Sorry if this comes a bit off-topic, but it is an important followup of a
thread 3 months ago.
For the record: 3 months ago I was complaining about the bad google
ranking of my wiki site (
) when looking for "good"
keywords (`bioinformatics wiki').
Today (hopefully it will last), it is the first time I see it appear in
the top 10 results (ranked 2nd after [[wikipedia:bioinformatics]] in
both Google search and Yahoo! search).
So I was not completely wrong in my expectations, and the concept of
"sandbox effect" seems to apply. It took approximately 6 months since the
opening of the site to get honorable Google results (see the date of one
announcement I made at that time:
On Sat, 11 Feb 2006, Gordon Mohr (@ Bitzi) wrote:
Martin Jambon wrote:
On Fri, 10 Feb 2006, Matt Morgan wrote:
Side note: I'm new to MediaWiki, but I'm
having a really hard time
googling for help. If I include "MediaWiki" in the search string, the
results seem to return every third wiki out there, rather than help
about MediaWiki. Any tips there?
I also noticed serious flaws in the Google results when searching for some
wikis. Google seems to have problems with the "wiki" keyword:
Shouldn't a search for `mediawiki wiki' (w/o quotes) return:
It does, in position #12. And the higher results are mostly reasonable,
Mmmmh, I don't feel that's usual Google quality.
`mediawiki' alone works, so that's fine, but not expected.
Their secret and proprietary ranking algorithms tell Google that people
searching for [mediawiki wiki] should get the other results higher, as
compared to plain [mediawiki] results. Tweaking the algorithms to do
otherwise might help some cases, but hurt elsewhere. They're optimizing
with regard to a lot of variables, so local anomalies aren't surprising.
And can we be sure their ordering for [mediawiki wiki] isn't optimal for
generating the most satisfied users (or most ad revenue)?
Overall, they seem to be doing great ranking wikis and URLs/content with
'wiki' in them. There's no evidence of any 'flaw'.
Maybe it's just a sandbox effect on my site as you suggest.
I was suspecting that all links from blogs and wikis tagged with rel=nofollow
would have a bad effect on the ranking of these new generation websites (as
far as I understand it's a widespread practice which is supposed to avoid
Google bombing http://en.wikipedia.org/wiki/Google_bombing
personally have a problem with the wiki I am running:
`bioinformatics wiki' should return some link to http://wikiomics.org
the top ten Google results, if not first.
How modest! :) You do appear somewhere around 60-70.
Come on, there are not 60 bioinformatics wikis!
Funny thing: I was about to say "come on, Google is not Altavista". So I
discovered that (1) Altavista still exists and (2) my site is ranked #2 on
Too bad nobody uses Altavista anymore.
And, you are named in the 'snippet' shown
for the #1 Google result, a
Wikipedia page, perhaps helped by the way you bumped your site to the very
top of the list of related links in an edit January 31st.
So I really don't know what I can do, I was
used to much better results
from Google for my non-wiki pages, without forcing anything. Does anyone
have a similar experience? Any fix?
It looks like your site is less than 3 months old. So step #1 would be:
just wait. It takes a while -- sometimes 6 months or more -- before Google
trusts new sites with a prominent ranking:
Very interesting, thank you. That matches exactly my symptoms.
Separate from the sandbox effect: as your site
gets deeper content, more
organic links from diverse sites (not just backlinks you've pushed out
yourself), and longer visits from happy users, you should also expect it to
find its "true" higher level.
Martin Jambon, PhD
, the Bioinformatics Howto Wiki
MediaWiki-l mailing list