WARNING: Fun with statistics ahead!
I've been wanting to make this post for a bit, but was missing a piece of the puzzle until NPD put out its public facing Press Release this morning. The report isn't incredibly descriptive, but it finally gave me the last bit of information I needed.
Many people are aware that, according to NPD's numbers, the iPhone 4S, 4, and 3GS were the top three selling smartphones in America in Q4 2011. If that were the only significant thing to say about the report, I would have simply linked to the NPD report and been done with it. Instead, though, I've noticed that a bit of fun with statistics can reveal some interesting numbers.
First, it's important to note that I do not, and will never, think that market share is in itself a valid assessment of whether or not a product is superior to another. My focus here is not in any way to declare Apple the "winner" of the quarter. So before I go on to make an entire post about how well Apple did, let me point out a two positive things about its competitors:
- While Apple sold the majority of smartphones in Q4, Samsung sold the plurality in Q3--that means in Q3 Samsung, not Apple, was the best selling smartphone vendor.
- In Q4, according to NPD, Android still outsold iPhones by 5 percentage points in the total smartphone market and by 23 percentage points in the first-time smartphone buyers.
We could discuss the significance or lack of significance around these points for ages, but that's not my point here. I'm not talking about overall market numbers for platforms or operating systems. My interest for the sake of this particular post is only to discuss sales of particular mobile phone handsets.
On that subject, NPD says:
Led by continued steady sales for Apple’s iPhones, the top five best-selling mobile phone handsets in Q4 were as follows:
- Apple iPhone 4S
- Apple iPhone 4
- Apple iPhone 3GS
- Samsung GALAXY S II
- Samsung GALAXY S 4G
So, from the outset we've established the iPhone was counted as the top 3 selling mobile phones in Q4. That seems pretty straightforward, but it's rather deceptive. It might already seem to speak volumes that all 3 iPhone models on the market outsold every other individual smart phone, but upon further inspection the disparity in single-model numbers is even further in Apple's direction. Before we go further, let us dissect iPhone sales a bit further.
From NPD's report:
"Attracted by a faster processor, improved camera and the Siri speech-driven agent, most iPhone buyers paid a premium for the iPhone 4S, making it the top-selling handset in Q4," said Ross Rubin, executive director, Connected Intelligence for The NPD Group. “The iPhone 4S outsold the iPhone 4 by 75 percent, and outsold the iPhone 3GS, available for free on AT&T, five to one.”
That last sentence is going to gives us the basis of some basic algebra. The awkward switching of phrasing to indicate sales comparisons will cause some issues, so let's simplify:
"iPhone 4S outsold the iPhone 4 by 75 percent" indicates:
iPhone4S = 1.75 * iPhone4
~57.1% iPhone4S = iPhone4
In the meantime, "and outsold the iPhone 3GS, available for free on AT&T, five to one" translates to:
iPhone4S = 5 * iPhone3GS
20% iPhone4S = iPhone3GS
So, in Q4 2011 in the United States:
iPhone4 = 57% iPhone 4S
iPhone3GS = 20% iPhone 4S
iPhone4 + iPhone3GS = 77% iPhone 4S
After a bit of algebra, we get the total breakdown of iPhones sold in the U.S. in the quarter (according to NPD) at about, roughly:
iPhone 3GS: 11.3%
iPhone 4: 32.2%
iPhone 4S: 56.5%
While the percentage of total iPhones is nice to know, the previous set of numbers comparing iPhone 3GS and 4 sales to 4S sales are more useful going forward.
What we've established now is the number 3 selling smartphone in the country (the iPhone 3GS) sold about 20% of the best selling smartphone (the iPhone 4S).
Here is where things get interesting.
According to AllThingsD's reporting of a consumer survey by Consumer Intelligence Research Partners:
iPhone 4S unit sales follow a predictable pattern, with the lowest-priced model being most popular:
- 16 GB: 45 percent of sales
- 32 GB: 34 percent of sales
- 64 GB: 21 percent of sales
It's important to note that CIRP announces 89% of iPhone sales as 4S sales (with iPhone 4 and 3GS at 7% and 4% respectively), but is specifically starting from the 4S's launch date half a month into the quarter. For two weeks all iPhones sold were iPhone 3GS and 4 units. Even then, I'm doubtful those two weeks bring the total 4S number down from 89% to 56.5%--but, according to Chris Rawson at TUAW CIRP's total numbers skew total iPhone revenue about 6% higher than Apple's--meaning they're a bit heavier on 4S sales than reality. High, but close enough to approximate individual capacity proportions.
Here's where my numbers are going to start getting a bit rough. Because I don't have exact numbers, and because my NPD-based numbers differ from CIRP's, I'm going to be approximating even more than I have been, but I'll be careful to justify any time I do.
Using my per-model numbers according to NPD with CIRP's 4S capacity breakdown, I get an average selling price of just $628.78, lower than the official one ($659) by a substantial margin. That means that, if NPD's per-device model numbers hold, the more expensive 32GB and 64GB models actually sold as an even higher proportion of 4S units than CIRP's numbers. In reality it's likely the real numbers lie between the two.
So with all this discussion surrounding the breakdown of sales per-capacity of 4S units, what happens if we no longer count the iPhone 4S as one phone, but three? Where do the individual units, especially the lesser selling 32GB and 64GB varieties, fall on the scale of phones sold in America? (This is fair because the iPhone 4 and 3GS are now only sold at 8GB capacity. There's no further breakdown of the two.)
Let's, just to make things especially difficult, continue to use CIRP's relatively pessimistic mix of total 4S units sold per-capacity (16GB @ 45%, 32GB @ 24%, 64GB @ 21%) with NPD's pessimistic mix of devices sold (iPhone 3GS @ 20% total iPhone 4S units, iPhone 4 @ 57% of total iPhone 4 units).
That puts us at:
(Be careful to notice the percentages offered are how each model compares to the total number of 4S units sold--NOT total iPhones. These numbers are used purely because they're a method by which to rank each individual split against the others--they obviously should not add up to 100%.)
- iPhone 4 - 57% total 4S sales
- iPhone 4S 16GB - 45% total 4S sales
- iPhone 4S 32GB - 24% total 4S sales
- iPhone 4S 64GB - 21% total 4S sales
- iPhone 3GS - 20% total 4S sales
- Samsung GALAXY S II
- Samsung GALAXY S 4G
Well. That list looks remarkably different now, doesn't it? Keep in mind this is definitely skewed towards lower price products selling in higher proportion than actuality--giving the aforementioned ASP of $628 rather than the accurate $659. Even then, every single different capacity of iPhone sold better than any other individual smartphone sold by anyone in this country for the quarter.
You might think now that that I've established that the iPhone could be considered to be the best selling 5 phones, rather than the top 3, that I'd have finished having fun with these numbers. If you did thing that, you'd be wrong. All I've finished so far is finding what can be solidly shown to be true in a "worst case" scenario when considering Apple's sales mix. What if we now move on from what is certainly true and begin to look at what is probably true?
What if we split the iPhone 4 into CDMA (Verizon & Sprint) and GSM (AT&T) models? This is a valid split because the 4S unified models with a common cellular chip which handles both CDMA and GSM--meaning the models themselves are the same. Now, instead of 5 models we have 6. According to Ars Technica:
Tccording to fourth quarter 2011 results, AT&T activated 4.1 million iPhones, while Verizon activated 4.2 million. Sprint would not disclose the number of iPhones it activated last quarter, but we feel safe in assuming that number is less than 4 million. Assuming Sprint was able to activate (perhaps a generous) 2 million or so iPhones, only a little over a quarter of iPhones were sold in the US.
with this correspndoning update:
(Editor's note/update 1/26/2012: The numbers used in the previous paragraph for AT&T's activations is incorrect, as they are from 2010. The fourth quarter 2011 numbers from AT&T show 7.6 million iPhone activations. Those numbers were released after this report was published, but they are relevant to the calculations in this article. With the new numbers, that means roughly 37 percent of iPhones sold last quarter were inside the US.)
So, assuming equivalent mixes of 4 to 4S sales across carriers, we're looking at AT&T being somewhat more than 50% of iPhone 4 sales in the US. Assuming an AT&T/Verizon/Sprint ratio of 7.6/4.2/2, we get a iPhone 4 GSM/CDMA mix of 7.6/6.2 or 55%/45%. That means GSM iPhone 4 hits about 31.35% of total iPhone 4S numbers and CDMA iPhone 4 hits about 25.65% total iPhone 4S numbers, changing our list to this:
- iPhone 4S 16GB - 45% total 4S sales
- iPhone 4 GSM - 31.35% total 4S sales
- iPhone 4 CDMA- 25.65% total 4S sales
- iPhone 4S 32GB - 24% total 4S sales
- iPhone 4S 64GB - 21% total 4S sales
- iPhone 3GS - 20% total 4S sales
- Samsung GALAXY S II
- Samsung GALAXY S 4G
So now we can see a way to easily see the iPhone as, most likely, the top 6 selling smartphone models.
But wait! There's more!
What if we ignore the previous breakdown, going back to 5 iPhones, but now split them by color? We then get 9 iPhone models (all the previous 5 except the 3GS are available in black and white--the 3GS is only available in black).
This is when we start running into trouble. You see, up until now I haven't even glanced at what the Samsung GALAXY S II's sales are because, well, they haven't mattered. The splits we've done have kept things above the 3GS, which we already know to be above the GALAXY S II. By definition then, these new split models have still been above the Galaxy S II. This is the first time we're looking at splits that would take us blow the 3GS's sales proportions (Remember! According to pessimistic numbers giving the 3GS a likely inflated number!) After all, at this point if we split 4S 32GB and 4S 64GB sales in to perfect halves per color to keep the lower numbers as high as possible, that still gives us new SKUs with numbers at 12% and 10.5% of total 4S sales--far beneath the 3GS's 20%.
Unfortunately, I don't have Samsung Galaxy S II sales numbers. Does that mean I'm finished with my thought exercise? No it doesn't. Here's where I blow your mind.
Up until now I've been trying to see how far I can stretch Apple's numbers by unfairly splittings its SKUs while not touching anyone else's. In doing so I've been comparing "sub-models" from Apple to full models from competitors. Obviously, this was to make a point. In effect, though, I've been giving Samsung an unfair advantage.
That entry on NPD's list at number 4 says "Samsung Galaxy S II". Let's hop over to Wikipedia for a second. What's this? An entire section called "U.S. Variants"? That's right. That number 4 item actually accounts for all phones flying under the Samsung Galaxy S II flag. These are:
- Sprint's Samsung Galaxy S II Epic 4G Touch - 4.52 inch touchscreen, capacitive buttons, no NFC
- T-Mobile's Samsung Galaxy SII - Significantly different processor, different cellular radio, same screen, same buttons, NFC
- AT&T's Samsung Galaxy SII - Same CPU as Sprint's, 4.3-inch touchscreen, thinner, NFC
- AT&T's Galaxy S II: Skyrocket - 4.52 inch screen, LTE, faster processor
- AT&T's Samsung Captivate Glide - Lower quality 4 inch screen, pop-out qwerty keyboard, slower processor
Those are a lot of differences. In fact, all 5 of those phones is more different from each of the others than the iPhone 4 is from the iPhone 4S. Those changes I made concerning capacity and color? Nothing in comparison to the differences here.
Would anyone care to bet that splitting all the above proposed sales units I made before being divided into two colors each would be above the Galaxy S II after it's split out into 5?
Exactly. It's certainly possible that a single one of those Galaxy S II variants will remain above the black or white version of a single one of the iPhone models split by color--but frankly it's unlikely.
So splitting the iPhone into 9 units makes it, most likely, the best selling 9 smart phones in America. Remember when I split the iPhone 4 into CDMA/GSM models above? If you'll notice, both iPhone 4 models were still listed ABOVE the lower-selling 32GB and 64GB 4S models. That means adding that back into the mix AND dividing it by color would still position the iPhone 4 network-split and color-split models above some of the individual 4S versions. That means that if the lowest selling 4S color/capacity variant still manages to sell above the best selling individual Samsung Galaxy S II variant (which is not guaranteed, but certainly possible given the above logic)--the iPhone was actually the top 11 selling smartphone variants in the United States in Q4.
Want to have your mind blown further?
I've kept saying best selling smart phones. Now go read the header to that NPD press release I linked to. What do you see? That's right, there's no mention of the word "smart" in the title. That list is actually the list of best selling "mobile handsets". That means the iPhone quite possibly accounted for the eleven best selling mobile phones--smart or dumb--in the United States for Q4 2011.
That is, well, insane. It's not often you can read a single player has taken the top 3 positions in a competition of any kind and then discover the announcement is deceptively understated.
-----
*Note* Once more, keep in mind I chose a combination of numbers on hand that would give me considerably more difficulty than reality in making this point. All of the arguments in the latter half get easier when you realize the real numbers, though unknown, are assuredly heavier on the products splits which fell lowest on my lists.