Google Florida Update New Hypothesis - Page 4

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View

Re: Google Florida Update New Hypothesis


Quoted text here. Click to load it


Free Search Engine Optimization, SEO and
Search Engine Placement Tips (updated 16/11/2003) /
Free SEO Mailing List

Re: Google Florida Update New Hypothesis

I previously posted that I believed Google's new algorithm was
targeting commercial sites. That was my knee-jerk reaction when I
first saw who dropped out of sight. More analysis shows I was wrong,
however those commercial sites in my industry that dropped have one
thing in common: excessive key-word density.

I analyzed the sites for my top 10 competitors, comparing their
rankings in October to today. I built a spreadsheet to compare key
word density for each site (total word count as measured did not
include URL, meta tags (except title), words less than 4 letters, and
common stop words).

The few competitors who remain in Google's top 10 today for our 2-word
key phrase have one thing in common: Each word of the 2-word
competitive key phrase is no more than 16% of the total word count (as
measured above) on the page.

Myself and my competitors who were dropped off the radar screen are
all running 20-25% density Therefore I believe a Google SPAM alarm
goes off at around greater than 16%-19% keyword density per page.

Tonight I will edit my site. I know many have said to wait and see
what Google does this month, but my competitors who remain now also
ranked well before November. So I see no risk to reducing my key-word
density. I think the new Google algorithm says in effect; let Google
decide what your site is about - don't shove it down our throat.

Ed Parker

Re: Google Florida Update New Hypothesis

Ed wrote:
Quoted text here. Click to load it

I've always run at 10% density and not a single site of mine completely
disappeared. I had 2 sites that kept disappearing then reappearing but
also during that time their PR was jumping from 5 to 6 then back to 5
again. In the end the PR6 remained and so did their position on the engine.

I too believe that keyword spamming is playing a part in the new changes.

I wonder if lowering the keywords on one page and spreading those words
over multiple new pages would make a difference?

I've always trodden carefully when it came to building a SE friendly web
site. Most of my sites start off basic and I slowly raise the stakes and
wait for the next crawl to see what changes have happened. All my
clients are extremely happy with the results even though it takes more
time. I've never been into keyword/anchor text spamming and all my sites
survived Florida (80'ish in total).


Simon Day
Free desktop wallpapers of Torbay at:

Re: Google Florida Update New Hypothesis

Quoted text here. Click to load it

I'm not sure this is universally applicable. I used scroogle to examine a
number of the folks, using my 2- and 3-word keywords, who dropped below 100
and did not find it applicable.

Also, in my case, for my 'best' 3-keyword-phrase, its true my denisty is
26%, but I only dropped from about 13 to about 24.(i.e. not off the board.)
There are competitors ahead of me with 50% density, and competitoirs after
me with <10% density. Additionally I note that all my other keyword
densities are below 14%, with the vast majority below 10%.

I grant I didn't due as rigiorous an examination as Ed describes, but a
"quick look" at least for my keywords doesn't seem to, by itself, explain
what's happening.

( And no umbrage taken if someone can show me wrong.)

Bob Kochem
| MinuteMan Systems          -        Project Management Software |
| |
| PO Box 152, Belmont, MA 02478  USA                (617)489-5639 |

Re: Google Florida Update New Hypothesis

my wife has about 4 years dedicated into her home built business
website and this whole florida thing has wiped her off the face of the
internet.  A legit business, trying to make a go of it.
How do we work out this density thing?
Much much appreciated
Ron (for my better half)

On 3 Dec 2003 13:35:24 -0800, (Ed) wrote:

Quoted text here. Click to load it

Re: Google Florida Update New Hypothesis

Yes I started along your line of thinking too...

search term horses  12 x but not dense
2xe-mail link
keyword in anchor link twice

However every single image on the page the file name:

and they all link...........

but then horses isn't a commecial term like horses for sale is it ;-)

then look at horses pictures

then pictures horses (shouldn't be different but it is) I like this result best :D

Re: Google Florida Update New Hypothesis

Quoted text here. Click to load it

So tell me what are we seeing now if its not an update?


Re: Google Florida Update New Hypothesis

Quoted text here. Click to load it

The above may well be true, partially true, or not at all true - here
are *MY* observations - and they are (of course) based on analysis of
searches for which I expect OUR COMPANY WEBSITE to perform well in -
as always, your mileage may vary...

Synonyms have been added to the backend databases queries.

It's that simple as far as I can tell.  I think a few of the spammer
filters have been dropped, but google is DEFINITELY applying synonyms
to searches for common terms.


google for: fort collins web hosts

google URL:

Look for: highlighed text!

In the results, you will see highlighted:


(I'm leaving out the fort collins web  - as these have NOT been
synonymed as far as I can see - not even ft or ft. for fort).

A month ago - you would NOT have seen anything but hosts highlighted.
Obviously you have to take that at face value if you can't remember
the details of your searches a month or so ago, but I'm sure many of
the gathered masses here remember that singular and plurals don't rank
the same in seaches - at least they DIDN'T USED TO.  It appears for
this particular search, various synonyms for "host" have also been
allowed to influence the outcome.

I strongly suggest you perform some of your stronger searches, look
carefully at the highlighted terms - see if anything SIMILAR to your
search phrase is highlighed, but something that doesn't match the word
exactly... you should find some - I've tried a number of various
phrases, and most I find some kind of related search being factored
in, where I'm 100% sure that wasn't the case before this "florida"

This (at least to my mind), has accounted for the generic searches
(those that do not have a geographic qualifier) being blown to atoms -
at least many of them have.  While the geographics for our business
have 95% remained the same (with some dodgy portal/links sites
creeping in to lower rankings - and 1 top spot).

I'm really surprised no-one appears to have noticed this... it's a
HUGE difference in the results - hiding right there in plain sight!


Greg Hewitt-Long

Re: Google Florida Update New Hypothesis

Greg Hewitt-Long wrote:

Quoted text here. Click to load it

People did notice, posted in their blogs, discussed online, and posted
to newsgroups...

But you're right, some seem to ignore this new spin on "Florida".

Re: Google Florida Update New Hypothesis

Quoted text here. Click to load it

ok, perhaps I should have said "no-one here".  While everyone appears
to be looking for some "magic" combination of alt-attribute and
link-text ranking change which will explain away the changes in
ranking that appear to have been made by this "Florida" update by
google, it's most easily explained (at least to me, by me), by the
stemming technology which some people spotted as a change on the
google help pages (I spotted it by looking at two search results and
reverse gathered that intelligence).

My point is that people too often look at ways to explain the changes
by plugging in external variables which are at best "guesses" - when
the most plausible theory and the evidence for it are sitting right
there in plain view on the results page.

Even the so-called expert's expert - Danny "search engine" Sullivan
only seemed to have half the facts in the articles I saw linked
here-abouts - not that it would surprise me, I'm more impressed that
there is actually some basis in the drivel he posted this time!

Quoted text here. Click to load it

I think (from evidence I'd seen, searches I've performed, and my
systems analysis background), that the stemming technology adoption
accounts for 90%+ of the changes arising from this "florida" update.

as usual, ymmv...


Greg Hewitt-Long /

Re: Google Florida Update New Hypothesis

Greg Hewitt-Long wrote:
Quoted text here. Click to load it

I agree with you making page coding changes probably won't do anything
except take our minds off the problem and give us something to do and
feel more relaxed about it all. But your idea with synonyms and stemming
technology is even stupier. The stemming technology was there before the
florida update for one thing and has nothing to do with how sites are
listed or ranked in google. It's there just to assist people if they
spell a word wrong they're searching for or to give alternative
possibilities. If anything the normal sites will just re-surface again
as if nothing happened over the next month or so no matter what anybody

Re: Google Florida Update New Hypothesis

Quoted text here. Click to load it

I take it you mean "stupider" - as in more stupid.  You're wrong - not
arguing with you here, but you are 100% incorrect.  The stemming was
added to many other engines as long ago as 1999 I think, but Google
only added stemming of search words VERY recently.  Do your homework.

Quoted text here. Click to load it

you are looking at the WRONG THING.  They provided alternative
spelling and search phrases in the "Did you mean?" section for ages,
the results being drawn from stemmed keywords is VERY NEW - and 100%
coincides with the "Florida" update.  This is perhaps where you are
confused - I'm not talking about the "Did you mean?" section, but the
result keywords highlighted - these are not limited to the words you
typed in, whereas before "Florida", they were - 100% - ONLY the
keywords you typed - they NOW contain stemmed words in addition to the
words you typed.   Read on.. I'll explain how to verify this for

Quoted text here. Click to load it

If that's what you care to believe, you are basing your theory on
major changes that worked themselves out, or largely got reversed in
the following months - I'm not sure this will be reversed out.

As I said, do the following search:

fort collins web site hosts

Now look at the highlighted words - these include hosts, host and
hosting - words stemmed from hosts base, which is "host".

Now turn off stemming using the -boguswordblah

search for:

fort collins web site hosts -boguswordblah

You will see that the results only contain the words "fort collins web
hosts" - no highlighed words will be host, hosting etc.  The stemming
is turned OFF by using a remove keyword qualifier.

You can verify this yourself if you choose to, or continue to tell
everyone that I'm wrong without actually checking it for yourself.


Greg Hewitt-Long
Happy Holidays - $10 Domains at Web Your Business!

Re: Google Florida Update New Hypothesis

Greg Hewitt-Long wrote:
Quoted text here. Click to load it

Greg's hypothesis goes against the "filtered keywords" hypothesis
-- but it makes more sense via Occam's razor, because it requires
less hand work. All google had to do was adopt widely available
stemming algorithms rather than make difficult, individual
editorial decisions about which keywords to filter.

Comments from the more experienced?


cat yronwode

Re: Google Florida Update New Hypothesis

Quoted text here. Click to load it

Yes, the filtered keywords that people talk about can be a "possible"
explanation - it's highly unlikey, given that the google help page now
mentions stemming, and it did NOT previously.

Here are a few more examples:

Google for: fort collins web site design
Highlighted stemmed words: designers design

Google for: denver marketing companies
Highlighted stemmed words: company companies

Google for: real estate agent new york
Highlighted stemmed words: agent agents

My hypothesis is that singular/plurals were the first stems to be
added, along with SOME stemmed words - this might back up some of the
filtered keywords theory (to some extent), in as much as they have
access to counts for popular search phrases - I'd put a TON OF MONEY
on them stemming the most popular phrases first - wouldn't you??   I
know I would!

Quoted text here. Click to load it

More than what?  You?  Me?  I have 7+ years doing this....!


Quoted text here. Click to load it

Happy Holidays - $10 Domains at Web Your Business!

Re: Google Florida Update New Hypothesis

Greg Hewitt-Long wrote:
Quoted text here. Click to load it

Than me! :-)

I only play SEO at google for a game, the way some folks go to a
backgammon server and pick up an opponent for a game. I've been
spending a few hours on SEO-gaming per week since 1995, but not
for pay, just for play. Google is the best SERP-server i have
found, because it updates so quickly that the rounds don't take
long. It is currently the only SERP-server i play.

I pick an actual two-word or three-word search term that appears
on my already content-rich pages and i try to run one of my pages up to #1 for that search term. The search
term must pre-exist at my web site; i do not create search terms
for the purpose of the game, as one might do when google-bombing.
The perceived commercial popularity of the search term is not as
important as winning the game and actually supplying great
content on the subject. No keyword spamming, faked-up
backlinking, or cheats like invisible text are allowed. All
coding is in html v.1 with no frames, blinking, pdf, shockwave or
flash. The game is played with NO QUOTE MARKS (i consider quote
marks a handicapping device, and i am too good to need the

Examples of chosen search terms for which a page is
currently at #1 in google's SERPs with no quote marks: rural
acoustic blues, evil eye, syd barrett lyrics, mojo bag, love
spells, sex magic, candle magic, hamsa hand, sachet powder,
secrets of the psalms, saint expedite, lucky tokens, penis
amulet, penis bone, ... and, well, there are a zillion of them.

Examples where i am currently stymied and have not (yet) been
able to rise above #2 or #3 for various reasons: hoodoo
[terminology conflict with the Hoodoo Mountains ski resort area;
i remain at #3], hoodoo rootwork [google/dmoz directory is at #1,
my page at #2], magic spells [i have been stuck at #3 and #4 for
more than a year], occult shop [at #2 and stuck for at least a
year], and lucky elephant [at #3 but still trying to rise and i
think i can make #1 on this one].

Examples i am currently campaigning actively and for which i am
not in the top 3 yet: folk magic [at #5], lucky coins [at #6],
occult store [at #6].

I campaign on a search term until i score in the top 5, and if
the way looks clear i go on to the "semi-finals" to crack the top
3, and if i make that, then i move into the "finals" rounds and
see if it is feasible to hit #1. If i win the jackpot #1 listing,
i pat myself on the back and start again with a new term. It's fun!

cat yronwode
news:alt.lucky.w --- discussions on folk magic, luck, amulets, charms

Re: Google Florida Update New Hypothesis

I like the way you did your site.

catherine yronwode wrote:
Quoted text here. Click to load it

Re: Google Florida Update New Hypothesis

Sam wrote:
Quoted text here. Click to load it


Credit where credit is due:

I owe one of my most basic design paradigms -- the linking of
body-copy -- to Shannon Turlington, a programmer and tech writer
who briefly had a small demo site on voodoo online way back in
1996. I ran across the site one night while searching Alta Vista
for something about Haiti and i realized that although the
content was good, the site actually had been placed online to
demonstrate the extensive use of internal links within blocks of
body-copy. I had never before seen such a consistent use of
internal links in body-copy at a web site, but i "got it"
immediately -- and so i started down that design path at my own site.

Incidentally, Shannon -- who has also written "The Unauthorized
Guide to the Internet" and contributed to tech books like
"Microsoft Office Desktop Reference" and "Teach Yourself the
Postal Service Exam in 24 Hours" -- took the voodoo demo site
offline and published it as a book ("Complete Idiot's Guide to
Voodoo"). Alas, although the informational content in the book is
the same, it was more navigable as an html site than when printed
on paper.


cat yronwode

Re: Google Florida Update New Hypothesis (Greg Hewitt-Long) wrote in

Quoted text here. Click to load it

Quoted text here. Click to load it

I don't think this is a complete explanation of the changes.  

Try 'weighing scales' - a search that's changed a lot post-Florida. The
cached version highlights 'weighing' and 'scales' but not 'weigh' or
'scale', with or without a -babubabu bit added.

However, the results are clearly different if you add the -babubabu.

One site that has moved from the top ten to nowhere for this search,
with and without -babubabu, is I am still trying to
work out exactly why it has suddenly fallen quite so dramatically.

Clare Associates Ltd /
01822 835802

Re: Google Florida Update New Hypothesis

Quoted text here. Click to load it

Sorry, just realised I misunderstood you.  You meant the bold text in the
main search page, not highlighted text in the cached version.

I still don't think this is the whole answer though.  I am still seeing
many pages optimised specifically for one term rating rather highly, and
singlular and plural searches producing very different results.


Clare Associates Ltd /
01822 835802

Re: Google Florida Update New Hypothesis

Quoted text here. Click to load it


I googled weighing scales and the terms weigh and scale were in the
highlighted results.

Unpaid result #3:

Operation of a Bulk-Weighing Scale. All official weighing of loose
grain is
performed on bulk-weighing scales. A diagram of such a scale follows.
... - 8k - Cached -
Similar pages

All variants of Scales, scale and weighing are highlighted.

Unpaid result #10:

Scales & Weighing,Scales & Weighing, Automatic Net Weigh Scales, ...
Scales & Weighing,Scales & Weighing, Automatic Net Weigh Scales,
Gross Weigh Scales, Manual Net Weigh Scales, Manual Gross Weigh Scale.
... - 9k - Cached - Similar pages

Tons of variants of scale, scales, weighing and weigh - all

Perhaps you were hitting a different database to me.

You mentioned the "cached results" - do you mean the cached pages
which appear when you click the "Cached" link?  I never once mentioned
these.  Even on a result with highlighted weigh and scale, the cached
version does NOT highlight anything other than EXACT search phrases -
this is NOT what I was referring to, and it's a totally different
issue what I refer to as stemming for providing the results


Greg Hewitt-Long
http://aaabusinesshosting - $10 domains, hosting from $4.99 a month

Site Timeline