Within the past couple weeks, we have had two major Web browser releases. Microsoft unleashed Internet Explorer 9, and Mozilla officially launched Firefox 4. Each browser had millions of downloads on its inaugural launch day. Firefox appears to have won the launch-day contest, but if you dig beneath the stats the battle was perhaps much closer than sensational news headlines might suggest.
Microsoft reported that Internet Explorer 9 was downloaded 2.3 million times during the first 24 hours. Not too shabby. But, Firefox 4 reportedly claims as much as 10 million downloads on its launch day...sort of.
Let's assume you are in a race to see how many paper airplanes you can fold in an hour. Does it seem reasonable to count the planes you started folding, or thought about folding -- but didn't complete by the time the hour is up? Or, how about if we count all of the planes you folded in the last week or so? No?
I imagine that most people would cry foul. It seems fairly simple and straight forward to say that the only planes that would count for such a contest are the planes that were started -- and finished -- during the hour in question. So, if we are measuring the number of downloads for a Web browser in the first 24 hours of its launch, shouldn't we also only count the downloads that are initiated -- and completed -- during the 24 period that begins with the time of the browser's official launch?
Mozilla is quoted in an Inquirer article stating, "Within 24 hours of Firefox 4 being announced it had been downloaded 7.1 million times, this is in addition to the more than 3 million people who were already running the release candidate that became our final version."
Hold up. So, Firefox 4 was downloaded 7.1 million times during the 24-hour window, but Mozilla is claiming 10.1 million by throwing in the RC downloads as well?
Well, that changes things. Microsoft does not have a breakdown separating RC downloads from the other beta versions, but IE9 had more than 40 million downloads prior to launch day. There were three public betas plus the release candidate, so...divide by four...10 million. OK. Using the "new math" where we count RC downloads in addition to the actual 24-hour count, I imagine there were roughly 10 million downloads of the IE9 release candidate, so we'll tack that on. That means IE9 had 12.3 million downloads and beats Firefox 4's 10.1 million.
Yes, I know that my math is silly. So is Mozilla's. The thing is, 7.1 million is already more than 2.3 million, and it is impressive enough on its own. Why bother jumping through math hoops and resorting to statistical sleight of hand to make it seem like more than it is? Perhaps Mozilla has a self-esteem complex about the fact that Firefox 4 fell below the 8 million mark set by Firefox 3 on its launch day?
One other small thing to consider -- many Firefox browsers are set to automatically update, or at least notify the user that an update exists. If Internet Explorer 9 were automatically installed to update all Internet Explorer 8 browsers, or if Microsoft pushed it out automatically with Windows Update, that first day figure would look much different.
Finally, consider this. Firefox is available for Windows...and Mac OS X...and Linux...and even FreeBSD. Firefox is essentially available for 100 percent of the PCs out there. IE9 is only compatible with Windows 7 and Windows Vista, so the potential pool is restricted to about one-third of the total PCs. If you view the launch day stats through the lens of the available audience, you have to divide the Firefox number by three -- or multiply IE9 by three -- to compare them on equal footing. In that light, Firefox 4's 7.1 million becomes 2.36 million and makes the two browsers more or less equal.
A Microsoft spokesperson also explained to me that, "Only about 10 percent of the 2.3 million downloads of IE9 RTW in the first 24 hours came from the existing install base, meaning it was largely people coming from another browser."
Launch day claims make for good headlines, but let's check back in a few months and see how the two compare.