Does sampling user data give an accurate picture of traffic?

If you've ever played the web traffic game, where big numbers can mean big advertiser dollars --and small numbers can mean going bust -- you're probably keenly aware of the lousy accuracy of "panel" based sampling methods used to rate and rank sites against each other.

BusinessWeek takes a look at how inaccuracy means trouble for some web upstarts, and at the ongoing struggle by some to straighten out the numbers game. Digg, for example, recently hired heavy-hitting web analytics firm Web Side Story to get to the bottom of a massive gap in their own numbers and those presented by traffic rating firms. Digg's own data showed 15 million uniques per month, while comScore put them at 1-2 million unique visitors for the same period. The difference in ad dollars for a gap like that are enormous, but percentage-wise they can't begin to compare to the plight of many smaller web-fish. "For many, traffic numbers decide not just the price of an ad but whether the advertiser inquires about an ad at all"

Tags: alexa, digg, neilsen, news, web traffic, WebTraffic