A rather dubious report has been circling the internet today, and since it’s been picked up by several large news providers, I thought it merited commenting on.
Facebook is not failing in Canada, despite what these misleading stats would have you believe.
The articles use stats without enough context and leave out key information to sensationalize a story. Aside from general seasonal fluctuations that have been commented on by others already, the numbers they are using are misleading or flat out wrong.
The numbers being quoted are from Facebook’s self-serve advertising interface, and represent total active users within the last 30 days; it is not a accurate list of membership. These numbers are estimate at best, as when you apply targeting parameters such as geographic areas, the reach in a given area will often exceed the actual population. This number represents all active profiles, included multiple profiles, and doesn’t differentiate between inactive users and deleted profiles.
In Canada, ComScore is the primary measurement tool for measuring online activity and it has reported growth in Facebook’s reach in Canada every single month since May 2009 (as far back as I have info on file). Their numbers actually exceed Facebook’s profile count, as they estimated a reach of 21.1 million users in May (84% of the estimated online population). The extra traffic would be non-registered users that still end up viewing Facebook pages (fan pages, pictures linked from other sites etc.)
The most glaring problem I have with this article is that they are using a slow down in new user growth as an indicator that the site is losing steam. As mentioned, Facebook already reaches 84% of Canadians, growth has to slow as a market reaches total saturation. Countries with high penetration simply don’t have any population left to add to the numbers. Look at Reddit.com in comparison: in Canada the estimate reach of Reddit increased by over 450% since March 2010 (noting that the traffic does swing wildly month to month), in the same period Facebook grew by only about 4%. Yet Facebook’s actual number increase was 937,000, compared to reddit’s increase of 345,000. Yes, Facebook’s percentage growth is small, but the actual increase in traffic is huge.
There are also glaring omissions from the source article when reported by newspapers:
“Bugs in the Facebook advertising tool that we draw this information from, seasonal changes like college graduations, and other short-term factors, can influence numbers month to month and obscure what’s really happening.”
…and this bit, covered by the LA Times, but ignored by Yahoo:
“Still, by the time Facebook reaches around 50% of the total population in a given country (plus or minus, depending on internet access rates in that country), growth generally slows to a halt, as we’ve noted before.”
That million user drop in Canada?
“Canada also fell significantly, by 1.52 million down to 16.6 million, although it has been fluctuating around that number for the past year.”
Facebook is an easy target and has generated a lot of backlash on privacy and security issues in the past; news sources know it makes a great headline. However, these articles are entirely based on a single report based on numbers generated by a self-serve advertising platform that is known to be a broad estimate at best. The source article admits these numbers are prone to fluctuation and even opens their article with the following quote, omitted by the newspapers:
“As we note below, we’ll need to wait to see what the long-term trends really are before knowing if Facebook is continuing to grow in the US and other countries.”
That omission speaks volumes about the quality of this content.
As a parting item, the poster of the original report has a follow up article today: http://www.insidefacebook.com/2011/06/13/available-data-shows-facebook-user-numbers-growing-quickly-or-slowly-or-falling/ This article indicates that other reporting source cannot corroborate the initial findings, and ComScore number in the US match the trends I’ve mentioned for Canada.