Comscore has finally caved to a decade of complaints from web publishers that its traffic estimates woefully undercount site traffic, thus producing a vast disparity between Comscore numbers and the numbers the sites have in their logs.
Now, Comscore is finally going to change the way it counts traffic, thus, hopefully, bringing the numbers more in line with reality.
(Comscore extrapolates overall usage from a relatively small panel of web users, which accounts for part of the difference. The rest probably comes from the same user using multiple browsers
to access a single site, which boosts the numbers in site logs).
For years, Comscore has insisted that it is right and that sites are delusional. But now, under pressure from competitors Quantcast and Google, which offer free tracking pixels via which sites can have their site-side traffic counted, Comscore is rebuilding its measurement system. Henceforth, for some sites, it will use a "hybrid" model of panels AND site-side tracking pixels.
That is an admission that sites might have been right all long. (Or at least more right than Comscore). But it's not what's infuriating here.
Comscore will give the tracking pixel away for free to paying clients, sites that pay thousands of dollars a year to get its detailed traffic information. All other sites, however, will have to pay Comscore $10,000 or keep having their traffic counted the old way.
Sites that can't afford Comscore research (it's extremely expensive) will have to fork over $10,000 merely to have their traffic counted on an apples-to-apples basis with that of their richer competitors. Because most sites think Comscore grossly under reports their traffic, this amounts to Comscore saying "If you want your higher traffic numbers, you need to pay us $10,000."
That's bad business. And it will make those who have already embraced Quantcast, Compete, Google Analytics, and other more reasonably priced traffic measurement services hate Comscore even more.