Down East 2013 ©
Hits and misses: My recent posting  on the top twenty-five news websites in Maine generated a lot of criticism from those who claimed their sites’ unique monthly visitors were undercounted by Compete.com’s  methodology.
They appear to have a point. But it also appears that while those monthly visitor figures may matter to their advertisers, they don’t make much difference in terms of their rankings among the top sites.
Most of the websites that complained cited visitor figures they had obtained from Google Analytics,  which showed consistently higher numbers, often double those from Compete, particularly for smaller sites (those with fewer than 100,000 unique visitors per month). The bigger the sites in my limited sample, the more closely the Compete and Google numbers tended to correspond.
Compete’s website has an explanation for the variations. It says its methodology  calls for excluding international visitors (which could have a major impact on Maine sites that appeal to Canadians). It also doesn’t rely on cookies to determine who’s visiting on multiple occasions. According to Compete, “Because of cookie deletion, return visits by the same person (with deleted cookies) wrongly appear to be a new unique visitor. In addition, if cookie implementation on the server side is done incorrectly with vague or inconsistent definitions, visitors will be overcounted.” Compete also filters out “spiders and bots” that can significantly inflate figures without any actual human involvement.
Is this more accurate than Google or other rating services?
I haven’t the faintest idea. I spoke with two experts (who asked not to be named because they do some work for companies involved in web-visitor measuring). They offered contradictory opinions on the value of Compete.com, although they agreed the numbers may have some uses for particular sites. They did agree that virtually all web ratings, including Google, are questionable to varying degrees. I lack the expertise to assess the more esoteric of their arguments as to which methodology is most accurate.
Nevertheless, Compete does seem to be consistent in terms of comparing one site to another. I spoke with several operators on my top-twenty-five list, and while all of them disputed the Compete numbers, they all agreed their sites were placed in roughly the correct position in the rankings. Most of them also said the Compete month-by-month numbers reflected the same gains and dips indicated by other measuring services. If there were major discrepancies, I didn’t find them.
In the future, I may revisit the top twenty-five, but if I do, I’ll be clearer about how I came up with the hard numbers and include a disclaimer about their accuracy.
Al Diamon can be emailed at firstname.lastname@example.org .