Published: Jul 05, 2008 - 07:07 am
Story Found By: ViperChill 1779 Days ago
Category: SEO
9 Comments
9 Comments
Search Engine Land produces SMX, the Search Marketing Expo conference series. SMX events deliver the most comprehensive educational and networking experiences - whether you're just starting in search marketing or you're a seasoned expert.
Join us at an upcoming SMX event:


Learn more about search marketing with our free online webcasts and webinars from our sister site, Digital Marketing Depot. Upcoming online events include:
Comments
Nice .. Id like to see more discussions based on experiments like this instead of listening to you or myself spew personal opinion based on hunches.
thanks Halfdeck. That is exactly why i dont like arguments from incredulity. "Doesnt seem reasonable" shoud always be followed by "but it should be tested".
I say boooo on the closing teaser paragraph :-) A few of us experimenter types should get together, make a list of things to test, and then split them amongst us. We could cover more ground in a shorter period of time and avoid duplicate efforts.
@Marios Yeah I feel bad about it, but it was a middle ground between not telling at all or sharing with the ones that i know will not abuse it. Huge thumbs up for the sharing idea. I know that there are things like that in some of the close forums (DavN sometimes writes about participating in things like that) but nothing I am a part of. So, maybe we can do that in private or something...On a different note, cant believe this thing went hot on 4th of July weekend! W00T! I was sure it will not be looked upon at least until Tuesday...Thanks to everyone sphinning and to Viper for submitting it in the first place... amazing community here
Great work, Neyne! Weve figured out a few ways to circumvent this issue, but I would love to hear about the technique you referenced at the end of this post.
Its sharing like this which makes Sphinn such a valuable resource to our community. Thanks very much for making the test public and GOOD GRAB ViperChill.
Having done extensive testing on this, we do occasionally get results where it appears that a second link is passing anchor text, but in the vast majority of cases it looks exactly like Neynes result.You have to do this kind of testing with more than one page, because you dont know effect what screen scrapers, RSS feeds, etc. may have - in other words, you cant see every link, and you cant rule out some sort of pollution, especially in a one-off test.Based on the evidence I have seen, the only reliable behavior is the first link passing anchor text.
@DanThies Point taken. I am definitely trying to diversify the types of sites that I perform my tests on.
Youre the first tester Ive seen (at least in public) addressing the need to test with different types of keywords. The reason why we launched a thousand tests on this was because we didnt have confidence that made up words (floblag glifflestop) that dont appear on the page would be handled the same way as real words.Search engines have to be looking for better ways to handle anchor text - does a link with the word "this" really add anything useful from the search engines perspective?