back
4 comments
This piece mentions "No manual intervention" and backs that up with some reasons why Google supposedly doesn't have any manual intervention in the ordering of search results.

If that's true, then does that mean the eval.google.com "secret" back-end application supposedly used by teams of moderators to manipulate search results was a fabrication or just not under the principle of "manual intervention"? Or, do activities two years ago not count in the principles shared today?

http://www.webmasterworld.com/forum30/29726.htm http://slashdot.org/article.pl?sid=05/06/01/1532216&tid=... http://www.searchbistro.com/index.php?/archives/19-Google-Se... http://www.davidnaylor.co.uk/is-eval-google-back-on-track.ht...

Marissa Meyer is also supposed (and a big stress on /supposed/) to have said that Google has 10,000 people checking search results for relevance:

http://www.25hoursaday.com/weblog/2007/06/26/GoogleScalabili...

I don't see any conflict. You suggest that that eval is used to "manipulate search results". It could just as easily be used to score results and flag up problems in the ranking mechanisms.

As the man says, they are constantly refining the algorithm, - so presumablt they need constant feedback on how well the algorithm's working.

No?

That's part of it, sure, but multiple sources use language like:

"Quality raters apparently spend their time checking search results, deprecating spam, moving the best results to the top of the search result stack, and (possibly) testing experimental Google features."

http://www.oreillynet.com/onlamp/blog/2005/06/behind_the_alg...

and

"They are paid to check search results of Google every day. Most of the employees, called international agents by Google, were recruited through universities all over the world. The aim is to avoid spam, to get the right sites at the top of the listing and to test new features, not shown to the public yet."

http://www.searchbistro.com/index.php?/archives/19-Google-Se...

"moving the best results to the top", "avoid spam" and "get the right sites at the top of the listing" suggested human manipulation to me at the time.

Of course, none of the several sources that talked about the rater hub project were Google itself, so determining the truth is difficult, but that was part of the reason I raised the question.

"Today, I would like to briefly share the philosophies behind Google ranking:

1) Best locally relevant results served globally.

2) Keep it simple.

3) No manual intervention"