Measuring Mobile Platform Churn in the US Market

The following chart shows the net gains in users for the major mobile platforms in the US. The data is derived from comScore’s MobiLens report.

The raw numbers for the last period (ending May) are:

  1. Android gained 2.1 million users
  2. iPhone gained 1.0 million users
  3. WebOS/PalmOS lost 96k users
  4. Blackberry lost 202k users
  5. “Other” lost about 95k users

The “churn” or percent of smartphone users who switched platforms is 1.22%/month  of the overall users (assuming users to left one platform moved to another rather than to non-smart devices.)

The share of the platforms and the overall composition of users is shown in the following two charts:

As usual, caution should be exercised when looking at share (top chart) outside the context of overall growth. The second chart may be more illustrative of what is happening. The light blue area on top is the smartphone non-consuming population.

Android has been a tremendous accelerant of smartphone adoption while iOS has grown as well but share has not grown significantly. iPhone growth has kept up with market growth. This cannot be said for the other platforms which have been “squeezed” between these two entrants.

So although the main competition is with non-consuming users and smartphones are winning in general, the loss of users by some of the incumbents is unsettling.

Microsoft lost just over 2 million users in the US last year while RIM lost a net of 1.5 million. Palm lost about half a million. Contrast this with 8.5 million iPhone additions and 23 million new Android users.

In particular, Windows Mobile/Phone is down to 5.8 million users (down from 18m in December 2009). With a loss of half a million per month, and share down to 4.5%, a lot is riding on the new platform.

I’ve argued that the post-PC era will be a multi-platform era. Although it may seem that the situation in the US is rapidly moving toward a duopoly, it may not end up that way. Note how much non-consumption still exists.

Note the time frame of the charts is only 18 months. The rapidity with which platforms gain ground could allow for continuing platform churn.

  • Travis Lewis

    "The light blue area on top is the smartphone non-consuming population."

    Is that correct? or should it be Non-smartphone?

    • asymco

      Perhaps not the best wording, but I meant non-consumers of smartphones.

      • Ziad Fazel

        Horace, I think the top of your second graph, the total installed base of cellular users in the US, should be an upward sloping line. Surely it has not been flat at 230m since Dec 09 with the booming sales of smartphones. Your graph incorrectly implies a 1:1 swap of non-smartphone to smartphone with no growth in the overall installed base.

      • asymco

        I use the data from comScore. They claim the total US subs has remained constant over this time period.

      • Ziad Fazel

        Horace, have a look at the CTIA data

        "CTIA-The Wireless Association’s Semi-annual wireless industry survey develops industry-wide information drawn from operational member and non-member wireless service providers. It has been conducted since January 1985, originally as a cellular- only survey instrument, and now including PCS, ESMR, AWS and 700 MHz license holders….

        The CTIA survey also develops information on the number of reported wireless service subscribers or “connections” for the responding systems, and an estimated total wireless connections figure (taking into account non-responding systems). Because the CTIA survey is a voluntary survey, it cannot compel responses from wireless carriers. However, the survey has an excellent response rate. For the December 31, 2010, installment of the semi-annual survey, CTIA received responses from companies serving 96 percent of all estimated wireless subscribers.
        Because not all systems do respond, CTIA develops an estimate of total wireless connections. The estimate is developed by determining the identity and character of non-responding markets (e.g., RSA/MSA or equivalent-market designation, age of system, market population), and using surrogate penetration and growth rates applicable to similar, known systems to derive probable subscribership. These numbers are then summed with the reported subscriber connection numbers to reach the total estimated figures."

      • asymco

        Great data. Thanks for the link. I'll post some interesting patterns from the data.

      • Ziad Fazel

        Glad you like. CTIA's research and reports are meticulously prepared and peer-reviewed in the industry. I used them a few times to debunk garbage from product managers using paid analysis. Many chops and bones for data hounds like us to chew and gnaw.

      • Travis Lewis

        Yeah, I read it too fast. Soon as made comment I realized my mistake.

    • Horace put that awkwardly. I think he means that the light blue area is that part of the population not buying/using smartphones. I would guess this includes both users of non-smart phones and those that do not use a phone at all.

  • Steve Erickson

    Excellent set of charts. I enjoyed looking at the same set of data from multiple perspectives.

    Here's a quick design suggestion. You should consider changing the colors of iOS and Android in the third chart. In the previous two graphs iOS was gray and Android was blue, but then in the third graph they got switched. It took me a second to realize things had changed.

    Aside from that, this is another great post. This is such an awesome site. Keep up the great analysis.

  • Omar grant

    Great graphs, seems iOS & Android OS will end up as the dominant giants for the U.S. I’m curious where windows mobile technology will be in the next 24 months.

    • FalKirk

      This is unadulterated speculation, Omar, (but hey, that's the fun part, right?) but I predict that Windows phone 7 is Microsoft's next Zune and the tablet running Windows Phone 8 is their next Kin. Twelve months after Microsoft's new tablet debuts, the shock waves from its dismal showing will spread to the phone, the desktop OS and to the very foundations of Microsoft itself.

      Now, of course, I don't KNOW this. Microsoft hasn't even displayed it's tablet yet. I'm just going by what Microsoft says it's going to going to be doing. They say they're going to put a fully functioning desktop operating system on a tablet. Does that sound familiar? Oh yeah, we heard it in 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 and 2011. I guess a whole decade of failure wasn't enough to convince them that they had it wrong. But I predict that the 12th go round will do the trick. By the end of 2012, EVERYone will know that Microsoft's approach is wrong – even Microsoft. And their shareholders. And their board of directors. And Bill Gates. And the one man who may still be holding out hope – Steve Ballmer – will finally be gone. (Lord, how I'll miss him!)

      • FalKirk

        Just to follow up my rant, above, this fresh in from Microsoft:

        ""We view a tablet as a PC."

      • As per Gruber, MS's conceptual framework for tablets is outdated and I, too, give them little chance to succeed unless they take a completely different approach. Just guess: there was a turf war, and the WP7 people had their plates full, so the Windows division won. For now.

        WP7's failure is not a given, however. It might take them a long time to become a serious competitor with Android and Apple, but MS has shown that it has the patience (with Xbox), and has the pockets to do so. The patent licensing deals with Android manufacturers is a money maker that can support WP7 development and marketing, and if WP7 starts getting any traction, MS might undercut its own Android license fees with lower cost WP7 fees.

      • Rudolf Charel

        Your prediction of MS products being the Zune and Kin is wrong in my opinion.
        The comparison may well be that Windows 8 is their next Longhorn and the result could easily be their Vista.

  • Rob Scott

    I am stuck record on this:

    This comparison is by enlarge meaningless.
    Android is the only OS with upgradeable base, every other OS has to get users from would be Android users. When you fact that in, there is nothing special about Android, but something very special about the iPhone for an example.
    Smartphone data without context is meaningless.

    Now, you might argue that this is important if what you are interested in are platforms, unfortunately this is not a platform discussion, where again iOS kicks Android's ass in almost very measure.

    To close, let me be clear, I work with this data everyday, and I was too was fascinated by these simple observation until I realized that they were not that great.

    This is probably my last post on this, will have to ignore future post similar to this one.

    • ScottyRad

      Rob, could you explain what you mean by upgrade-able base? It seems to me like every OS is competing for the same people, to various degrees of success.

    • FalKirk

      With all due respect, Rob, take another look at the big sky blue portion of the third graphic identified as "non-consumers". No one has to take users from Android. No one has to take users from iOS. No one has to take users from anyone else. It's a land grab. There's more than enough non-conumers to go around.

      To mix and mangle most every metaphor possible, Android is winning the manufacturing war (they are making more devices than anyone else), Apple is winning the profit share ware and although a rising tide raises all boats, RIM, Nokia, HP/Palm and Microsoft continue to sink towards the bottom of the Bay.

      This data is NOT meaningless. To be fair I don't know what it means… but I refuse to believe that it's meaningless, damn it! That's my story and I'm sticking to it.

      And despite your most earnest entreaties, I will NOT ignore your future posts. You're just going to have to get used to disappointment.

      • name99

        To add to this, both Android and Apple seem to be aware of this.

        Apple has made their public statements about how they will not cede the low-end market (which have been analyzed to death in this forum as to whether this means non-contract phones, cheap phones for the US, cheap phones for Asia, etc).

        Android, meanwhile, has had advertisements very deliberately targeting non-smartphone users. Think, for example of that "Harold and his big-boy phone" TV ad from a few months ago, with a supporting cast of muppets.

      • westech

        Android doesn't manufacture hand sets, so they can't win the war. Apple is winning the manufacturing war. The Android OS hand set manufacturing is fragmented, which is why individual manufacturers (Samsung, HTC, Motorola, etc) can't come close to Apple's low manufacturing costs. The manufacturers who use the Android operating system(s) compete with each other so they don't share information. Each moves down its own learning curve.

        I believe that Apple is winning the mobile OS war as well. The metric here is not manufacturing cost but ecosystem dominance. Although I believe that Apple is clearly winning I don't have a metric to measure it. Apple's iOS learning curve encompasses not only all the mobile products they make but also Apple TV and the learning contribution from their MacOS products, especially the Macbook Air.

        The data ARE meaningful but you must make the right comparisons.

    • kevin

      Guess I must've missed your prior entreaties on this topic. But why is "Android the only OS with upgradeable base"? Please do clarify.

      Since the US market is largely (90%) built on postpaid contracts, almost all of the handsets should be considered upgradeable every 18-24 months, and thus, for the most part, people are free to switch platforms (even if they don't switch carriers). AT&T even let me upgrade my iPhone for the heavily subsidized price within one year (yes, of course, they knew Verizon was coming soon…)

    • EWPellegrino

      Ok, I'm assuming that by 'upgradable base' you mean the fact that existing dumbphone users are often already using handsets that are made by Android OEMs?

      If that is what you mean then to an extent you're just wrong because of course there is Nokia, which is still the largest maker of dumbphones and is a WP7 shop going forwards, and because the big Android OEMs often also sell WP7. Ignoring that though, there's a good question in there – how loyal are handset owners to the OEMs, or put another way, what is the churn between handset makers.

      In my (entirely anecdotal) experience, the only handset maker before Apple that had significant loyalty was Nokia.

      • jonshf

        In my also entirely anecdotal experience the loyalty to Nokia was mostly toward their dumb phone platform which had a simple, consistent and well designed user interface. When someone upgraded to a new Nokia dumb phone they instantly knew how to use it (we're not talking about geeks here) and the old chargers could be used (European homes typically had lots of them spread around the house).
        Again, it's the platform that people are attaching themselves to. Nokia's Symbian platform never got this kind of attraction as the user interface was horrible.

      • EWPellegrino

        Agreed, so the conclusion must be, if even Nokia with it's army of loyalists couldn't guarantee to keep them on board going dumbphone->smartphone, then Samsung likely won't. This was the key point that allowed Apple & RIM to break into the smartphone sector without ever competing for dumbphone share.

      • Nokia's S60 user interface on their Symbian based phones was almost identical to S40 or even S30. The similarity was deliberate so that people upgrading had some familiarity.

        That's where they went wrong. They kept hold of the old dumbphone UI on their smartphones yet asked increasingly more of the phone. The menu system is labyrinthine compared to an iPhone, even though that's partly because the iPhone does a lot less and is less configurable than a Nokia.

      • EWPellegrino

        even their feature phone UI drove me nuts, it seemed like every iteration they moved something – while still leaving it looking essentially the same. Moving to iOS was a huge relief. As you say, I bought Nokia back then because they made solid handsets that didn't feel like they'd break if I squeezed too hard.

      • Yep. That annoyed me too on S30/S40 and again on S60. Each release of the OS moved things around. It was equally confounded by carriers moving things around and Nokia itself moving things around depending on if you had an E series or an N series.

        At least on S60 you can move things back to where you like them.

        Apple simplified and that's why many people like them. Personally, I find them TOO simplified.

        Still, it's just frankly puzzling that Symbian ended up with S60's UI when there were much better alternatives like Nokia's own S80/S90 UI or SE's UIQ interface that also sat on top of Symbian.

      • Nokia's loyalty is almost entirely because of excellent hardware – bombproof handsets and excellent signal.

  • EWPellegrino

    That's surely only a lower limit on churn? In any period you will have people switching between say Android & iOS (in either direction) who are not being measured because both sectors are growing overall.

    • asymco

      Perhaps, but you have several mitigating factors: 1. Few Android users have used their first devices for longer than their expected lifetime so they have not had the need to replace the device. 2. Stickiness for iOS is so high that very few users switch out of it.

      • EWPellegrino

        Yes but I was only using those platforms as a 'for instance'. Consider RIM, even in the quarters when it gained subscribers you can be sure that it lost others. In the quarters where it lost subscribers it probably gained a few corporates, masking greater churn under the surface.

        It's still useful to have a lower bound on the churn rate though.

      • Kas

        Yes, based on comments from friends who are part of this Android surge, I think Apple will do well when all of these new Android users get out of contract. Once bitten, twice shy.

  • OpenMind

    iOS in smartphone seems stalling. In an old saying, you either go up or go down in a stream, but never stall. Will iOS go up or go down from here?

    • asymco

      This is why it's a mistake to look at share data in a growing market. It's not a zero sum game.

    • You don't fish for trout much do you? They will stay in a localized part of a stream for a long time;-)

      Even then, iOS's market share is showing slight gains even with Android's relentless march. I suspect the next iPhone will open the iOS platform to Sprint and T-Mobile increasing the potential customers in the US by 70% or so.

      • jonshf

        Not to mention the possibility of a China-mobile iPhone which would have huge implications. Of course, that is not part of the US market discussion which raises the question: Why are we looking at numbers for only one country which isn't even very representative of the world market? It's almost meaningless.

    • mbaDad

      Actually by these charts Android is stalling, as iOS is growing rapidly. But all of them are growing so fast it is almost useless data at this point, and will likely change quickly and often.

    • Rudolf Charel

      These figures are for the US market, which is a different one one from the international one. We all know that the US market is totally different from the one in other countries. Considering that Apple derives more than 50% of their sales abroad their “home” market pays a diminishing role in the overall picture. Therefor stalling, if that is really the case, does not necessarily apply to iOS overall.

  • relentlessfocus

    I also find market share by OS meaningless. It doesn't tell us or the firms involved anything about profitability, it says nothing about the focus of developers or the quality of the apps they build ( which might be reasons to look at Market share IF it correlated with developer interest), it doesnt tell us anything about the "stickiness" or brand loyalty and it's misleading in that people infer these things from Market share numbers.The market share metrict seems to be a dumbing down of business strategy to something that's graspable by a lay public more into taking sides and rooting for their side than understanding market place dynamics. of course on a gross level it's easy to see Microsoft and rim and HP have hills to climb but you don't need graphs or metrics to see that.
    And lastly when you have competitors with very different OS strategies, it's hard to understand how each is benefitting in their chosen strategy.

    I'd love to see a profitability by OEM chart, a stickiness graph and aspirational data if it can be found.

    • Market share, by itself, is a meaningless number. I think most here would agree with this. The numbers you are talking about are posted on this site every 3 months about 3 weeks after the various players financial quarters.

      Look here for some of the work Horace has done for specifically these numbers:

      Basically, you have to be careful not to get too centered on making only profits at the expense of market share. Market share decline is not sustainable. Likewise, if you are sacrificing profit to gain market share, you will find your self in trouble once the market becomes saturated and your price erosion forces you to lower quality to stay in the game. What these plots show is Apple is being able to maintain its market share. In the next 1-2 weeks, we will start to see if how the ARPU has been maintained by the various handset makers.

    • kevin

      Platform marketshare is one useful metric, among many. It's the score at the end of each month or each quarter, because it documents just how many users are using the platform. (Comscore measures users not devices sold.) But it is a backward-looking metric. It documents what has already happened. At half-time, it tells you the score at the end of the half, but doesn't clearly tell you who will win at the end of the game.

      And we (and the stock market) are more interested in the future, so we are looking for trends and deeper analysis of the marketshare metric. The marketshare metric is a good starting point for analysis. But you must do the analysis if you want any forward-looking value.

      Here's a beginning: Apple is still selling iPhones as quickly as it can make them, and iPhone sales growth closely correlates with the number of carriers and countries in which it is distributed. Android-based handsets are made by many more manufacturers, and distributed by a greater number of carriers and countries, and thus its growth is significantly outpacing iPhone. Apple is choosing to "slowly" increase its capacity (and by slowly, I mean close to doubling every year, which might be considered rapid in many other industries) and not adding more carriers and countries than its supply and customer service can support (with high-quality). The effect of this action is that at some point in the future, Apple will need to be getting switchers from Android. Is that a likely possibility?

      Not being measured here is iPod touch and iPad. What role will that play in the future and in the likelihood of net switchers into iPhone?

      The profit margin for Android handset mfrs is small, and getting smaller (with each royalty agreement). How will that affect future supply (or handset quality) and the likelihood of switchers? Etc. Etc.

  • bossjet

    The first chart is a little hard to read, perhaps a quarterly view would be better and understand macro trends while retaining enough granular data.

  • Tom Ross

    In the iPhone 3GS era (2009/2010) iPhone share went up first, then down, staying the same on balance. This year iPhone seems to have gone nowhere but up (albeit ever so slightly). Apple counter-programmed the cycle with the late Verizon release. There will likely be a sharper rise of iPhone marketshare in q4 of this year than we have seen in the last 2 years.

  • N. Yelpat

    One must not ignore the subscriber froth that is certainly taking place below the surface of these "in service" sub counts provided by Comscore. At best they can help us understand the net subscriber change for any given platform. However it misses all of the metrics that help us understand "how" the change occurred. Is Windows churning more subs than previous trend (likely), or have sales fallen off a cliff and are no longer adequate to mitigate the churn they have always experienced (also likely – but which is the greater culprit? What about customers who abandon the category – whether by choice, by not paying their bill, or other "natural causes"s? I have a hunch that your churn rate estimates are too low and we are missing the more interesting part of the phenomenon of the US market… thoughts?

  • Interesting graphs for the US. UK graphs would be interesting by comparison. Here RIM have just overtaken Apple. BBM is a killer app for kids.

    The drops in Windows Mobile in the last two quarters must be cause for concern for Nokia if they still intend to continue with their stupid move to Windows Mobile, even in the light of overwhelming praise for the N9 and finally some signs they've got their UI sorted in Symbian Anna and Belle. If Windows Mobile can't even pick up customers in it's heartland, it's surely in trouble.

    I'm wondering if Android's rise is a bubble. Anecdotal I know but I've bumped into quite a few people, generally older, who've been upgraded to Android smartphones by their carrier to find them too complicated, buggy and power hungry. These people are switching back to non-smartphones. I've also come across iPad owners who have ditched their smartphone for a cheaper, longer lasting non-smartphone as the iPad fulfills the app potential.

    There's lots to play for yet.

    • addicted44

      You are right about BBM. I see that a lot in India also. I wonder if iMessage can reverse that tide.

    • jonshf

      I wonder if those generally older people who have been upgraded to iPhone feel the same need to switch back to dumber phones. My guess would be not so much.
      Assuming the rest of the market that hasn't yet switched to smartphones are in general a less tech-savy group, what will that mean for the next market "land grabs".

      • You're probably correct with your guess but I'd guess that the 'problem' is that people are getting pushed the cheaper Android handsets as upgrades on the same tariffs as they had non-smartphones here in the UK without understanding what they're getting. iPhones are always on much more expensive tariffs (£25+) than non-smartphones whereas Android phones are in the £15-20 range.

        For some people an Android handset is a bit overwhelming, bigger and has crap battery life compared to their old non-smartphone.

        One of my relatives (he's a deputy head teacher in his 50s) had an Android upgrade (HTC Wildfire) from a Sony Cybershot phone and was complaining about essentially complexity and battery life. He sent it back and now has a Nokia C3-01 S40 phone which is simpler and lasts over a week on one charge.

        That's why I'm wondering if there's possibly a low-end Android bubble.

  • Pingback: What’s your prediction for the future? (We’re probably both wrong) « David McElroy()