Google recently introduced a new keyword phrase filter during its most recent update. Some phrases were unchanged, but many highly optimized and highly competitive phrases were drastically altered. Some webmasters saw their sites drop from top listings to not being in the top 1,000 sites. Unlike other filters, this spam penalty does not affect a page or sites overall perceived value. The end effect is lowered rankings for various specific searches for that page.
If you have a couple commonly paired words on your website such as “search” and “marketing”, the new filter will check to see how the words are placed on your page (and possibly in incoming links). The new algorithm is looking for more naturally existing context to prevent search engine optimized websites from dominating search results. Essentially the goal of this filter is to get webmasters to adjust their content to be written for the users and not search engines. The problem with this logic is that good SEO typically falls in line with good site and page structure.
The filter is selective in that it is only enabled for some terms on certain searches. The filter can be overridden by placing a – character for the terms. If you search for search marketing you will find Commision Junction where my site was once listed. If you search for search-marketing you will see my site. If you search for search marketing info, you will see that the filter in not enabled for that search.
Off the start the filter was set rather corse, but it appears to be shifting more in line with reality. Some of the decent optimized sites are appearing back in the place where spam was just filling the results. Just this morning I saw a clients website pop up for two seperate phrases he was tripped for.
This filter marks the first time in search engine history where webmasters have not been able to reengineer exactly what the search engine did. Currently I do not recommend making any chages while Google is still fine tuning the filter.
Some business cant just wait and see what is happening…and a good ranking means money !! I think that the filters may be shaped but not removed again, also for making adwords more interesting for companies who dont like to invest time after time in SEO. I think is good to optimize a site already now…bring down the KWD if you have many links with good phrases….i noticed that this STILL WORKS !!
regards, hans
website ranking service
If you read this article talks about Google Bombing and how this type of activity is ruining Googles results you will start to rather than floating the business they might have to try and stop it sinking.
This article starts by talking about PR0 an earlier attempt by Google to stop the spam problem but goes on to talk about BadRank and explains how this might work.
http://pr.efactory.de/e-pr0.shtml Ignore the PR0 stuff and scroll down to read about BadRank.
If Google applied BadRank to a page based only on certain keywords then they would have created a Web spam filter which uses their patented PageRank technology applied in reverse.
If I were a Google engineer it would really appeal to me to use exactly the same technology, the same premis of the Web being a uniquely democratic place to fight against spam. It so simple it is brilliant. If you can build up more PageRank by linking to and having links from authority pages then using exactly the same premis in reverse you could build up an unreported BadRank by linking to and having links from “bad pages of pages” which might be used in a number of ways within the algorithm that produces the final SERPs for a given term.
So how do you build a list of terms to target and how do you identify what a bad page is. I think that Google’s engineers would look for an automatic solution to this. I suppose they might have used a hand built list based on spam reports but it doesn’t fit with their way of thinking and it means that Google would always be playing catch up when they should have enough data to stay ahead of the game.
Here’s how I would do it.
1. Identifying terms to target. Google just has to look at what searches are being done. Whilst it might be fun to Google Bomb some stupid term if no one is searching for that term then it does no harm so it can be ignored. If there is a certain level of searching for a certain term and/or the growth in the use of that term reaches a certain critical level it is automatically added to the list.
2. Identify bad/spam pages. If a page has links out including a particular search term but no links in using that term then it is probably bad. If it has many links out each using terms in the automatically generated target list, and no or a very low level of links in using those terms then it is probably very bad. If it also does not have any authority pages linking to it and it has a high density of the amalgum of words and phrases in the target list then it is probably a spam page.
3. Decide on a level of back linking to spam pages that will trigger a penalty “Google’s Matt Cutts’ remark on this issue is: “If someone accidentally does a link to a bad page, that may not hurt them, but if they do twenty, that’s a problem.”” Quoted from here http://pr.efactory.de/e-pr0.shtml What if just a few reciprocated links to pages which Google’s automated system has now identified as very bad indeed trips the switch and you page gets a temporary zero weighting on all algo factors for that search.
4. What mechanism would you use to store this information. In the same way that PageRank is stored following the monthly recalculation a Badrank figure could be stored but not reported. The match is then very simple if your page has a particular level of Badrank and someone searches for any term that is in the list being targeted at that time then it is filtered. If the term is not in the filter list then your page will still appear in its natural position for that term. The system does not need to store which terms your page/page is bad for, it can assume that it is bad for all of the filtered terms.
Bad page + filtered term searched for = apply null algo
Bad page + non-filtered term = apply normal algo
Good page + any term = apply normal algo
This form of rolling term specific filtering is scalable to the web, it produces a low level of false positives and it defuses most Google bombs before they go off. It is so simple it is brilliant, If this is not what they are doing then they need to stop what they are doing and start doing this as soon as possible.
Hi – big thanks (great site!).
Spent some great time in your site, really enjoyed it
Just wanted to say hi, thanks and bye
hi, nice site, bye
Thanks Bob. Great information..