Revised Smartphone Installed Base Numbers, Global and Related Handset Stats - Based on recent Pew survey data

I've been reporting quarterly the installed base of smartphones for a long while now and I believe this blog has been the only source in the public domain to give that count, also including the market shares by operating system. That is probably the most relevant number to most who read data on smartphones ie the developers and various players in the ecosystem. Only the handset manufacturers are really interested in the raw sales numbers and market share that is reported by the other industry analysts quarterly, and all others would then need to try to estimate that number relating to the installed base.

Now that we had that wonderful Pew survey of international mobile phone ownership by adults (ie 'unique owners') and I did some deep analysis of that data projecting from it global numbers and adjusting for the missing youth phone ownership for total global per-capita unique owners, we have some very rare data on the total planetary ownership of mobile phones. But that data also is a check on the total installed base of all phones in use, whether smartphones or dumbphones, which included multiple phone ownership. And the data I've been reporting here quarterly has recently had a creeping error, which I estimate started around year 2003. The recent first-time mobile phone buyers are far poorer than those of the past (who also often buy very low-cost phones whether smartphones or dumbphones; and often buy second-hand phones) They tend not to afford to replace their phones as often as richer buyers, and that means, the total installed base has seen a big growth in older phones still in circulation. Mobile phones are often kept longer and they are also increasingly passed down to family and friends (especially when a teenager gets his or her first smartphone, if often is a hand-me-down) and thus the age when a phone exists the installed base has grown longer than my model. I did a massive readjustment to my model and have updated now my installed base to reflect a base of 2.6 Billion smartphones in use worldwide (this is total smartphones, including those who own more than one) which is in line with the Pew data for 2.3 Billion unique smartphone owners at the end of 2015. So 300 million of all smartphones in use are in pockets of someone who owns two smartphones. Thats obviously 13% of all smartphone owners.

Now I expect that my model was accurate (enough, to say 100 million) in year 2011 so I've adjusted my installed base number and market shares back to 2011 and am now sharing those with you. My quarterly market share blogs in the future will be built upon this number, so if you look at past blogs, there will be a jump but this is the way to adjust those numbers for those who want to have time series numbers for trend analysis.

GLOBAL SMARTPHONE INSTALLED BASE AND OS MARKET SHARES 2011-2015 (REVISED APRIL 2016)

Rank . OS Platform . . . . Units . . . Share 2015 . . . . . . . 2014 . . . . . . . . 2013 . . . . . . . . 2012 . . . . . . . . 2011
1 . . . . Android . . . . . . . 2,079 M . . . . . . . . 79 % . . . . . . . 76 % . . . . . . . 67 % . . . . . . . . 51 % . . . . . . . . 32 %
2 . . . . iOS  . . . . . . . . . . . 505 M . . . . . . . . 19 % . . . . . . . 19 % . . . . . . . 20 % . . . . . . . . 22 % . . . . . . . . 19 % 
5 . . . . Windows* . . . . . . . 31 M . . . . . . . . . 1 % . . . . . . . .  2 % . . . . . . . .  3 % . . . . . . . .  3 % . . . . . . . . . 3 %
4 . . . . Blackberry . . . . . .  15 M . . . . . . . . . 1 % . . . . . . . .  1 % . . . . . . . .  3 % . . . . . . . .  7 % . . . . . . . . 12 %
5 . . . . Symbian . . . . . . . . . 2 M . . . . . . . . . 0 % . . . . . . . .  1 % . . . . . . . .  5 % . . . . . . . .15 % . . . . . . . . 31 % 
Others . . . . . . . . . . . . . . . . 8 M . . . . . . . . . 0 % . . . . . . . .  0 % . . . . . . . .  1 % . . . . . . . .  2 % . . . . . . . . . 3 %  
TOTAL Installed Base . 2,640 M . . . . . . . . . . . . . . . . . 2,184 M . . . . . 1,651 M . . . . . 1,172 M . . . . . . . 812 M
Total all handsets . . . .  5,605 M . . . . . . . . . . . . . . . . . 5,497 M . . . . . 5,323 M . . . . . 5,096 M . . . . . . 4,707 M
Smartphone migration . . . 47 % . . . . . . . . . . . . . . . . . . . 40 % . . . . . . . .31 % . . . . . . . 23 % . . . . . . . . . 17%
* Windows includes Windows Phone, Windows 10 Mobile and discontinued Windows Mobile
All annual data installed base reflects market on December 31 of that year
Source: TomiAhonen Consulting April 19, 2016 and TomiAhonen Almanac 2016
This data may be freely used and repeated

So overall compared to my old model, I have raised my total number of global smartphone installed base as of December 31, 2015, by 7%. That has mostly gone to Android where the obvious growth has been mostly in the past 3 years as iPhone's unit sales market share has declined from its peak of 20% in 2012 to the 15% it was in 2014 and it was last year. Note that in my model I already had a very strong long-life factor for the iPhone and the iPhone installed base has for this whole decade been above the actual unit sales market share - due to the long life span of iPhones, I could not really adjust the iPhone number up by much but I did increase it by 2% vs 7% overall for the industry. (Also there is of course a less than sold impact adjustment for Windows as so many of those phones are not ever activated) Still, the iPhone total installed base is 505 million today, good for a corrected installed base market share of 19.1% (vs 20.1% using my older model). iOS fans and developers need to remember that is installed base of iPhone units specifically, ie smartphones, not other iOS devices like iPads and iPod touch media players and Apple Watches, which would of course grow the total accessable market as an OS platform similar to how Android tablets, smart watches, netbooks and other gadgets increase the total also for the Android OS market.

TOTAL INDUSTRY METRICS 2015

So now with these adjustments, we have revised numbers for the full year 2015. Total humans on the planet alive at the end of 2015 was 7.3 Billion people. There were 7.6 Billion active mobile phone accounts ie subscriptions (usually SIM cards) either post-paid or prepaid, combined. That means a top-line metric for our industry of 104% per capita mobile phone account penetration rate. Note in comparing such stats, this is not 'per household' number nor is this a 'by adult population' number. It is per-capita so counting all humans alive from babies to great great grandparents. That 7.6 Billion includes machine-to-machine subscriptions of about 400 million (5% of all active mobile phone accounts are now connecting machines).

As per the Pew survey sample of adults in 30 countries, adjusted for world population we get total unique mobile phone ownership at 5.0 Billion. That is 68% of all humans alive, and this is the 'best measure' if you want to know how far can mobile reach. Yes more than two out of every three humans alive on the planet has both a mobile phone in their pocket and its connected to an active account so the person can be reached for example by SMS text message. The unique user count, easy to remember this year, its an even 5.0 Billion people who have a mobile phone. And calculating out the multiple accounts, 34% of all mobile phone accounts in use, are second or third accounts (whether connected to a second phone or not; some save money owning one phone but swapping SIM cards into it and can easily have four SIM cards and one phone). If you consider the error in using total subs vs unique users, that is now 52%. You OVERCOUNT the actual market by 52% if you talk about total mobile subscriptions when you mean to talk about unique users.. Its time to learn those numbers and use the one that is relevant to you and your business. Sometimes you do want to reach the target on any device he or she is using, so the total subscription count is not an inherently wrong number, just that it is increasingly misleading now in this industry. Usually the unique users number is the best for measuring the actual mobile audience total size.

Some of us have more than one phone. So what is the total phone population in use? 5.6 Billion so 12% of us have two phones. For many of us that is two smartphones. For some who recently became smartphone users, it may be that the older phone is still a featurephone. And for some who just need basic connectivity (say a taxi driver) then both phones (or all 3 or 4 phones) can be very basic phones just that customers on any network can call or text that taxi driver directly. 

As it comes to smartphones, yes 2.3 Billion unique smartphone owners have 2.6 Billion total smartphones in use. Thats 13% of all smartphone owners who have two smartphones. Among dumbphone owners (2.7 Billion people with 3.0 Billion phones) the number who have two phones is of course less as a percentage, its 11%. That will keep coming down rather fast as most who can afford two phones will be migrating to smartphones.

WORLD MOBILE CUSTOMER STATS 2015
World Population . . . . . . . . 7.3 B
Mobile Subscriptions . . . . . 7.6 B
 - of which M2M subs . . . . . . 0.4 B
Mobile Phones in Use . . . .  5.6 B
 - of which Dumbphones . . .  3.0 B
 - of which Smartphones . . .  2.6 B
Unique Mobile Owners . . .  5.0 B
 - of which Dumbphones . . .  2.7 B
 - of which Smartphones . . .  2.3 B
Source: TomiAhonen Almanac 2016
The above data may be freely shared and quoted

So that is what our world looks like today. Feel free to share these numbers with anyone you want and put into infographics and quote in your presentations and stats. Please refer to the upcoming 2016 edition of the TomiAhonen Almanac as the source if you need a formal source or you can also refer to this blog of course. Meanwhile for those who need over 100 such stats for mobile industry, the best stats source every year is the Almanac, and my 2015 edition came out late last year. So the freshest mobile stats on everything you ever wanted is all here, in one volume: TomiAhonen Almanac 2015.

Comments

[abdul muis]

How can BB number be that high with only less than a million unit sold per quarter? That's like BB from 3.5-4.5 years ago still 100% in use.

[abdul muis]

I'll repost this here....

You're iPhone will expired in .....

http://www.financialexpress.com/article/industry/tech/your-iphone-will-expire-in-three-years-says-apple-inc/238233/
"Apple has said the iPhones have a life expectancy of three years and so does the Apple watch while AppleTV devices expire in four years."

[abdul muis]

(sorry to put this here)

intresting report.
http://static.googleusercontent.com/media/source.android.com/en//security/reports/Google_Android_Security_2015_Report_Final.pdf

Data Collection: decreased over 40% to 0.08% of installs
Spyware: decreased 60% to 0.02% of installs
Hostile Downloader: decreased 50% to 0.01% of installs

[Tester]

No, Wayne.

Tomi already said that the average upgrade cycle is 18 months. The rest is second and third hand use.

Don't twist the numbers around to your liking. 3 years on average sounds reasonable for smartphones - not just for Apple but for the industry as a whole. Many last longer, but a significant quantity won't even survive so long because their users mishandle them.

My old Android phone is 5 years old now and still good enough for someone who doesn't need to have internet access all the time. But that old thing is far more robust than all those shitty 'premium' metal cases that have become so popular - not just at Apple. I guess the real reason for their popularity is that they are so fragile that they bend and break easier, i.e. shorter replacement cycle.

[Tomi T Ahonen]

Hi abdul, Lullz, Tester, Wayne

abdul - the Blackberry number like all installed base, reflect all still in use and because of Blackberry's rapid decline in its new sales, that is mostly older phones 'still in use'. But is 15M reasonable? Just adding the last 2 years, Blackberry sold 13.7M units and essentially all of those will be in use as anyone who had a BB before but didn't like it, will have gone long since to rivals and these are very hard-core loyalist BB users. But it now also has a meaningful slice of very old phones still in circulation. Yeah am very confident 15M is about the right number.

BTW to all - I wanted to include the Symbian number just for contest (and rub it in with Elop haha)

On the iPhone ownership and 3 year period, most who own an iPhone won't hold onto it for 3 years but many will. Think of your own smartphone ownership, us geeky users will want new tech fast and 2 years is an eternity to hold onto a smartphone when the newer ones keep getting better tech. The global average is no longer 18 months, its up to 21 months and growing for the average replacement cycle for smartphones - that means many replace smartphones FASTER than in 21 months, some will replace their phone every year. That is more true of rich people for whom a 600 dollar cost is peanuts, than for low income people who recently bought their first smartphone. So those who replace every year skew STRONGLY to iPhones vs Androids. But equally - the total life of iPhones is FAR longer than those of average cheapo Androids - partly as Apple products also are VERY well made - and being expensive - they also are often better taken care of (and not given to kids as readily as cheaper phones).

With bad phones we want to get rid of them fast. So some of the brands with problems like say HTC on the Android side have had users quickly get rid of them and wanting to replace as soon as it becomes practical.

But yeah, Apple has the longest replacement cycle and has the best rate of surviving to the second-hand market (something that used to belong to Nokia) which extends the total usage of a relevant fraction of iPhones in use. With all that, Apple sold 15% of new smartphones in 2014 and 16% in 2016, that is down a lot from the peaks at 20% so even with the long life span, iPhone share of installed base is not expected to grow, it should come down at a gentle pace but will continue to stay well above the new sales market share level.

Tomi Ahonen :-)

[abdul muis]

@Tomi

OK. It feel like already 3 years since BB decline around 1 million device/quarter.

[abdul muis]

@Tester

"My old Android phone is 5 years old now and still good enough for someone who doesn't need to have internet access all the time. But that old thing is far more robust than all those shitty 'premium' metal cases that have become so popular - not just at Apple. I guess the real reason for their popularity is that they are so fragile that they bend and break easier, i.e. shorter replacement cycle."

I still have Samsung Galaxy Nexus (Dual Core / 1GB RAM). It still good for playing games, Hangout, Gmail, Web browse, Youtube, etc. The only problem is whatsapp. If I installed whatsapp, after 1-2 months of receiving message/picture/video, the phone will lag beyond useable.

[abdul muis]

@Lullz

iPhone... "It may get slower with OS upgrades but that's another story".

Since you mention it, I was reading the article about Apple sabotage the old iPhone to be slow with OS upgrade & the article from financialexpress (the link in my previous post), but forget to put that link too.

Anyway,

I was wondering have you try using the 3GS with whatsapp in 2016. Really using it? Like having a 8-10GB file (photo, chat, voice, video) in Whatsapp folder? Because I'm not talking freshly installed whatsapp. I'm talking about a very active whatsapp usage with Galaxy Nexus with about 8GB-10GB files in whatsapp folder.

[abdul muis]

http://www.digitimes.com/news/a20160420PB200.html
Global smartphone shipments reach 292 million units in 1Q16, says TrendForce

* Combined shipments (including exports) from China-based brands reached 125 million units in the first quarter, surpassing the combined shipments from Samsung and Apple for the first time, and also accounting for 42.9% of the global smartphone shipments, up from 41.5% in the previous quarter.
* Apple posted its largest quarterly decline ever for iPhone shipments, totaling only 42 million units in the first quarter, plummeting 43.8% from 75 million units shipped a quarter earlier.
* Samsung's first-quarter shipments exceeded expectations and reached 81 million units, up 2.5% from the prior quarter. Samsung boosted its smartphone sales by launching its flagships Galaxy S7 and S7 Edge ahead of schedule and stepping up its promotional activities.
* Huawei's first-quarter shipments came in at 27 million units, or a drop of around 20% compared with the prior quarter. Nonetheless, Huawei retained its rankings as the top brand in China and the third-largest vendor worldwide.
* Lenovo smartphone shipments reached 17 million units in the first quarter, translating to a small quarterly decline of 5.6%. Lenovo will be focusing on foreign markets this year and have assigned 80% of its total shipments for exports.
* Xiaomi shipped about 16 million units of smartphones in the first quarter and is now in close competition with Lenovo for the number two spot in China, TrendForce said.

[abdul muis]

@Lullz

"The new Android versions have also been slower then the new ones. What's the point?"

The point is, if Apple did make the old iPhone slower, it will be a good reason for that person to upgrade to iPhone SE or big iPhone. Otherwise, there is no reason to upgrade at all.

"I have a iPhone 3GS 32GB here and can't really understand what would be the problem with having 10GB of files on the phone. The files can be stored there but what's the point?"

Then, it's not apple-to-apple comparison. I could put lots of files in my GNex, and it won't slow down. I could install WA and use the storage for other things till end, and no slow down....

AGAIN, what I said... THE problem of slow GNex is the bad programming of WA programmer, that make WA hog the memory if there's a lot of WA content in the phone.

[Tester]

@Lullz:

"This behavior is one reason why iPhone has been gaining more installed base than market share. Android phones have this problem that they became really slow if you use certain apps."

Perfect FUD and completely wrong. Of course old phones become slow if you fill their limited storage, or organize data in a way that managing it brings out performance issues. That has nothing to do with Android but only with the software asking for more resources than such old hardware got - which ultimately boils down to sloppily programmed software. Android's problem here lies elsewhere: Java makes incredibly easy to write such software!

And what people like you fail to mention is that an iPhone of the same age is completely obsolete because Apple deciding to end support for it. For this old Android phone I still can install most of today's software.

[abdul muis]

@Lullz

Bravo bravo...
You're answering me without answering me...
Well done

[NO ONE WANTS A WINDOWS PHONE]

The piece of garbage Windows Phone is dying away... even our microsoft astroturfers, including Baron95, are not telling ups how wonderful it is any more ....LOL!

Remember: NO ONE WANTS WINDOWS ON A PHONE!!!! ...they never did LoL!

[Wayne Borean]

LOL. My first iPhone lasted three months. Our Evil Siamese Cat TM decided that it would make a good water bowl ornament. My wife and I saw her shove it off the kitchen counter, where it was charging. Right into her water bowl!

The replacement had later five years, and is still in use.

We also have an iPhone 5 and an iPhone 6 in the house, along with two cheap androids. The cheap androids are less than two years old, and no longer in use. Everybody hated them.

But that's us. We think Apple makes a damned good phone. Of course if we'd tried a top of the line android we might have liked it, but since I'm on disability we tend to be really careful with what we spend our money on.

[abdul muis]

@Wayne

Sure,
I have BMW 750 & Mercedes Benz S500, I also got a dozen of Tata Nano, and we all hate the Tata.... Nice comparison....

[Kamehameha]

@abdul

"Apple posted its largest quarterly decline ever for iPhone shipments, totaling only 42 million units in the first quarter, plummeting 43.8% from 75 million units shipped a quarter earlier."

Good info buddi.
iSheep lullz start cry now.
iPhone crash

[Per "wertigon" Ekström]

If Apple sold only 42M units, that's a catastrophic collapse of units since they sold 61M units last year (e.g. 30% unit collapse)

I'm waiting for the official numbers however, that number seems suspiciously low. I'm anticipating a lesser decrease in numbers. Again however, my rough forecasts for anyone that cares can be found here:

Data - https://docs.google.com/spreadsheets/d/1hIbcjggIqiYF9lS2LbLWMcOypA6NoqTs-7EawsAEyfw/edit?usp=sharing

Apple vs Global market - http://imgur.com/jOKGj8d

Apple moving average marketshare - http://imgur.com/VWFCAQf

[Per "wertigon" Ekström]

@Lullz:

If you don't trust me then take a copy of the data. That copy won't change now will it?

But yes, I will revise those documents as Tomi and Apple fills in the details.

It is also a rough forecast and I don't sit on all data so it's bound to be off by as much as 10M or so for Apple.

Tomi has a far better analysis basis than I do, but for the general trend - it's down for Apple. :)

[Per "wertigon" Ekström]

@Lullz:

I redraw my model after I get facts that tell me the model is off. So the numbers I give today are constantly revised as the facts move along. Most analysts work this way.

To expect me not to change and revise my model is simply silly. About as silly as expecting a boxer cannot dodge any punches.

What does not change however is the direction Apple is headed. They had a great year with iPhone 6 - But it was a one-off. Now it looks like Samsung will be strong again. :)

[Tomi T Ahonen]

Hi everybody

On the Apple pattern, its now going to change and be far less the 'staircase' model of one huge jump in sales going to Q4 (in Calendar Quarters not Apple Fiscal Quarters, so Q4 is Oct-Dec and Q1 is Jan-Mar quarter etc) and then have three relatively flat quarters as the peak recedes, and as Apple launches its next new model just in the last days of the weakest of those three ie Q3. With the dual launch dates now, one in the Autumn and the other in the Spring, the extreme variation within any 12 month cycle will be significantly dampened but still the main new phone launch will produce again a peak as before.

With that, the past four years have established a rather consistent pattern of iPhone sales. If we consider the calendar quarters, Q1 is typically down 16% vs previous year's Q4, then Q2 is down another 21% vs Q1. Q3 is up 6% with the new launch and first days, but then Q4 is up 68% vs Q3. And this pattern continues. Its been pretty close to that for the past four years (obviously am excluding the year when the launch window was changed ie 2011 5 years ago where the long delay to the new iPhone meant a prolonged wait and thus a higher spike).

For 2015 four quarters, the above average model would suggest (Calendar) Q1 of 2015 should have been 84% of the Christmas sales of 2014 ie 62.8M (reality was 61.1M). Q2 then as 79% of Q1 actual should have been 48.2M (reality 47.5M), Q3 should have been 50.3M (was 48.0M) and Q4 should have been 80.5M (was 74.8M).

As you can see, the model is quite close to what happened and on most quarters it would predict the actual iPhone sales better than most 'expert' analysts of Wall Street (and helped me be usually among the most accurate iPhone unit sales forecasters, when I'd take the other factors into consideration than strictly the mathematical average in that pattern).

Now. That model suggests iPhone should be selling now at 84% of the Christmas iPhone sales level ie 63.1 million - and nobody is expecting the January-March quarter to be anything like that good. We received repeated news from just about all sources that iPhone factories are not as busy, that parts are not in short supply, that the market share numbers of iPhone is consistent with bigger decline etc. This is part of why clearly Apple rushed its Nano to the market - had Apple had this launch date well planned in advance (ie known all year 2015 that the new model WILL launch in 2016 and that the launching pattern will be changed from one annual launch date of all iPhones to two launch dates) then the iPhone 5SE would have been ready for sales on the last four or so days of March to boost the year's first calendar quarter fiscal reporting at Apple. That will of course become the pattern from next year so the first sales date for whatever iPhones are launched in the Spring will appear in stores on the last days of March.

Once again, a decline in iPhone sales from Christmas quarter 2015 to the January quarter 2016 is not in any way 'bad news'. The whole INDUSTRY sees a downturn in smartphone shipments in this quarter. But the SIZE of the decline, in particular the hit to market share that is likely to come - that is news. But now Apple has already announced its answer to this 'problem' ie iPhone 5SE - so it should not be hysteria all around.

For us.. this is the last quarterly result where the maximum volatility in the iPhone unit shipment numbers was in play - but we all know how to counter that - we use the 12 month moving average as the measure of how Apple is doing vs the competition - that means always every quarter we factor out the within-year-seasonality which is exceptionally volatile with Apple's launch pattern. So we are not alarmed either way, when the big spikes come, or when the big drops come.

And luckily from now on, going into the April-June quarter and ahead, Apple will see a far less extreme pattern to its sales.

Tomi Ahonen :-)