There has been quite of bit of controversy over the past few days arising from the new LDA based tool recently released by SEOmoz. While there may have been some very well thought out, compelling arguments against giving this tool any credit whatsoever, I have to tell you that in my opinion no argument, no matter how well worded, is going to win over a good old fashioned demonstration.
I am a big one for testing, and test this tool I did. Now, I know, I may have voiced some opinions in the past as to my doubt of the sincerity of Rand Fishkin and the folks who run things over at SEOmoz, but regardless of what I said before, for me seeing is definitely believing. I plugged both the url for the post introducing the tool itself, along with the phrase [made up statistical bullshit], into the tool’s interface, and sure as hell this is what the tool showed me:
(click to enlarge)
I mean, c’mon now… those words weren’t used anywhere in the article, yet this tool was able to accurately determine that at least half* of everything that Rand said was relevant to that phrase?
I don’t know about you, but I’m convinced. 😀
Yer a funny man Michael… a damned funny man. I saw the tweet (to the post) and headed over right away with the belief that I’d find a morning giggle. Ye didn’t dissapoint as always. Kudos!
BTW I also got some more out of my system here; http://seobullshit.com/lda-google-games/
Thnx for the laugh brother… always appreciated!
‘The New SEOmoz Tools Is At Least Half Accurate. We just don’t know which half.’ 😉
Excellent test phrase. I’m a believer
Michael… the correlation is remarkable
That Wikipedia data sure does improve the accuracy.
Maybe they should try scraping Mahalo for their next data set?
You’ve convinced me [before the tool was out]!
If you only had a pretty graph to illustrate this Michael, I’d be willing to plunk down $1997 somewhere for additional unicorn training.
You’re funny Mike, hehe 😀