Comscore's Fishy Facebook Numbers
Comscore published some web site usage metrics for Facebook yesterday. While the aggregate visitor growth of 89%, year over year, from 2006 to 2007 was impressive, the demographic information was puzzling, and just doesn't add up. … Age Segment — Facebook.com
Context & Ripple Effects
Comscore reported a flood of new Facebook traffic — 89% year-over-year visitor growth following the site's open registration — and within a day blogger Alec Saunders flagged that the accompanying demographic breakdown simply does not add up. The dispute lands at a delicate moment: Facebook opened its platform to third-party apps in May 2007, spawning both an active market debating what an app is worth and commentary comparing Facebook to AOL.
First-order effects
- Advertisers and media planners buying against Comscore's Facebook age segments are working from numbers one analyst calls internally inconsistent, which puts ad-targeting decisions on the site on shaky data.
Second-order effects
- Comscore's rivals in web measurement get an opening to attack its panel-based methodology just as Facebook's post-registration traffic surge makes accurate demographic reads more valuable — and harder — than before.
Third-order effects
- If open-registration social networks keep breaking the assumptions behind demographic panels, the measurement industry faces pressure toward census-style, server-side counting over sampled panels — a methodology fight that will shape whose audience numbers everyone trusts.
The trend: As social networks blow past the population samples used to measure them, third-party traffic data is becoming the contested ground between platforms, agencies, and the measurement firms themselves.