Forum Moderators: Robert Charlton & goodroi
[webmasterworld.com...]
Many of us were rightly fearful that this could likely become yet another tool for SEs to use against webmasters, rather than help us.
As that thread shows, Wikipedia was one of the first to begin implementing the "no follow" in all external links immediately.
Consequently, after the Allegra update in Google, it appears that the Wikipedia articles have indeed dropped from SERPs.
Our fears appear to have been proven correct, with Google Allegra. Google strikes again.
I have seen it firsthand: a Wikipedia article that used to be 3rd in SERPs (only for the last 6 months) in one particular keyword -- mostly because it linked out to the true authority sites (which Google itself refuses to allow those sites to be seen in the same high SERPs anymore) -- has now disappeared. {NOTE: This particular Wikipedia article did not ever link to the useless sites that Google has absurdly rewarded with high SERPs.) Notwithstanding that a Wikipedia article is probably not the best result to be in top SERPs anyway (even if using the correct authority sites in outbound links), this new finding still does offer possible insights into coping with Allegra and the issue of using "no follow".
Outbound-linking with "no follow" could possibly be self-destructive to SERP position. Or, linking to true authority sites could be necessary for high-SERP position (even if Google refuses to allow those authority sites to be high in SERPs themselves.)
Does anyone else have some other ideas as to what the ramifications of this information might be?
Or, for that matter, has anyone else also seen a Wikipedia article disappear in the SERPs like this, following Wikipedia's implementation of "no follow" and Google's Allegra update?
Searchers will find what they want to through 1 of 6 ways
Organic traffic from SE's
Word of Mouth
PPC
Links from other authorities
PR
and finally just because they bookmark and go revisit
All the big SE's decided that they wished to provide websites with the ability to NOT BE SPAMMED but also not to lose genuine content and implemented the nofollow attribute
allowing websites to make their own decision
This was not a Google only policy
wikopedia made a decision to implement nofollow and from my small amount of checking it does not appear to have hurt their traffic and now many more may follow
just my own little uninformed view of nofollow and Wikipedia
steve
I'm curious, how did you exclude other variables and isolate nofollow?
Also, can't we form a completely different (and reasonable) hypothesis about nofollow given wiki's sudden rise in traffic as reported by Alexa [alexa.com] and others?
Also, can't we form a completely different (and reasonable) hypothesis about nofollow given wiki's sudden rise in traffic as reported by Alexa and others?
Was that rhetorical?
I have obvserved outbound links to help a good deal. For the most part it's always been links to authority sites, but it's difficult to say which of these criteria that are helping:
1. Linking to an actual topic authority.
2. Linking to an actual topic authority + using the same anchor text as the targeted search term.
3. The targeted search term being used as outbound anchor text regardless of where the link goes.
I'm omitting other possibilities for lack of time. Most of them would probably be stripped of benefit by adding a nofollow.
The sites (not just mine) that I have seen this happen to are all authority and well established sites that started being affected by this during the re-ranking on Aug 14th last year and then massively on the 14th of Dec.
Plus, to say fact is fiction when the fact I reported IS a fact only shows one's hostile bias rather than being willing to check out the reasonable possibility.
I simply observed a fact and shared it here with the intent of being helpful. (I even went so far in my original post in this thread to refer to G$ as Google instead. Doing that was my way of further trying to show even that much more than normal that I was only seeking to open an unbiased legitimate discussion and research without anyone's pro- or anti-G$ perspective. Since I have been attacked by G$ cultism here though, it has forced me back into pointing out the "G$" handle.)
Anyone attacking is only out to destroy what others might benefit from this discussion. After all, what IF my hypothesis IS right? Then to try to attack and refuse to consider it only seeks to prevent WebmasterWorld users from learning what could very well be an important variable to consider.
I'm not the bad guy here.
Let's keep this thread reasonable and stop the attacks and trolling.
Now back to the topic...
But that is not what I am talking about in this thread here. This is a completely separate issue.
I even went so far in my original post in this thread to refer to G$ as Google instead.
I thought I had made some useful contribution but, that's OK, I'll stop "attacking" you and I'll keep out of your thread MM$. :)
[edited by: Macro at 4:11 pm (utc) on Feb. 14, 2005]
The wikipedia sites have all sorts of redirects going on. There are scores of 301s and other server responses. There are some other technical issues at play too.
Examples...
.com 301s to .org
www. 301s to en.
I wonder if they made any major recent changes to URI structure...
So I don't think this is anything to do with "no-follow" which has been implemented site-wide. Or for that matter, Allegra. I would look elsewhere (loss of inbounds to the page etc).
TJ
Of course, not all Wikipedia traffic comes from search engine results! It's quite a sticky site.. and actually quite addictive if you start browsing it.
explanation of how some wiki articles might have a better possibility of surviving this possible "no follow" consequence in Allegra.
I think there is almost certainly likely to be a consequence, but just how drastic it is remains to be seen.
I'm sure that quality inbound links will diminish it's effect - you're either authoritative or not. I always viewed a hub-structure to be a variation on authority, not an antidote.
TJ
At Feb 1st balam first discovered the wikipedia fall in the SERPs and at Feb 5th I saw the same for one of my keywords. Look at [webmasterworld.com...] in the Link Development forum.
I also noted that at one of my sites I had one page which was not spidered for several days and the cache date in Google stayed the same where all other caches where updated regularly. This "freezing" page happened to be the only page on my site with the rel="nofollow" attributes. I removed the rel="nofollow" and now the cache in Google is updated daily again.
Did anyone try site:www.wikipedia.org or site:en.wikipedia.org? Only very few pages have a cache and snippet. I didn't check the site:wikipedia.org searches before Allegra so I am not sure this is an Allegra effect, but it looks very suspicious.
The lack of outgoing links to quality sites is a likely contributing reason.
A certain decrease by some small amount in linking *to* wikipedia pages by sites listed on those pages is a likely contributing reason.
The godawful mess that wikipedia is in terms of SE-friendly webmastering is always a wild card... redirects, century long load times, etc. With a site so search engine unfriendly it would be hard to pinpoint any cause for anything.
I recorded 10 Wikipedia search results for 10 random non-competitive phrases just before the rollout of "nofollow" on the Wikipedia site. My intention was to see if the new attribute would have any effect (I also recorded the # of OBL's on each found Wiki page).
Of course since then, we've had this big update so it will be impossible to determine what effect the new attribute had on the changes.
Wikipedia is a special case because of the vast number of duplicates on the web. Plus they seem to have server config issues ie. www.wikipedia.com = en.wikipedia.com. Therefore I think WP is probably a bad example for testing anything other than dupe content related issues. Regardless, here are the results:
KW# OBL PR SER1/SER2 DIF
001- 01 5 001 / 001 0
002- 03 5 003 / 003 0
003* 05 5 013 / 016 +3
004- 06 5 072 / 072 0
005- 08 4 026 / 280 +254
006* 14 5 028 / 017 -11
007* 15 5 001 / 002 +1
008* 17 6 017 / 022 +5
009- 19 6 002 / 001 -1
010* 26 5 010 / 009 -1
Note: The KW05 result has a drastic drop but in my opinion it is warranted. Meaning the original 26th placing for the wiki article was not deserved because it was only losely related to the KW. The Wikipedia article now at 280 is actually a different (but more relevent then the previous) article.
When you take KW05 out of the equation there really isn't a significant overall change.
*There's also something wierd going on. These results above were all found using the Opera Google search tool. The ones marked with an *asterisk came up as "title only" results. The articles with an asterisk were entirely missing from the SERPS when I did the same search using the Firefox Google search tool. I then did a little checking and found that 4 out of 5 of these "title only" articles are Mediawiki "redirects" meaning it is more than likely a dupe content problem.
Only 3 of the 10 results were identical in both Firefox and Opera....!?
Does this mean that the results haven't settled? Or is there always such a difference between browsers? Are the "title only" results on the way out, or are they the 'real' results, on their way to re-aquiring a snippet? I will keep an eye on it and see what happens.
I can't really conclude anything from this little test. Except:
1) I can say that the wiki results haven't changed much in this small sample (with one deserved exception).
2) "Title only" results are probably the result of dupe content issues because of the poor way Mediawiki handles internal redirects.
I don't know what else these results indicate. It was much too small a sample.
