Life after Launch of Firefox 3 – Revisited

Ken Kovash

8 responses

A couple months ago, we speculated on what the Firefox 3 universe would look like approximately one month after launch.  In that analysis, we took our download and usage numbers from the time period following Firefox 2’s launch in 2006 and extrapolated from there.  We’re now exactly one month removed from Download Day, so we wanted to update that discussion and see how close our guesstimates were for Firefox 3.

Specifically, how many cumulative downloads of Fx3 have we seen since Download Day and were does the “Active Daily User” number for Fx3 usage stand today?  Below you’ll see this snapshot analysis, along with a comparison to the launch of Firefox 2:

The numbers above match fairly well with our original predictions.  So, how else can we interpret these results?  One pattern that stands out is the conversion percentage, i.e., how well downloads are translating to active daily users.  Dividing the “red” number by the “blue” number, you’ll see that the conversion rate was approximately 38.5% during the first month of Fx2 and has been approximately 33% during the first month of Fx3.  This could merely be a seasonal effect… or perhaps this tells us that there’s a little room for improving the user’s experience from download to install and from install to long-term usage.

8 responses

  1. Charles Broam wrote on :

    You need to keep getting more numbers longer. The same time that FF3 came out, so did Opera 9.50. While you came out with FF 3.0.1, Opera delivered 9.51.
    People just want to give you a fair shake, and at the same time they give other browsers a fair shake as well. Be patient, and keep collecting that data. The long term answer will be the real deal. By the way, some people like myself prefer the SeaMonkey Suite (keyword here: “Suite”). I couldn’t wait for SeaMonkey 2.0 so, I already have SM 2.0 pre-alpha 🙂

  2. Dorus wrote on :

    You shouldn’t confuse opera’s market share with that of firefox, considering firefox got about 25 times more users. I think the biggest difference between active users and download come from people downloading firefox multiply times. I myself downloaded firefox 3 both at my laptop and my own computer, and might even have download it 1 more time when i switched from using firefox 3 as a test browser, to mainbrowser, when i reinstalled it (first i didn’t had it installed in program file), and finding the installer back on my disk would take more time than downloading it again.

    Anyway, the growth in active daily users look nice. What i’m missing is the amount of daily users still using firefox 2 a month after launch of firefox 3, and the same for firefox 1.5 users after 2.0 launch. I can imagine these numbers will be quite big considering auto update is not yet turned on for mayor updates at that time.

  3. Jesse Ruderman wrote on :

    Maybe Firefox 3 has a lower conversion percentage because a lot of people who aren’t already Firefox users are trying it 🙂 Are you able to break the statistics down by the browser used to download the new version?

  4. kkovash wrote on :

    Jesse — great observation! for the “cumulative downloads” from the chart above, I looked at what percentage of those totals were attributable to IE and Safari users. For the Fx2 launch, 48% of those downloaders were IE+Safari users, and for the Fx3 launch, 51% of those downloaders were IE+Safari users. So, you are correct: Fx3 saw more people who weren’t already using Firefox trying it.

  5. Havvy wrote on :

    Hey, yet another thing that can skew your results: Download Day

    Eight million times Firefox was downloaded to create a record. How many times was is downloaded multiple times by one person to bump up the results?

    Anyways, all the other results I’ve seen have been looking nice.

  6. Jackson wrote on :

    “there’s a little room for improving the user’s experience from download to install and from install to long-term usage.”
    The best way to do that is to encourage users to provide feedback.
    I’d expect that most users who stop using Firefox are the sort of people who’d simply prefer using, say, IE.
    I think Mozilla needs to find ways to ask these people what features they like in their preferred browser, and/or why they didn’t like using Firefox. Find new ways, because these people won’t be interested in going to feedback.mozilla.com.

  7. fcp wrote on :

    Hi, this blog is very interesting!

    I have one question. How do you count the “Active Daily Users”? If you count the accesses to the Application Update Service (AUS) as explained on http://john.jubjubs.net/2007/11/27/mozilla-firefox-market-share/ , I am afraid that the observed number might be much smaller than the actual number because of Bug 407875: https://bugzilla.mozilla.org/show_bug.cgi?id=407875 . Bug 407875 states that if Firefox is installed as a privileged user and run as a non-privileged user, the user is not notified of updates in some environments. Bug 407875 is about notification of updates, and I do not know whether it affects AUS pings, though.

    If Bug 407875 affects AUS pings, then it is probably bad for the statistics because Bug 407875 affects some systems but not the others, depending on the OS and the Firefox version. For example, on Windows Vista, it affects Firefox 3.0.1 (and I believe also 3.0.2 and 3.0.3) but not Firefox 2 (I do not remember the minor version, probably any).

  8. kkovash wrote on :

    fcp,

    good question! Your guess is correct. The numbers we used here were from AUS. We realize that that count is far from perfect and simply provides us with a proxy for understanding the size of the Fx user base.

    In regards to the bug you point out, those commenting on the bug are much more qualified than me to discuss the particulars 🙂

    Ken