Almost every quarter, without fail, you hear through some Mac news website that some data collection service has determined that Apple's market share for personal computers has increased and OMG APPLE IS NOW THE THIRD LARGEST COMPUTER MANUFACTURER. Or perhaps every month, you hear through some Mac news weblog that some other company that focuses on data collection reports that Safari's market share for web browsers has decreased sequentially month-over-month and OMG APPLE MUST BE FAILING AND THEY ARE DOOMED.

The problem is, you hear various market share numbers thrown around from a couple of different sources, and they all get their own data and they all come up with their own figures, and it's really hard to tell what's accurate and what's not. Things are especially complicated when the personal computer market is segmented, either between consumers and enterprise, laptops and desktops, etc. Ostensibly, Apple has both a larger share of the market in laptops as compared to desktops, and in the consumer market as opposed to the enterprise market (and Apple certainly has a broad presence in education).

Comparing Apple's current market share to its historic market share is also very problematic, because comparing numbers from two different sources probably doesn't make too much sense, and the data in previous years is spotty enough that you may be forced to do that if you want to draw any conclusions whatsoever.

So how do you make sense of all the data? For example, this month's pet market share number for Apple is 8.1%. Oh, wait, no it's not, it's 6.3%. Question: who's right? And more importantly, is Apple's market share really increasing dramatically like everyone says it is?

Well, I took a little time today to do some compiling of data from a bunch of different sources, and created two neat little graphs of Apple's U.S. and worldwide market share. You can also download the raw data itself. But first!

Some caveats about the graphs and the data: many of these numbers are extrapolated from the scant info I could scour from the web; this is on top of the fact that much of this data is collected by companies that are also estimating unit sales numbers for Apple and for the worldwide market. There is a high (and un-calculated) margin of error, here.

Any cell that's highlighted in red in the spreadsheet is data that's been calculated using other numbers (which are possibly averages), and the data in these cells should be considered less reliable compared to the data in other cells. Cells with a blue background are data that's taken directly from the source (Apple, IDC, Gartner), or calculated directly from data from the source.

Because all this info has been compiled from lots of different sources (MacRumors summaries, Macworld articles, IDC press releases, Apple press releases, etc.), you might think it'd be hard to verify where I got these numbers from. Not so! I've hyperlinked the values in each cell in the downloadable data files so that it takes you directly to the webpage which gives you the value itself or the data used to calculate that value. The text doesn't look like they're links, but trust me, they work -- even in Preview. Your cursor will change to the Mickey Mouse pointing hand when you mouse over them.

So, with that out of the way, here are the graphs (click for larger versions):


You can download the spreadsheet data that created these graphs (in order of decreasing preference of file format): Numbers, PDF, Excel.

For the U.S. market share, there are three values: Apple's share as calculated solely from IDC's numbers, solely from Gartner's numbers, and from both Apple's and IDC's data. This last set of data is highly experimental: I took Apple's published worldwide unit sales from their press releases, multiplied it by one minus the percentage of revenue for that quarter that was derived internationally (to estimate the number of units that were sold domestically), and then divided that by the number of computers shipped in the U.S. by all vendors according to IDC. Compare this to IDC's actual numbers.

What's interesting is that you can see a real uptick in Apple's market share starting around October 2005. I had always heard that the Mac had been gaining for several years now, but it seems that the real catalyst for the market share increase was the switch to Intel processors (whether directly or indirectly).

Worldwide market share is much more spotty, especially because available data doesn't allow the calculation of worldwide market share from 2003 to 2005 or so. The main dataset here is derived from both Apple's and IDC's numbers, so it's dubious as to how reliable it is. The 2001-2003 data is also extrapolated from averaged yearly unit sales numbers (instead of quarterly sales numbers), which makes that section especially unreliable. Also, Net Applications' data is measured by web traffic and not by sales, so while it's worldwide market share, there's likely a bias in favor of Apple due to the proportion of computers bought that are actually regularly used to surf the internets.

However, from Net Applications' numbers we can still see that there's been a noticeable uptick in market share beginning around the same time as the U.S. market share (October 2005), and especially so once Apple's product line was completely converted to Intel processors (October 2006). You can probably convince yourself that a bump is also happening in the worldwide figures derived from both Apple and IDC's numbers, but it's ever so slight if it's actually there.

The conclusion I draw from this exercise is this: Apple's market share has been and is still increasing, but it's pretty much only been doing so in the last two years, and not so much before then.

Lazyweb request: if you're interested in fleshing out this data, I would greatly appreciate pointers to places that would fill in gaps in my dataset. I pretty much need any quarterly IDC worldwide/U.S./Apple unit sales numbers for FY05 and earlier, and pretty much any Gartner data you can get your hands on.