there are between 175m and 177m smartphones in use in the US.
We also have to assume that Amazon’s Fire Phone will show up on Chitika’s network as often as any other phone - so that their web use is comparable (ie one iPhone in use will show up on Chitika’s network as much as one Samsung or LG or HTC or Amazon Fire phone). We’ll look harder at that later.
Based on the ComScore data:
• 0.015% of 175m = 26,250 Fire phones in use
• 0.015% of 177m = 26,550 Fire phones in use
Split the difference, and it’s reasonable take 26,400 as a median value.
Indexing: the over and the under
But what if Fire Phones “under-index” on Chitika’s system - that is, they don’t show up on its ad networks as much as other devices, because their users are (for example) all spending time on Amazon buying things, rather than browsing the parts of the web that Chitika supplies ads for?
To explain “under-indexing”, imagine there are 100 owners of device A, and 100 owners of device B, and 100 of device C. If owners of A and C use their devices online an equal amount of time, but device B owners spend only half their time online, a system that measures web impressions will report that there are half as many owners of device B. One would say that device B “under-indexes”. Alternatively, if you have equal numbers of A, B and C, but device A owners spend twice as much time online as B or C, then device A “over-indexes”.
A Chitika release in August seems to suggest, for example, that the iPhone over-indexes: it had 52.3% of Chitika web traffic, though iPhones are about 40% of smartphones in use in the US.
Android devices (including Fire Phone, which uses a forked version of Android) do under-index - though not much. The same Chitika survey that shows the iPhone over-indexing gives Samsung a 22.8% web use share, and HTC a 3.1% share . Correction: Chitika’s US data in August shows Samsung with 26.4% of web usage, and HTC with 3.0% of web usage.
That compares to ComScore’s data, which shows that Samsung smartphones are 28.6% of the installed base, and HTC are 4.8%. (BlackBerry devices under-index on Chitika, making up 0.5% of browsing, but having 2.4% of the installed base.)
On that basis, you could argue that if the Amazon Fire Phone under-indexes, it probably isn’t by much; you could multiply the number by 25%, based on the average of the Samsung and HTC figures. That takes you up to about 33,000 devices.
Therefore even allowing for margins of error, it seems unlikely - based on Chitika’s data and the ComScore data - that there were more than about 35,000 Fire Phones in use after those 20 days.
Amazon had not responded to a request for comment on the calculation by the time of publication.
Lots of caveats apply: this is a calculation based on two non-congruent sets of samples, though both are large enough to be robust.
So is 35,000 phones good or bad? Bezos said that Amazon was in it for the long run and that it intends to be patient. Chitika comments that “While Amazon certainly is looking to make the Fire Phone a hit, current conditions show this being more realistic as a long-term goal rather than a short term one.”
Observers felt that Amazon had to get into e-commerce through phones because it risked being disrupted itself as more shopping - even the physical sort - moved to apps.
So far, though, the Fire Phone looks more like a curiosity than a barnstormer.
Correction: the calculation for under/over-indexing originally used the UK dataset from Chitika but compared it with the US installed base.
guardian.co.uk © Guardian News and Media Limited 2010