Yahoo Hand Tweaking Search Results Pages - How & Why?

2 comments
Thread Title:
Seem Google will chase SE spammers with new remote workers
Thread Description:

Danny Sullivan, a man generally held to be in the know, provides some insight into the reasons behind the major SE's hand tweaking certain result sets as well as a practical example of Yahoo! doing this (alledgedly :-)

It doesn't take a search genius to realize that hand tweaking goes on quite alot, particularly on the kind of queries that were they to be poor, would attract negative media attention. Greg Boser made a good post about this detailed here.

Although TW is hardly a search engine, i often hand tweak the recent posts list by choosing which posts to reply to in which order :) It's a natural enough thing to do when the normal algorithms or mechanics of a system fall short of ideal.

It's a great technical post by danny, go check it out.

Comments

Had been watching this for a

Had been watching this for a bit. Hit a serp repeatedly only to have the site drop second page after a week. It was like the serp knew my sites didn't belong. The top ten all are coded with H=1 tags. I also watched it happen to several other, um, creatively optimized sites over a period of 6-8 weeks. After some research, we found a few categories where you would expect spam to be, but instead, you have a front page full of "official looking" H=1 results. Maybe an H=2 thrown in here and there. Can I say for sure it's happening - no, I don't work at Yahoo. But, as far as I'm concerned, there is merit to the theory. Just my opinion though ;).

Yahoo editing by hand thread at WMW

It could well be that Yahoo are manually altering the SERPs. It is especially relevant on "high worth" keywords.

Think pharm or gambling where you can see the following in place.

The "tracking URLs" that are in the SERPs contain a variable called H. When the keyword delivers results where the H is NOT zero, it appears that those that have H=2 rank higher than H=0 and where H=1 they rank higher than H=2.

A common libido enhancing drug has H=1 results higher than would be suggested by an algorithmic answer but common sense says that the type of information these specific pages have should put them prior to the more commercially orientated results below.

Hand editing ? I don't know but if it sure does look like it!

A personal thanks to goodroi at WMW for starting the thread there in the supporters section.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.