Strategy Analytics released numbers that clearly show Samsung beating the pants off of Apple in smartphones and closing the gap on tablets. But it is getting pounded by some because these numbers may not be reliable, and given that many may short Apple's stock as a result, this has broad implications. I believe Samsung is outselling Apple, but Apple is making far more money off its phones than Samsung and that is the more important metric. However, I also agree the numbers aren't accurate, but I don't wholly agree with this post calling them out. Confused yet?
The fact that the comparative numbers are sourced differently would be enough to render them, for comparison purposes, invalid. Now this doesn't mean you can use them to show that Apple is beating the pants off of Samsung. Both companies' financials do support that both are doing well in this market. However, being in first place or dropping to second place can have an impact on how companies are valued and if the numbers aren't accurate but believed, they still will have a material impact on the stock market and perhaps even sales (people tend to gravitate to products that appear to be very popular).
Let's talk about the mess we are dealing with here. It isn't Strategy Analytics' fault; it is simply dealing with a mess as best as it can.
Shipped vs. Sold
In the referenced piece above the author takes exception to the concept of shipped vs. sold and while he has the right idea, he gets it wrong. When a manufacturer ships something, it is sold. It is sold to the retailer, but it is still, from the standpoint of a Samsung or Sony, sold. Apple, which has its own stores, can ship product to those stores before it is sold. In its case, shipped is not sold because it sells to the end user. So Samsung can use shipped to mean sold, but Apple can't because of the Apple stores, making the same term used for both vendors misleading.
Sell To vs. Sell Through
A number of us have been trying to get the terminology changed so that it is more accurate. "Sell to" means sell to channel or sell into retail (same thing said two different ways). This is more accurate than shipped because it focuses on what retailers bought, though still a problem for Apple due to its stores. "Sell through" is what the consumer has bought and is the best way to measure the actual popularity of a product.
For instance, when Microsoft's Zune initially came to market, its sell-to numbers were very strong because retailers didn't like the deal they had with Apple or that they were competing with Apple stores with Apple products and bought a lot of Zunes. However, sell-through numbers were horrid because consumers didn't buy them. By looking at both numbers you could see a major inventory problem and, as an investor or a consumer, reach the conclusion that Microsoft had a problem and that there would likely be some great deals on Zune hardware in a few weeks.
To me, that is what reports should be capable of doing: helping people make informed decisions. So why don't they?
Vendors and Numbers
I'm one of a handful of analysts who have been formally trained in market research and analysis and I won't touch numbers with a 10-foot pole. I started out doing numbers for Dataquest (now a part of Gartner) with the goal of making them accurate. In a few months, I became famous and was the only analyst who was told that he would never work as an analyst again by virtually every major technology CEO in the market but one. I was even called out on the stage as being an idiot even though my forecast was, in fact, the most accurate. This is because the vendors fund numbers services and accuracy is only the friend of the guy in front.
Forecasts better show everyone gaining share or the folks on the downward slide will cancel their services and try to have you fired. Add to this that the vendors often put someone who is ethically challenged in charge of supplying the numbers (I found one guy who took his company's actual numbers and multiplied them by three). Yes, in most of the reports, the source for the numbers is some unnamed guy the research firm calls and asks, "So, hey, how much did you ship/sell last month?"
And it isn't as if these mystery employees (who are never called out) give you the real numbers anyway. You end up playing this "bigger than a breadbox" game where you guess and they say something like "close enough." I have for years been waiting for the SEC to step in and have what amounts to a hissy fit on the entire process if only because it violated disclosure rules to high heaven.
Interestingly, only Apple would basically say, "If you want my numbers, read the published financials because that is all you are going to get." So if this thing ever blows up, I expect only Apple and anyone else taking this approach to emerge unscathed.
All Numbers Aren't the Same
NPD, for instance, actually surveys retailers, which are arguably a better way to capture sales; however, it doesn't measure all retailers and it doesn't measure worldwide. For smartphones and other connected devices, there may be ways to automate the devices that are actually turned on much as we measure browsers by Web activity. This isn't yet ideal because each Web method is typically tied to sites that aren't universally used, which introduces bias. But this approach does seem to remove the tendency for a vendor to manipulate the result.
Wrapping Up: Take Numbers with a Grain of Salt
Or maybe a quart of whisky because the numbers will generally imply things that can't be proven. Many will argue that they are better than nothing, but I'm not in that crowd. I believe there should be, like there is for any other financial report, a requirement that these numbers be accurate and there should be an auditable disclosure process that assures accuracy. Eventually some pissed-off investor or government agency will blow this process up and force accuracy into this process. Until it does, just figure you are reading some level of fiction and stay as far away from numbers as you can. That's what I'm doing ... well, sort of.