July 31, 2009

why we count

The 1 billion Firefox downloads celebration has generated a lot of buzz and some questions around browser metrics. Rather than addressing each question individually, I figured I'd just wrap them all up in a single post that I hope explains why we count things.

Right now, we keep a pretty close eye on three global measurements. The first, and all the buzz today, is the Firefox download count. The second is active Firefox users. And the third is market or usage share.

We count downloads, and we're celebrating 1 billion downloads today, not because it's a good proxy for how many people use Firefox or how much they use Firefox. It's not. We count and celebrate download milestones because it's something our community of advocates can directly contribute to and see the actual results of their work.

Firefox doesn't ship on every new PC like IE or on every new Mac like Safari. Those guys don't actually have to do any work at all to put their browser in front of users. That's not the case with Firefox. Virtually all, more than 95%, of our growth comes from people downloading Firefox.

We've never had the budget to rely on advertising as a primary driver of Firefox downloads the way other successful software companies do. As a matter of fact, more than 80% of our users say they found out about Firefox from a friend, co-worker, or family member.

So, people helping other people download Firefox is critical. It's how we've achieved the success we have. Take that away and there is no Firefox phenomenon. In that light, it would be a huge insult to all of the Firefox contributors that have been helping other people get Firefox if we didn't celebrate their work.

The second number we watch is active Firefox users. This is a bit more difficult because Mozilla has never and never will track individuals using Firefox. At the same time, it's fundamental to our mission that we empower people with more control over their Internet experience. If we don't keep an eye on how many people we're actually helping, it's somewhat more difficult to know if we're getting better or getting worse.

So while the methods are indirect, we have come up with a pretty good way to get an approximate number. Every Firefox browser makes a security check with Mozilla each day. This mechanism is fundamental to keeping users secure, but it can also tell us how many users were running Firefox in a 24 hour window. From that number, we can extrapolate how many people use Firefox in a given month.

Today we have about 300 million active users. That's actually a pretty large percentage of all of the people on the Web and we're rightly proud of that because it says that for 300 million people and growing, our mission of promoting choice and innovation and empowering participation is succeeding.

The final number we watch is called either market share or usage share. I think "usage" is a better way to describe it so as to avoid confusion with the more traditional definitions of market share which are usually measured quarterly and in terms of sales.

Usage share, more precisely, the percentage of Web traffic that comes from a particular browser, is reported by several large analytics companies and widely reported on in the press. For browsers with increasing share, it can have a positive impact on raising awareness, but more important, I think, is that it helps us understand if we're actually improving the Internet as a whole.

The Web moves forward when Web developers believe that a new Web capability is widely deployed enough to start building sites and apps that can take advantage of that capability. It's not enough to know how many users a particular browser has, though. The Web is growing by about 100 million new users a year so you actually need a percentage here. Sites measure that kind of platform deployment with analytics packages that can identify different browsers and what percent of their traffic comes from those different browsers.

But not all projects have representative sites where they can make those measurements themselves so they turn to the analysis from companies like Net Applications and StatCounter. These companies aggregate website statistics from all over the globe and publish some or all of that data publicly.

We look at the numbers for almost the same reason as a Web developer would. We're interested in how much of the Web's traffic comes from people who are using a modern, standards-compliant browser. Because we're not just about improving the online experience for individuals, but also advancing the entire Web platform, the bigger that percentage, the better we're doing in our mission.

So, each of these measures has a somewhat different purpose, but all of them help us measure and celebrate our accomplishments and taken together, and over time, they help us understand what kind of impact we're having on people and the Web.

Posted by asa at 2:45 PM