Lies, Damn Lies, and Statistics

Amazing how a 15% market share growth in five weeks can be interpreted as bad news

If Firefox was growing linearly, the percentage growth over constant time would be expected to decrease as the installed base got bigger. If the percentage value was maintained, we’d in fact be growing exponentially. The truth is probably somewhere between the two.

7 thoughts on “Lies, Damn Lies, and Statistics

  1. Well yeah, but Firefox is growing at something around that rate is old news. Firefox growth slowing is a different news story, and therefore more likely to get picked up. If you’re doing regular press releases about stats (as websidestory are), you have to say something different each time.

  2. on Onestat:

    4. Netscape (1.11 %)
    5. Opera (1.09 %)

    Netscape usage higher than Opera O_o; damn, that’s quite a kick in the teeth for Opera – that they can be beaten by a POS like NS >_

  3. If you number-crunch their stats *without* assuming exponential growth, you get steady linear growth from early december to late february, albeit at a lower rate than in november (but then, november had 1.0 and a big publicity push)

    My numbers are here.

  4. It depends on whether the browser market is a closed system or not.

    If we have an open system and IE’s number of users stays constant, then firefox’s browser marketshare would be x/(x+a) where “x” is firefox’s number of users and “a” is IE’s number of users. It’s clear that each new marketshare percentage point corresponds to more actual users than the last, so you would expect the percentage growth rate to drop.

    (it is also worth noting that the growth will have to stop at some point — if IE has a fixed number of users, Firefox would never be able to reach 100%, no matter how many users it had).

    Now if you consider the browser marketplace as a closed system, then every new Firefox user corresponds to one less user for another browser. In this case, a slowdown in growth of marketshare would correspond to a slowdown in growth of users.

    Without knowing the details, it is a bit hard to tell which model is correct. You’d need to know if the number of web browser users is changing or not.

  5. James: Globally, it’s open. The number of internet users in the US and some parts of western europe has (more-or-less) stabilised, but its still growing everywhere else in the world. What I’d really love to see would be some sensible browser stats out of (say) India or China.

    In terms of growth rates, however: go read my piece – I calculated their growth percentages from their market share figures, and the formulae they are using have a built-in assumption of exponential growth. To them, growth from 4% to 5% (25% growth) is slower than from 3% to 4% (33% growth) in a similar period. Except their periods aren’t that similar, either.

    On their figures, firefox’s market share grew at an average of 0.021% per day from December to February.