How Many Firefox Users Customize Their Browser?

There’s a relatively simple question we’ve been asking for quite some time… of the total Firefox user base, how many people (or what percentage) use at least one add-on?

blog-firefox-bundle

Thanks to some recent work by Simon Krueger, our metrics intern, we’ve been able to arrive at a rough approximation.  Here’s my thought process for the important data points Simon’s been able to extract:

  • We know that there are roughly 300 Million Firefox users, and on a given day, we see about 100 Million active users of Firefox
  • Narrowing our focus to just a single day’s worth of data (we simply picked June 22nd), we turned our attention to estimating add-ons usage for just that one day, so that we could compare something on a relative basis to our overall Fx metric of 100 Million active daily users
  • We saw Firefox add-on pings from 32.8 Million unique IP addresses
  • There are a couple issues at this point: (1) There could be multiple users from the same IP address, and (2) There could be a handful of add-ons where adoption isn’t fully user-initiated (e.g., people happen to have it, but don’t know what it is)
  • Balancing these two issues (i.e., they may come close to offsetting), I think it’s safe to say that roughly 33% of Firefox users (32.8/100) have at least one add-on
  • So, of our user base (roughly 300 Million), perhaps about 100 Million people use at least one add-on

I want to conclude with a few words of caution.  This methodology is far from perfect.  It’s conservative and there are also potential holes in it (some of which we haven’t yet thought of).  It also remains unclear as to how useable this slice of data will be as we see it trended in the future.  Moreover, the data above is wide open to interpretation.  The add-ons team and other folks across the community may arrive at a completely different interpretation (read here for Justin Scott’s point of view).  Regardless, a fruitful discussion around this topic should hopefully ensue.

(photo from mynameisideal.blogspot.com)

8 responses

  1. Greg K Nicholson wrote on :

    Wow—the Firefox 0.8 logo! Old skool.

  2. Tony Mechelynck wrote on :

    Only 33%? That sounds indeed conservative to me; but of course I don’t have the means for a scientific counter-experiment. I’d be interested in how many addons those 33% have (on average, or as a frequency distribution) but I don’t know how one would go about measuring that.

  3. Asa Dotzler wrote on :

    One thought.

    If among our 300 million users there is a group that opens its browser every day and a group that only opens its browser only every few days or only on weekends, it might be that those groups are very different kinds of users.

    If we look at just one day, we might be seeing mostly people who open their browser daily and those could be more “power user” types who would be more likely to use add-ons.

    Is it possible that we’re seeing “most” add-on users in a one day sample but only about 1/3rd of all Firefox users? If that’s the case, then we’re talking about more like 10% of Firefox users being add-on users and not 33%

    Or, another way of saying this is that the ADU to total user conversion might not work for add-on pings.

  4. Erwan wrote on :

    I guess there’s one more issues – addons with a custom update URL, that won’t ping AMO for their update. It’s probably not much, but there may be some missing addons users there.

    Another potential bias is that of the 300 millions, addons users could be over-represented in the daily users if they are more active than others.

  5. Tom wrote on :

    What??? I diden’t understand the math, can you just contact mozilla and ask them?
    For a bonus here is my favorite add-on
    https://addons.mozilla.org/en-US/firefox/addon/10548

  6. Kurt (supernova_00) wrote on :

    What about users that have Firefox set to not check for extension and theme updates?

  7. Marco wrote on :

    I believe more active users are more likely to install add-ons. In other words: for the 200 million that didn’t show up June 22nd I expect a lower percentage number of them to have add-ons. Therefore in my eyes extrapolation doesn’t apply.

    Now you can either prove that your assumption was right – or find a correction estimate for my assumption.

  8. Marco wrote on :

    @Asa Dotzler: sorry, I was paraphrasing your post #3 in #7 – but when I wrote mine only #1 & #2 were displayed and the rest invisible (probably waiting for “editorial approval”…)

    @admin: if you want to spark a discussion between your readers you need to change that. (if your fear is spam: check Mollom for help)