iPhone XR Accounted for an Estimated 32% of U.S. iPhone Sales in November

[0388631]

I'm not surprised. Though I presumed the take rate would be higher than what's reported. The XR at 128 GB of storage is definitely a good buy for those who'll keep their phones for more than two years with a battery swap at some point. With some minor drawbacks, I don't see a damn thing wrong with the phone.
[doublepost=1545885983][/doublepost]

Long term, the XR will do very well and it should, as I think Apple wants this phone to be the real contender against the competition.

Speaking of that, I think I read a headline and paragraph stating the XR outpaced a developmental version of the upcoming S10 on several fronts, and lost on some. It's fantastic value for money, especially with trade-in.
[doublepost=1545886038][/doublepost]

That's why Apple is actively making an effort to keep older iPhones usable. This is the new Apple strategy.

Old. Old iPhones working years later is not a new trend.

 

[iZac]

It makes me sad to see the SE blip out of existence.

Interesting to see the percentage mix of 'new' models is less this year than last.

Also aggressive trade-ins are the best idea in a saturated / mature market, because it get's a unit out of circulation that will likely otherwise be handed down or sold 2nd hand, which eventually is one lost sale.

 

[Spizike9]

I don’t know anyone with an XR but I saw one at the Apple Store the other day. I only played around with it for a few minutes but it seems like an awesome phone. Imho, I still believe the iPhone 8 Plus is the best phone currently offered in terms of features/pricing.

That’s interesting I’ve seen more XRs than XSs. However I probably seen more 7 and 8 pluses more than anything else. I work in a bank so I see quite a wide demographic from young to old. My dad and my friend and his wife just bought XRs though.

 

[john123]

Aside from the title being completely misleading, conducting a survey of only 165 customers of the thousands if not millions that actually bought iPhones in November is just ridiculous. Who is supposed to take such a survey seriously?! It is "not a large sample size". True. It is completely irrelevant!

lol at 165 customers ... i am sure if I’d pick 200 apple users in my class, about 2 would have an XR.

Sample size 165? Like .000000000000001% of users. Gosh, how scientific.

Good God, people. If you don’t understand statistics, that’s fine. But then keep your mouths shut on the topic.

Whether the underlying population has 10000 or 1 million people is nearly irrelevant when talking about sample sizes. Yes, <200 is a rather small N, but if (big if) it’s conducted scientifically, it’s more than adequate for ballparking a reasonable estimate.

According to another study*, 0% of iPhone 7 owners upgraded to a newer model in 2018. So that study is clearly wrong.

*Study based on a sample of 4 owners of iPhone 7 I know of, including me.

Which is not at all the same thing.

165 people for a survey! Some people don’t even try anymore. Let me know when it’s a few thousand.

A few thousand would be a colossal waste of money The increase in precision from 1000 to 3000 is tiny—about 1.3 percentage points actually.

 

[Luke MacWalker]

The sample size of this survey is 165 people and it's USA-only. Given the size of the smartphone market, it's hardly what I call meaningful. The most accurate I've seen in terms of iPhone adoption trends is this one by Mixpanel

https://mixpanel.com/trends/#report/iphone_Xs

At the bottom of the referenced graph above:
"THIS REPORT WAS GENERATED FROM 1,192,663,560,936 RECORDS"

Wow! That sure beats 165 !

 

[dampfnudel]

Good God, people. If you don’t understand statistics, that’s fine. But then keep your mouths shut on the topic.

Whether the underlying population has 10000 or 1 million people is nearly irrelevant when talking about sample sizes. Yes, <200 is a rather small N, but if (big if) it’s conducted scientifically, it’s more than adequate for ballparking a reasonable estimate.



Which is not at all the same thing.


A few thousand would be a colossal waste of money The increase in precision from 1000 to 3000 is tiny—about 1.3 percentage points actually.

Well, 165 is undeniably a waste then. I’m sure you agree based on your reply.

 

[nexusrule]

According to another study*, 0% of iPhone 7 owners upgraded to a newer model in 2018. So that study is clearly wrong.

*Study based on a sample of 4 owners of iPhone 7 I know of, including me.

Again, ignorance about how statistics work doesn't make anyone point valid.
[doublepost=1545905496][/doublepost]

Well, 165 is undeniably a waste then. I’m sure you agree based on your reply.

Nothing on his reply implies that. I really don't understand why 1. People have to talk and even judge and criticise things they don't know anything about 2. Even when someone tries to educate them, they need to keep judge and criticise thing they clearly are not able to understand. To be able to admit one is wrong is a sign of intelligence guys.

 

[CarpalMac]

Study some statistics and then come back.

To get a vaguely precise sample you would need at the very least 300 people and even then that'd still be fairly low accuracy. Indicative at best. 500 is better, 1000+ from a diverse demographic, then you are really talking.

There is between 7 to 11 devices being researched there. Even taking just the lowest number of 7 devices (being kind), that is an average of 23.57 people per device or put another way, each person representing 0.606% of their results. So, if one person swapped from one device to another that is a 1.2% swing. This makes it a far too sensitive a measure to be of any real meaning.

Perhaps you and some others in this thread should take your own snarky advice, rather than dishing it out.

 

[WatchFromAfar]

I thought Apple was doomed?
Maybe not then.

It's 32% of iPhone sales not moble phone sales. Taken to its extreme Apple could have 1% of the mobile market, of that 1% 32% could be Xr sales and the article would still be true....

 

[nexusrule]

To get a vaguely precise sample you would need at the very least 300 people and even then that'd still be fairly low accuracy. Indicative at best. 500 is better, 1000+ from a diverse demographic, then you are really talking.

There is between 7 to 11 devices being researched there. Even taking just the lowest number of 7 devices (being kind), that is an average of 23.57 people per device or put another way, each person representing 0.606% of their results. So, if one person swapped from one device to another that is a 1.2% swing. This makes it a far too sensitive a measure to be of any real meaning.

Perhaps you and some others in this thread should take your own snarky advice, rather than dishing it out.

What do you do for a living? I am a user experience design, but my background is in social sciences. I paid my studies doing design and when I finished my studies I moved to UX design that's 50% user research through qualitative and quantitative methods. I think I know what I am talking about, and not because I look for some basic notion than I then apply without the proper background knowledge to reply to an internet post. I can tell you that you don't know how the sample has been selected, so if you were serious you wouldn't criticise without having this information. Also, your computations are quite meaningless anyway, considered that they totally ignore factors like demography, geography, psyography and behaviour that are equally important when defining a sample. I won't contribute to this discussion further because I see it's a total waste of time.

 

[john123]

Well, 165 is undeniably a waste then. I’m sure you agree based on your reply.

You'd be wrong because there's no additional context provided as to why 165 was selected. But a little common sense and inference gives the answer. The 165 mentioned is likely a sub-sample from a larger survey—specifically what’s known as a tracker. The sample size used for their survey for the fiscal quarter that ended March 31, 2018 was N=500. The 165 covers a 30 day period. What’s 165 times 3, which would cover a full quarter? 495. Boom.

A small sample doesn't make a survey a “waste." It simply means the estimate will be noisier (less precise).

The tracker approach to survey research is very common. It’s standard. It’s methodologically sound.

I wouldn’t read too much into these exact percentages, but I also wouldn’t dismiss them outright.

 

[haydn!]

The sample size of this survey is 165 people and it's USA-only. Given the size of the smartphone market, it's hardly what I call meaningful. The most accurate I've seen in terms of iPhone adoption trends is this one by Mixpanel

https://mixpanel.com/trends/#report/iphone_Xs

Interesting, and tells a slightly different story. Looking at that the XR has sold faster than the XS to reach a similar market share in less time.

Looking at the full last 12 months though, none of the new models have yet reached the same market share the three iPhones launched last year achieved by December 2017. Loosely supporting the notion the 2018 lineup is underperforming.

https://mixpanel.com/trends/#report/iphone_Xs/from_date:-365,report_unit:month,to_date:0

 

[john123]

To get a vaguely precise sample you would need at the very least 300 people and even then that'd still be fairly low accuracy. Indicative at best. 500 is better, 1000+ from a diverse demographic, then you are really talking.

There is between 7 to 11 devices being researched there. Even taking just the lowest number of 7 devices (being kind), that is an average of 23.57 people per device or put another way, each person representing 0.606% of their results. So, if one person swapped from one device to another that is a 1.2% swing. This makes it a far too sensitive a measure to be of any real meaning.

Perhaps you and some others in this thread should take your own snarky advice, rather than dishing it out.

Wrong. First off you’re throwing out numbers like there are magical cutoffs. There aren’t. Second, “diverse demographic” is meaningless. If you care about analyzing sub-samples with precision, yes, then you need more respondents. However, you can construct a perfectly valid stratified sample and get unbiased overall estimates with a small N.

Statistical estimators—good ones—have three major properties: unbiased, consistent, and efficient. People here are confusing those properties.

It’s made worse by conflation of data with interpretation, and worse again when people attach nonsensical descriptive terms to those interpretations—like “real meaning.”

This is why those of us who do this for a living don’t use point estimates much. We use distributions and ranges. But non-stats people don’t understand that (heck, they don’t even understand point estimates—see this thread for proof). So there’s a tension between dumbing results down to make them readable and accurately describing the data.

But the important thing for most purposes—the property we care about—is unbiasedness. And despite all this blather from everyone about sample size, that’s been almost entirely overlooked in this thread.

 

[Candykane75]

I thought Apple was doomed?
Maybe not then.

40% of sales are old models and Xr is cheaper iPhone so 70 % of all phones are budget choises of people. These are percentages and is small sample size so its saying nothing about total global sales . It does show that most people buy the cheaper models , so if Apple wants to grow this gives you a indication what people want no $1000+ IPhone

 

[B4U]

Wondering if they have kept the SE as a lower priced option and see how the sales would be.

 

[jonblatho]

Those of you commenting on the sample size need to read up on statistics.

165 sample size with a population well in the millions gives a (95% confidence) margin of error of about ±7.6%. There are also concerns regarding the sample itself. How was it selected, and how representative is it?

I have read up on statistics, and that's a lot. This likely resembles the real sales mix, but there's a healthy amount of uncertainty in the data and I'm reluctant to draw too many conclusions from it.

 

[john123]

165 sample size with a population well in the millions gives a (95% confidence) margin of error of about ±7.6%. There are also concerns regarding the sample itself. How was it selected, and how representative is it?

I have read up on statistics, and that's a lot. This likely resembles the real sales mix, but there's a healthy amount of uncertainty in the data and I'm reluctant to draw too many conclusions from it.

See my post above. The reason this is only 165 is (probably) that it’s one-third of a quarterly tracker. So your instincts are exactly right, but we should also get more clarity in a couple months—subject to the caveat with any tracker that while you’re ostensibly gaining precision, it’s also possible that the underlying trends are moving as well.

 

[dampfnudel]

You'd be wrong because there's no additional context provided as to why 165 was selected. But a little common sense and inference gives the answer. The 165 mentioned is likely a sub-sample from a larger survey—specifically what’s known as a tracker. The sample size used for their survey for the fiscal quarter that ended March 31, 2018 was N=500. The 165 covers a 30 day period. What’s 165 times 3, which would cover a full quarter? 495. Boom.

A small sample doesn't make a survey a “waste." It simply means the estimate will be noisier (less precise).

The tracker approach to survey research is very common. It’s standard. It’s methodologically sound.

I wouldn’t read too much into these exact percentages, but I also wouldn’t dismiss them outright.

So at the very least you do agree that these surveys should be taken with a grain of salt, but not dismissed outright. Good enough for me.

 

[gijoeinla]

Here in Los Angeles, I sampled 5 large sizes AT&T stores asking three questions.

1) How are the iPhones selling? Hands down the answer was very briskly.

2) What model is most popular? Hands down the iPhone XSMax

3) What brand of phones is your top seller? Apple? Samsaung? LG? Apple - by a long shot.

As a side note 4 out of 5 stores reported that they struggle to keep stock on the Apple Watch 4’s. Very popular.

 

[john123]

So at the very least you do agree that these surveys should be taken with a grain of salt, but not dismissed outright. Good enough for me.

I'm not gonna say grain of salt, but yeah, tomato/tomato. My main objection is the chorus of people screaming about the 165 and implying that there's some magical threshold or, worse, implying that a much bigger sample size is needed because there are "millions" of iPhone users. Both are false.

And I'd be much more interested in potential bias effects, including self-selection bias. In the market research work I've done, that's far and away the most nefarious and difficult factor we face. Everything else is really just balls in urns.

 

[GoodWheaties]

It makes me sad to see the SE blip out of existence.

Interesting to see the percentage mix of 'new' models is less this year than last.

The only reason the SE isn’t in this graph is because Apple doesn’t sell it anymore, so how could it be there? I still know lots of people, including myself, who use it and are quite fond of it.

I do have to wonder if the price of phones isn’t making people keep their phones longer.

 

[iZac]

The only reason the SE isn’t in this graph is because Apple doesn’t sell it anymore, so how could it be there? I still know lots of people, including myself, who use it and are quite fond of it.

I do have to wonder if the price of phones isn’t making people keep their phones longer.

Sorry yes, that was the underlying message, I'm sad that Apple axed it rather than commit to a market segment that no one else plays in anymore (in terms of handset size)

I guess they're conflicted because their intent was to create a lower priced model for emerging markets with a dated design, but many still hold this design as the Heir to the iPhone 4 design pinnacle (myself included). Especially when compared to the 6 design that took (purely subjective opinion) three years of design tweaks to 'fix' until it was acceptable with the 8.

Personally - device performance has kept me with my SE too - it still holds up, but my 2G, 3GS and iPhone 5 never aged as well.

If there's a silver lining, Apple seem to be open to resurrecting seemingly dead products, if the iPad mini rumours are anything to go by. It's just a shame they can't do it without having insultingly long gaps between releases for many 2nd tier products.

 

[macdragonfl]

40% of sales are old models and Xr is cheaper iPhone so 70 % of all phones are budget choises of people. These are percentages and is small sample size so its saying nothing about total global sales . It does show that most people buy the cheaper models , so if Apple wants to grow this gives you a indication what people want no $1000+ IPhone

Plenty of people want and buy $1000+ iPhones. This was a new top tier that Apple created. They also have options for people who do not.

 

[Candykane75]

Plenty of people want and buy $1000+ iPhones. This was a new top tier that Apple created. They also have options for people who do not.

If 30 % is Plenty in your eyes then yes , but if 70% of your market is going for a cheaper or older model. Then in my eyes you need to ajust your strategie. if 40% of your market are buying older models , your new models are not atractive enough. 20+ % are buying phones that are 2 or more generations old ...
say Apple would sell the Xr of 600 dollars i know they would sell a lot more new phones , sure the margin would be less but still a lot Better then selling iPhone 7

 

[john123]

Plenty of people want and buy $1000+ iPhones. This was a new top tier that Apple created. They also have options for people who do not.

If 30 % is Plenty in your eyes then yes , but if 70% of your market is going for a cheaper or older model. Then in my eyes you need to ajust your strategie. if 40% of your market are buying older models , your new models are not atractive enough. 20+ % are buying phones that are 2 or more generations old ...
say Apple would sell the Xr of 600 dollars i know they would sell a lot more new phones , sure the margin would be less but still a lot Better then selling iPhone 7

Yes, 30% is more than "plenty." As macdragonfl was saying, there are options—so it's not an either/or choice. The reason companies make product models and lines is so they can have different products that appeal to different segments.

The XR can't sell for $600 based on the teardown estimates. It's simply too expensive to make, and one thing Apple (like any mature company) isn't going to do is sacrifice what have been remarkably stable gross profit margins (percentages) for increased revenue. People around here seem to have the idea that Tim Cook pulls pricing out of a greedy hat, but that's simply not how it works in mature businesses in the real world.