Thursday, February 28, 2008

Is Wordtracker accurate?

I've been looking at the keyword tool and at and it seems quite impressive. Put in a term like "SAT Prep" and the results seem right, but you'll notice two things: on the SEOBook site, the MSN and Yahoo! estimates always follow the Google estimates exactly (how can that be?), and that the numbers seem perhaps lower than you thought they were (although without other numbers it's hard to say). Second, on the WordTracker site, there are never less than 3 hits for any search, and the hits are NOT strict multiples of 3 or 4, but in the low hits they are always 3 or 4 (which looks suspicious, as some of those searches are pretty weird).

Anyway, I shall not digress. The point is not how we can determine that the results are suspicious. The point is that the WordTracker estimates are completely bogus. Why? Because if you look at their methodology ( you'll see that they don't really have the data! They don't have data on Google nor on Yahoo!. They have data on Dogpile and Metacrawler. Do you search on Metacrawler? On DogPile? These account for less than 1% of all searches. And from those they extrapolate keyword data for all search engines.

I know that in Marketing like in Sociology, drawing massive and probably erroneous conclusions from irrelevant data is a popular sport. But that doesn't make it right.

Highest-quality SAT video prep

No comments: