Forum Moderators: open
I have checked all 24 pages with a fine toothcomb and there are no nasty tricks at all and the 2 terms concerned are moderately optimised in a not too competitive category.
Both pages are PR3 and the Home Page PR4, which seems about right as no one has ever sought out any incoming links except a few minor directories and a listing in ODP.
I am reliably informed that "red widgets" did feature at about #25 and blue gizmo at about #35. Checking the opposition and the 2 pages concerned this would seem about right.
In an exercise to improve these 2 terms from #25 and #35 respectively, the exact terms were inserted as alt tags on one graphic only in the 2 pages concerned.
To check the expected improvements, Web Position Gold was used and apparently the URL for both pages was inserted in the "URL Verification" field as well as the 2 search terms across 6 or 8 Search engines. This took place several times (could have been 3 or 4) spaced out by about a week.
After some time (no-one can be specific) both terms deteriorated suddenly to their current low position. I assume at the next update or the one after.
The terms that include a geographic locator are still featuring at #1 and many other terms both generic and generic+locator are apparently not affected.
My question is:
Could Google have detected WPG and because the URL was checked to confirm indexing, penalised the site for the 2 search terms reported on by WPG only leaving other search terms unaffected?
Could Google have detected WPG
Yes. I wouldn't risk running WPG on your sites with Google. Many people have reported their sites, or pages from it, have been sucked into a big black hole after using it.
Have you ever noticed on many sites that the page they advertise WPG on is without a rank or 0 ranked? Google knows all! ;)
I would never use Automated Reporting especialy WPG, but this is a new client and it is a case of finding out the what and why that has caused this problem from previous undocumented site management.
There are only about 8 incoming links (only to the Home Page) and all are prety innocent with the link text "Business Name" except ODP.
All other search terms are achieving positions as deserved with a fistfull of #1's and loads down to #40. These 2 phrases however are totaly buried with no visible reason.
Both pages in question perform better for other search phrases that don't even appear in the <title> tag.
Are you sure you have seen a site penalized because of too high a proportion of links with the same keyword? If so, what would you estimate the percenetage to be?
The high profile case of Google Bombing was Microsoft and "go to hell". they certainly did not have a large percentage of their links with this text. I would think that the best way for GG to prevent bombing would be to check if the test in the link is also on the page it is pointing to.
Agreed. Otherwise I would definitely buy WPG to start checking my competitors. I guess that many of us would do the same, most probably without telling anybody ;)
Dan
As sure as I can be. I had it happen to several sites and saw it happen to 2 competitors. All had top rankings for pretty competitive terms for over a year, all got buried (just for terms in the link text) at the same time. All other terms continue to have excellent rankings and some are far more competitive than the ones that sank. The only thing that the sites had in common was the link text contained the search term (including in Yahoo and DMOZ). As for the percentages, I have a theory which I am testing - but it could take a few months to see if I am correct but it seems to change with the relative 'popularity' of the search term measured by the number of results Google returns. The higher the number, the lower the threshold - but this is an observation, not a proven fact.
In conclusion, I obviously checked for other factors that could have affected the rankings, such as new sites, better linking of competitors etc., etc., but this was the only common denominator I saw across all the sites where the problem was seen.
I understand you may not want to say what a safe percentage may be, but what percentage of links were the same when you feel you were penalized? Like 80%+?
Thx
All had top rankings for pretty competitive terms for over a year, all got buried (just for terms in the link text) at the same time.
How do you know it was just for hte terms in the link text?
Did you remove the link text and they got reinstated?
Does having the text cheap widgets and buy widgets count for the same penalty on the term widgets or is it an exact match you found to cause problems?
Because the pages rank well for other competitive terms which are not using words in incoming link text.
Yes, it is exact matches which cause the problem.
mmt - This means "The Red Widgets Company" wouldn't rank well on red widgets. That to me is a pretty drastic step for Google to take. It also doesn't correspond to what I'm seeing on sites I'm involved with, but I haven't done any sort of controlled comparisons. I'm still seeing this anchor text being very key in rankings.
Exactly so. An excellent example of what I have seen happen - but only when:
a) red widgets produces a very large number of results.
b) a very high proportion of links include that text - but if links are over a certain number and the majority have a certain PR then the affect is nullified.
Unfortunately, it is a bit of a moving target in that number of results given changes the proportion of links needed to kick-in the penalty - plus the quality of link then has to be factored.
There is a fellow member here who ranked for years for a top SEO phrase because that is the name of his company and all links have that term in the incoming links. Great PR and no rankings for that term after this happened!
Please forgive my skepticism. This to me is a very basic shift in how I see the algo, so I'd love to get additional confirmation.
Some thoughts:
Old fashioned natural domain names such as brandname.com, will on average have an X percentage of links towards the indexpage using brandname, or the full url brandname.com as link text. 100-X percent will be descriptive anchortext in natural language terms.
Now take: blue-cheap-widgets.com, Google could decide that, of all the anchortext links towards that indexpage, X percent are disregarded anchortext-value-wise.
Also mentioned here: [webmasterworld.com]
Maybe someone's links could be "over-optimized" as well.
I think as of the sept/oktober update, Google did take surrounding/heading effects into consideration.
From the originating link page and the destination linked to page.
For example, would you be penalized for "widget" if your links all had widget in them but also online widget, cheap widget, etc? It seems a bit unfair if this is the case.
Page 1 title "Red Gizmos"
"Red Gizmos" off the radar and buried #464
"Red Gizmos location" #1
"Big Red Gizmos" #1
Page 2 title "Big Widgets"
"Big Widgets" off the radar and buried #1,350
"Big Widgets location" #1
"Very Big Widgets" #1
Almost any word from within the body text added to the 2 words from the title tag produce top 10 position.
I have been to the wayback machine at archive.org but the last stored version is September 2001. So that holds no clues.
The people involved have poor memories about what exactly happened prior to the trouble but I have found no evidence of bad practice except the use of WPG.
The incoming links are very few and anything but optimised. So I am inclined to rule them out.
If there is indeed a selective penalty, these sites are not getting it applied to them. Most of their links are form pr 5 and 6 sites
Good points - the number of incoming links was pretty small, Yahoo, DMOZ and a few other directories - all with the keyword-domain-name as the link text - so, yes - would sort of stand out a bit :)
It did work wonderfully though for a couple of years! Oh well, onwards and upwards!
I've checked the only site I'm involved with where this characterization might apply... a keyword-keyword domain... and it's still doing very well. It also has good on-page optimization and relevant content.
>>the number of incoming links was pretty small, Yahoo, DMOZ and a few other directories<<
>>The incoming links are very few and anything but optimised.<<
Well... this isn't what I was talking about. I'm talking about a decent number of incoming links with relevant anchor text from good referrers.
It may be that in both cases, you guys are also seeing the effects of competition. If you don't have many inbound links, all other things being equal, it's very easy for competitors to pass you by quickly.
One thing I've noticed about incoming links that I request but don't control...
- for a keyword1keyword2.com domain, I can request link anchor text in the form "Keyword1 Keyword2" (with a space between words)
- but I'll still get a fair number of the links in the form "keyword1keyword2.com," even though the company name is clearly two words on the site.
So I don't think that that going after keywords in the anchor text is going to get you penalized, but I definitely wouldn't depend on it as my only strategy. Just my opinion....
