I wish I could tell you. I have a big issue with google and their cache right now. Very outdated..
The most common reason is that the actual phrase appears in the anchor text of pages that LINK to your competitor. Or even in some text that is near the link, or the title element of that linking page. I've even seen some evidence that this kind of influence may go back a few "generations" (pages that link to pages that link to the ranking page.)
There is also some semantic secret sauce going on in Google, where closely related phrases can occasionally get triggered by their related single keyword -- but in most cases, it's much more likely to be the influence of the text on linking pages.
This can be a little "crazy making" because Google doesn't give you anything near a complete backlink report. If you want to track down the details, you need to use another search engine to research the pages that link to your competitor.
If your talking about results that have nothing to do with the 3 word keyword phrase then yeah I'm seeing a lot of that for searches.
Maybe it's just another filter?.
|King of all Sales|
The semantic secret sauce you refer to is being used by Google extensively with Adwords and they have been doing it for awhile. It allows them to show ads a lot more often and hence, make more revenue.
I did check other search engines and I am listed correctly in first position. I checked the source code of the sites in the Google results to be sure the keywords do not exist. There is also no Adwords Ads for the other sites.
Forgetting that it effects my sites position it pisses me off that many searches I did today for a project I am working on yielded bogus results without the keywords.
btw - that should read 1-4 above.
Tedster - how many generations back would you say does it affect the results?
|The semantic secret sauce you refer to is being used by Google extensively with Adwords |
I agree with you -- especially in Adsense content network placement.
|how many generations back would you say does it affect the results? |
I pretty much proved it for two generations last year - to my own satisfaction at least - but I never looked further. If it's there for two generations, it seems mathematically that it must be iterative beyond the 2nd gen. But the effect I saw faded rapidly with just the second generation, and it also seemed to be less strong after Jagger. Not worth tracking back any further.
Somewhat related oddity: for a short period when BigDaddy first began rolling out I saw some really wacky sites high in the SERPs.
The cache page said the term appeared only in links pointing to that page, but it definitely did not (not even on a 2nd gen link page). It was a completely irrelevant and off-topic result, with only a typographical similarity to some letters in the keyword. Almost looked like stemming gone bad, but the cache was reporting that there was a linking page connection.
Imagine a page about "playing hookie" coming up high for a search on "okie dokey" -- that's like what I saw. I'm glad it was short lived -- that bugger was a big global brand with over 100,000 backlinks
what a stupid idea
So for all those sites that dont have a keyword domain that have links from sites that simply say "get your widgets at abc.com" with the abc.com as the anchor they havent a hope in hell of ranking!.
amazing - another gem from google currently
Well, I wouldn't say that -- there's plenty of examples of non-keyword domains ranking beautifully. It's not just anchor text that matters, it's also title element of the linking page, text that's nearby (proximate) to the actual link -- and I wouldn't be astonished that some other measures of the linking page are folded in as well.
We're talking about one little corner of the big picture here. and Google's total algorithm is complex. But when it misfires for a while, you can sometimes get a deeper glimpse into the anatomy or structure of the algo. (And when it misfires for a long time, it can make you crazy.)
In looking at this further the only common thing I can find between my site and those now above mine (without any of the keywords) in results is we are all in the same Google-ODP directory.
So evidently keywords in the meta tags, titles or in the article do no matter when performing a search. As long as they are in the Google directory that has the keywords in the directory name they show in results.
I would not draw 'evident' conclusions from the limited data that you mention.
do you check the 'quality' of a potential link 2 gens back? Or do you just go by current SE ranking of the linkpage you are targetting?