How many smartphones did Samsung ship in Q2?

50.2 million.

That’s the estimate from IDC. So why bother asking?

Because that is an estimate. Of the 104 million Android phones shipped in the quarter (itself an estimate from another, possibly different methodology), I could only account for 7 million actually reported. That figure comes from a close reading of an investor presentation from Sony. HTC does not report their shipment numbers. It stopped some time late last year. Neither does Motorola now that it’s a part of Google. Huawei is silent except for setting targets and ZTE published a press release citing IDC’s estimate of its own shipments.

But most glaring of all is the absence of any mention by Samsung of its performance. The company stopped reporting any data on either overall phone shipments or of smartphones within that total since Q3 2011. We may have been able to estimate Samsung if we had more competitor actuals, allowing us to back into a figure. But we don’t.[1]

Which leaves us with IDC’s estimate[2]. But the problem is that since the industry is growing so quickly it is very sensitive to assumptions. Consider how difficult it is for consensus estimates for Apple’s iPhone shipments to come near the actuals–and that’s for one quarter, and knowing all the previous quarters with precision. The absence of visibility into the assumptions made by market analysts (or their methods) should lower confidence in the results.

Consider that IDC specifies 50.2 million Samsung smartphones, implying an accuracy down to 100,000 units. Is this accuracy believable?

There is reason for doubt.

I start with the following graph:

It shows the estimates for Samsung global shipments (latter four quarters are estimates without the benefit of company reports.) I separated the component of shipments that Samsung reported as part of its submission to the Samsung v. Apple trial in California. Note that this (orange) segment consists of most US smartphone shipments. It’s not all US Samsung shipments. It notably excludes the Note, the latest Nexus and the Galaxy SIII.

Without these “non-accused” phones, it would appear that Samsung’s US volumes dropped to 4% of global totals from a 10% average[3]. In order to return to 10%, the US totals would need to be adjusted as shown below:

 

I added 1.5 million units in Q1 and 4 million units in Q2 (to account for Galaxy SIII).

So far this seems consistent with the total estimated global shipments of 50.2 million from IDC. US units would result in 12% of global shipments.

However, there is more data we can test against. First, we are also given pricing for the US phone portfolio, or at least we have revenue per unit sold. This allows us to get an idea of what revenues can be obtained for the smartphones in the US, and possibly world-wide[4]. Knowing prices for smartphones and knowing global sales (which Samsung reports) and finally knowing global total phone shipments allows us to price the average non-smart phone.

Excluding the non-accused smartphones, the average revenue per smartphone in the US for Samsung in Q2 was $339. Adding in the estimated Galaxy SIII (4 million) at an average price of $470 gives a new “ASP” of $413.

This is where the analysis goes off the rails.

If we assume the global ASP is equal to the US ASP and knowing everything else about the mix of smartphones (54% of total) and revenues (as reported) yields a negative ASP for feature phones. Assuming the revenue figure is correct, either the number of smartphones is wrong (50.2 million) or the global ASP ($413) is wrong. The total phone number is also an analyst estimate which can be called into question.

There are many assumptions being multiplied here so I have very low confidence in the result, but we should ask “what do you need to believe” in order for the 50.2 million figure to be real. The following scenario for example has to be true:

1. Global Samsung phone shipments total 93 million units (which means 42.8 million feature phones sold as well as the 50.2 million smartphones).
2. 6 million smartphones shipped into the US, equivalent to about 50% increase y/y, roughly in-line with market growth
3. 4 million Galaxy SIII shipped (but not necessarily sold through) at an ASP of $450. This results in a US total smartphone ASP of $413
4. The global smartphone ASP is significantly lower (at 80% of the US or $330)
5. Global revenues of $18.14 billion is essentially all from phones
6. Global feature phones will have an average price of $36

If these conditions are met then it’s possible that 50.2 million Samsung smartphones did ship.

The figure for feature phone ASP is difficult to accept as it would be lower than Nokia’s. Also, the low global smartphone ASP is also difficult to accept but it might be possible to find a scenario with the above variables which makes the 50.2 million figure believable.

Unfortunately this is an unsatisfying analysis. I could propose another figure to answer the question in the title, based perhaps on more “believable” assumptions. But I have no means to defend my assumptions. The only option I have remaining is to caution that the total smartphones shipment estimate for Samsung remains precarious. The new evidence from the US trial places more doubt on it but it is not enough to disqualify it entirely.

—

Notes:

1. This cone of silence over 93% of the most popular mobile platform is somewhat puzzling in contrast to the perfect information we are used to having about iOS, BlackBerry and even Symbian. It seems that the more open a platform, the less is revealed about its market performance. This extends to the licensed Windows Phone whose shipments are only revealed through Nokia’s reporting of Lumia shipments.
2. In the PC market, vendors also selectively publish their units. The norm is, however, not to. As a result IDC et. al. are the only sources for global market data. However, the PC market can be sampled more easily than the phone market. PC retail measurement methods are more reliable compared with all the points of sale for phones.
3. Note that Apple sells about 30% of iPhones in the US.
4. Apple’s pricing world-wide is very consistent with pricing in the US which makes sense from a channel management point of view. One would not want to create incentives for cross-border arbitrage of phones.