UBS Finds 31% of Android Users Likely to Switch to iPhone

| News

Investment bank UBS issued a report Friday showing that Apple’s iPhone crushed all would-be competitors when its comes to “stickiness,” referring to whether or not owners of a device would stick with that device the next time they upgraded their smartphones. Not only was Apple #1 with an 89% retention rate, well ahead of #2 HTC with a 39% retention rate, the survey found that 31% of Android owners were planning on ditching the little green monster in favor of an iPhone with their next purchase.

UBS surveyed some 515 consumers to measure smartphone stickiness by asking consumers, “which handset OEM they expect to next purchase a handset from,” according to Barron’s. As you can see in the figure below, Apple’s retention rate was far higher than any other OEM, with #2 HTC truly representing a “distant second.” 

UBS Smartphone Retention Survey

Chart by The Mac Observer, from UBS data

The survey also found that when you add up gains and losses for each OEM that 50% of those looking to switch smartphones planned on buying an iPhone. Both Samsung and HTC were also narrowly on the side of net gainers.

Which means there much be a big loser, and in this case it’s two big losers, Nokia and Research In Motion.  Nokia scored at the bottom of the survey, with a 24% retention rate, while RIM was third with 33%, but that’s down from 62% in just 18 months, a precipitous drop.

“When we look at all consumers who are considering changing handset Original Equipment Manufacturer (OEM) provider, Apple remains a significant net beneficiary — alongside Apple, only Samsung and HTC also appear as net beneficiaries (more users won than users lost),” UBS analyst Gareth Jenkins wrote in a research note for clients. “[RIM] and Nokia are the largest losers.”

He added, “Retention rates appear to be falling for most of the OEMs. The largest loser recently has been RIMM whose retention rate has dropped from 62% to 33% in 18 months. Relatively, Apple’s retention rates have held up incredibly well even as its market share has risen.”

At the same time, when the survey focused on platforms instead of hardware, Android performed better, with 60% of respondents saying they intended to stick with Android. While still far behind Apple’s 89%, it’s better than the retention rate of any individual Android OEM. 

That suggests there’s something about Android that some customers like, they just don’t like the way it’s being packaged by the OEMs.

Sign Up for the Newsletter

Join the TMO Express Daily Newsletter to get the latest Mac headlines in your e-mail every weekday.

22 Comments Leave Your Own

Haha

515 Costumers!! Great research you should get a Nobel prize!

Mark

So what? This is old news, last year same research also showed Apple was having loyal customer when Android still trail iPhone. Does it stop Android taking over iPhone and make Android having double market share as iPhone? No!

Bryan Chaffin

I’m guessing that neither of you understand statistics and the concept of samples.

Just clarifying

I believe that his point was that 515 consumers is *not* a valid sample size. the second poster probably meant to say that android users are growing faster than iOS users atm, which is also true.

dhp

A sample size of 515 is enough to have less than a 5% margin of error with a confidence level of 95%.

furbies

The way I look at it is unless they surveyed every soul on the planet, every survey is open to scrutiny to some degree.

wab95

That suggests there?s something about Android that some customers like, they just don?t like the way it?s being packaged by the OEMs


A fair statement, and perhaps yet one more reason why Google decided it was time to take some control over that packaging via Motorola. Looking forward to seeing how that turns out.

wab95

The way I look at it is unless they surveyed every soul on the planet, every survey is open to scrutiny to some degree


True, every survey is open to scrutiny, but the maths behind sample size calculation are fairly straightforward. That does not mean it is always easy. However, every survey should have a published methodology section, in which the assumptions on which the sample size, notably, background exposure variables, power (the 1 - the likelihood of a Type I error i.e. saying there is a difference when in reality there is none) and precision (1 - the likelihood of a Type II error i.e. saying there is no difference, when in reality there is), are stated and justified.

And, by definition, surveys are samples, with quantifiable sampling errors, which are quantified measures of uncertainty. That is the beauty of science. We aim to be precise in quantifying our uncertainty.

Bryan Chaffin

I believe that his point was that 515 consumers is *not* a valid sample size.

Yeah, I know. To clarify, that’s why I accused him of not understanding statistics and the concept of samples.

the second poster probably meant to say that android users are growing faster than iOS users atm, which is also true.

Aye, this is true, and I personally believe that it will continue to be the case. Fortunately, neither UBS’s survey nor the fact Android is growing faster than iPhones are mutually exclusive.

Macaman

I would have enjoyed if they did this same report taking out the people who own macs. I am sure it would vary considerably.

androcks

Lemme guess this survey taken at and around apple stores, right?

Windsor Smith

That suggests there?s something about Android that some customers like, they just don?t like the way it?s being packaged by the OEMs.

Yeah, and the bar graph you chose to publish exaggerates Apple’s dominance, because people who prefer Android can switch among OEMs, but those of us who prefer iOS can get it only on Apple hardware.

(I love my iPhone, but it’s not ideal for everybody. I, for one, wish it had a little LED that flashed whenever there are new voice mail or text messages in the queue. If I could get iOS and the LED from another OEM then I’d definitely consider switching.)

BurmaYank

”... the bar graph you chose to publish exaggerates Apple?s dominance, because people who prefer Android can switch among OEMs, but those of us who prefer iOS can get it only on Apple hardware.”

No, the bar graph chosen exaggerates nothing.

The chosen bar graph is what it is; it is an accurate measure of the Apple iPhone?s true dominance among all smartphone OEM device providers.

The chosen bar graph is NOT what it is NOT; it does NOT pretend to give a relative measure of iOS’s dominance/non-dominance among all smartphone OS’s, as you seem to be suggesting, since it doesn’t even distinguish the various leading smartphone OS’s among those listed OEM providers.

BurmaYank

Thanks again, Bryan, for such good bulletins as this one is, too.

Windsor Smith

The chosen bar graph is what it is; it is an accurate measure of the Apple iPhone?s true dominance among all smartphone OEM device providers.

I understand; my point is that comparing smartphone device providers in this way is flawed because it doesn’t disentangle the hardware from the platform. A substantial number of customers choose smartphones primarily on the basis of platform, not hardware. Apple has the fans of iOS all to itself, but HTC and Samsung must compete with each other for Android fans and effectively divide them up. Thus, this type of comparison is inherently biased toward manufacturers who have proprietary platforms. It doesn’t necessarily reflect the popularity of the hardware itself, because in some cases (Apple, RIM) you can’t separate the choice of hardware from the choice of platform.

Besides, that’s not the whole picture. The article (and the survey behind it) doesn’t just compare smartphone device providers; it also compares smartphone platforms. (The title, after all, doesn’t even mention a manufacturer; it mentions only platforms). So when I used the word “dominance” I was talking about dominance in both areas—hardware and platform. TMO’s choice to include only a graph comparing hardware, and not a graph comparing platforms, gives the casual reader an exaggerated impression of Apple’s overall dominance.

“There are three kinds of lies: lies, damned lies, and statistics.”

wab95

I understand; my point is that comparing smartphone device providers in this way is flawed because it doesn?t disentangle the hardware from the platform.


Windsor:

You are right to point out that platform and hardware comparisons are different, and should not be conflated, however I would argue that TMO (Bryan Chaffin in particular) have been quite precise in distinguishing comparisons of hardware vs comparisons of platforms.

That said, there is nothing inherently flawed about comparing devices by manufacturer (hardware). You may prefer to look at other data, or additional data, such as platform disaggregated from manufacturer, and then equipment separately, but the survey measured what it purported to measure, namely ‘equipment’.

It is not clear to me, from your argument, how this results in an observational bias; that is, whether you intend an informational bias from incorrect determination of exposure (availability of handsets) and outcome (choice of next purchase) as would occur if the survey were conducted at Macworld; or if you intend confounding, as would occur if the supposed exposure (availability of handsets) and the outcome (choice of next purchase) were related to a third factor (specific access, such as polling people in an Apple retail store or perhaps an AT&T store with limited handset availability).

Granted, I am being technical, but the term ‘bias’ is a technical term in surveillance. Can you clarify how Barron’s survey, or the data that Bryan graphed from that survey, reflect ‘bias’?

Windsor Smith

That said, there is nothing inherently flawed about comparing devices by manufacturer (hardware).

Here’s the flaw: With some hardware manufacturers the customer’s choice of hardware is independent of the choice of platform, but with other hardware manufacturers it’s inextricably linked. Putting them all on the same graph is a bit like comparing apples and oranges (no pun intended).

The graph measures customer retention by hardware manufacturer. If I have an Apple phone, and I decide to buy a new iOS phone, then I have no choice but to buy it from Apple; Apple will retain me as a customer. But if I have an HTC phone running Android, and I decide to buy a new Android phone, it’s much easier for me to switch to, say, Samsung, in which case HTC would lose me as a customer, and its retention statistic would go down. Since Android customers have more options, and since many (possibly most) customers like to stick with the same platform, it will be significantly harder for HTC and Samsung to retain customers than it is for Apple, simply because HTC and Samsung have to compete with each other in the Android market, while Apple has no competition in the iOS market.

Granted, I am being technical, but the term ?bias? is a technical term in surveillance. Can you clarify how Barron?s survey, or the data that Bryan graphed from that survey, reflect ?bias??

I didn’t mean “bias” in the statistics sense, so I probably should have used a different word to avoid confusion. Let me rephrase: Thus, this type of comparison inherently favors manufacturers who have proprietary platforms.

Bryan Chaffin

Thus, this type of comparison inherently favors manufacturers who have proprietary platforms.

In a phrase, and your point is?

By which I mean that you seem to be frustrated by the fact that UBS’s survey addressed the market place as it is, rather than some other state. That is how I’m interpreting your comments, at least.

It turns out that Apple’s platform is proprietary?if that in any way confers something favorable or unfavorable to Apple, the survey’s results reflect that and still represent the market place as it is.

In addition to that, the other proprietary OS/hardware platform in the survey, BlackBerry, performed just as poorly as the Android device makers did. Is RIM somehow not subject to the same unfair advantage as Apple?

In addition to that, as noted in the story, the survey also asked about platforms independent of hardware, where Android did much better, if still nowhere near as well as iOS.

As Wab95 noted, there are many different ways to look at different aspects of the mobile market. Handset makers, platforms, OS, profits, ad views, ad clicks (or taps), data consumed, apps downloaded, free apps or paid apps…

There is no end to the ways analysts can (and do) slice up this industry.

That there may be another way to look at retention rates doesn’t necessarily invalidate the way that UBS chose to measure it. In this case, I don’t think it does so at all.

wab95

I didn?t mean ?bias? in the statistics sense, so I probably should have used a different word to avoid confusion.


Windsor:

Your intellectual honesty is appreciated, and I concur, what you describe is not bias in the sense of survey design.

Here?s the flaw: With some hardware manufacturers the customer?s choice of hardware is independent of the choice of platform, but with other hardware manufacturers it?s inextricably linked. Putting them all on the same graph is a bit like comparing apples and oranges (no pun intended).

What you describe is not a flaw in survey design, but in the interpretation by anyone who fails to appreciate that the iOS has but one manufacturer, Apple, whereas Android has multiple. The survey is not at all at fault. Indeed, as Bryan notes, Apple’s portrayal is no different than that of RIM, which faired poorly. If anything, many pundits, as well as readers of TMO, have argued that this is the iOS’s ultimate weakness, namely that it is only available on Apple’s hardware, while the Android platform has multiple OEMs and would trump iOS on several metrics, notably marketshare. That it has compared unfavourably to iOS on the metric of repurchase or ‘stickiness’, which the survey reported for the Android platform in total, doesn’t invalidate the analysis, any more than surveys that show a greater overall Android than iOS marketshare. In short, these hardware as well as platform comparisons, like the consumers themselves who support them, are more nuanced.

What researchers often do in such cases as you describe, is to lump multiple entities that may represent a common component into a single category, in this case, Android OEMs, thus making Android manufacturers appear as one, similar to Apple. That would address the issue you raise, but would undoubtedly and justifiably raise objections from other readers that different vendors may perform differently, and therefore it is unfair to lump all Android OEMs into a single category. No one study will satisfy all readers, or address all issues.

Still, the comparison, as designed and reported, is a valid one, even if it has limitations. All well-designed surveys do.

Windsor Smith

In a phrase, and your point is?

By which I mean that you seem to be frustrated by the fact that UBS?s survey addressed the market place as it is, rather than some other state. That is how I?m interpreting your comments, at least.

Not exactly; I’m disappointed that the only graph in the article is one that shows retention by hardware manufacturer. To me, that graph is not very useful because when I review it I have to carry around all these asterisks in my head—like the possibility that Samsung’s number is substantially depressed by the fact that it’s easy for Android users to switch OEMs, so RIM’s higher score doesn’t necessarily mean its customers are more satisfied than Samsung’s.

The graph I really want to see is one that shows retention by platform, because (1) it would be easier to interpret, and (2) I think it’s more relevant and what readers are more interested in.

Probably the ideal approach would be to include both graphs, giving a more complete picture.

Windsor Smith

What you describe is not a flaw in survey design, but in the interpretation by anyone who fails to appreciate that the iOS has but one manufacturer, Apple, whereas Android has multiple. The survey is not at all at fault.

Agreed. I haven’t read UBS’s report (as far as I know, it’s not available to the general public), so I was trying to direct my comments toward the TMO article as presented.

Krabbie5

The way I look at it is, all apple folk have the same available choices as the restnof the pack and 9 out of 10 choose staying. Android folk have the same choices and many are changing for their reasons. It is soooooooooo much harder (read more expensive ) to retain a customer than to grow new ones. Apple gets it and the others don’t seem to get retention as much. That is how I see the graph and article.

Log-in to comment