September, 2002 Google Update Discussion - Part 1

Discussing the major changes that took place

[dauction]

How on earth can they justify dropping sites that were ranked in the top 10 and are now page 20 and NOTHING at all has changed on the sites from the last month?

The biggest thing is they move the toilet mid stream without a hint they are going to do it...(change the rules)

Googles a joke..

tired of their games..

off to support ANY other search engine..enough of this every month change the rules nonsense..good bye Google ..Good riddence..

[pmac]

I see a change, but I don't see a complete disaster for any searches I try. Don't get me wrong, I certainly see the odd glitch here and there, ie: I just did a search for "city" kw, and the number one result has nothing to do with the query in any way, shape or form. The query is not in the title, text on the page, description, and I dont see any links show up for the page. I also see some mirrors and doorways creeping into the SERPS but at least they are relevent ;-)

Maybe I'm not looking hard enough for the real bad stuff I keep hearing about. Feel free to sticky me a SERP or two of some bad ones, I would love to see them.

[4eyes]

I am seeing some very strange results.

Some very competitive phrases that I monitor have had almost all the top ten drop of the scale. Fortunately only one of my customers appears to have been hit.

The sites replacing them are most definitely NOT spam, some appear to be untouched by a search engine optimiser!

In general, they appear to be less relevant but not spam.

[steveb]

I know this is a bit specific, but I just did a literal search for "city". The top result is a search engine, but the URL is literally "city". It redirects to the search engine hompage. Excellent example of how keyword in URL is critically important this month... which means any site that titled its URL bizarrely will show well in the "wrong" results. I don't think this is a major error on Goggles part as it seems reasonable to assume that widget.com should have something to say about widgets. It appears though that they are giving the URL *too much* credit in the algo and that should be toned down.

[jbage007]

this is precisely what I mentioned in a previous thread--the searching public will soon (!) grow weary of the atrocious results provided by the "new and improved google"

we shall have the last laugh

good thing I didn't buy any google stock
(as if I could)

[austtr]

I didn't say the top results are spam... not implying that in any way. What I'm seeing is top 10 results that seem to be offering the viewer results that are much less focused than they used to be.

An example..... a search for "placename accommodation" produces results in which those words do appear, no doubt about that. But in one case, the placename bit appears as part of an email address and the word "accommodation" appears in random text somwhere else on the page. The site is actually about a town on another continent.

I'll put some of the others sites in the form of fictional examples:

Assume we want "german accommodation". What we get is French accommodation because "accommodation" appears in the text and there is something about France being a neighbour of "Germany".

We are also offered Black Forest accommodation. Same logic where "accommodation" appears in the text and the Black Forest is in "Germany".

Obviously I can see the connection, but top 10 results? I don't recall Google providing those type of results before... or am I just becoming befuddled and starting to dribble a lot?

I have no idea whether these results are due to forward, backward, inward, outward, stand on head links.. or any other permutation of the SEO craft and/or algo tweaking.

The point being that Joe Public has no interest in all those things and just wants to be served the closest match sites available. Based on what I am seeing I wonder if these latest results are providing that.

[vitaplease]

Its a pity Google does not have a www4.google around where you can search for results as they were before the update ;)

There have always been strange results in Google and it is difficult in hindsight to compare.

Not that I can compare, but I agree that single word searches tend to be less perfect though.

Do a search for "god" or for "america" (I hope generic enough to use as an example) and you will see not so perfect results.

[JonB]

well, there seems to be some strange results, yes. also dmoz imporntanceseems enourmos. too mcuh in my opinion since at dmoz little editors rechek all sites if they update or not.. i just did few searches and third search was "chinese food".well, jsut check the first site, kw in url, mentioned in description btu check out the page - it is closed like SEVERAL months (says will be reopened june 2002 that means it was closed BEFORE june and i would say few months before that).it is few sentences page that si all! how on earth does this belong to #1? definitelly not good for "average Joe" 2 surfer...

of course thee are some great results (well, at least i think so becasue it is possbile better sites are buried deep donw so we cant know this) but #1 is really funny. all sites below this have same pr and a lot more backward links...

that is of course only one term i searched , 3rd that came to my mind actually ,i dont have to search deeper, inet is expensive here :)

Jon

[Paully]

Taxpod,

Thanks for the great, honest comment. I also agree. I see very little change in the quality of search engine results. If anything it looks a little cleaner this month.

[subway]

I am seeing a big change in SERPs that have 1 or 2 sponsored links at the top and adwords running along the side - i.e. the money keywords.

The sites that would have been up top for these keywords have been replaced by "information" sites with an obvious absence of words such as "buy here", "book online" etc. forcing me to click on one of the paid ads if I in turn am looking to buy or book something.

I am still getting good honest results from all "non money keyword" searches, e.g. "Aunt Mary's Chocolate Brownie Reicipe".

[chiyo]

In information type queries, many SERPS have improved, some have stayed similar in relevance terms though the "players" have shown some small change. Very few showed no change in the top 10.

Our corporate site is getting less referrals, but this may be because we have changed strategy recently (reflected in our website) and some of the old incoming links with hypertext referring to "old" business foci no longer count.

I beleive however in high competitive high commercial sites (only from reading comments here from people who have been hit) the SERPS seem to be having problems. Im not sure what part of that is emotional over-reaction.

So after viewing the Google referrals for Monday (ezine info news sites showed no change or very small increase for Monday averages, and the coporate site was around 15% down) and reveiwing many keyword queries, Iid say overall a good update for "information" type queries at least.

[chiyo]

Hi subway. I know what your last post was trying to imply! ;)

I beleive that Google is not as good at ranking commercial sites as primarily information sites. Basically the basis of PR and link popularity just does not work with competitive commercial sites. (eg: such sites dont link to their competitors while information sites generally to link to related information sites)

Im many posts here before Ive suggested that targeting Google is no long term solution for web exposure and promotion for highly commercial sites just because of that. Its better to use Overture Adwords and many other avenues if you want reliable exposure.

In this update it seems Google has incorporated some changes that has impacted commercial areas in particular. Maybe things like downgrading value of reciprocal links, freshness, and downgrading off topic incoming links. Some commercial sites who couldnt get PR or incoming links naturally have used "workarounds" that Google's algos have just caught up with in refining their systems. Things like negotiatiing "artificial" reciprocal links that acheive nothing than increase PR. Maybe buying links and negitiating links for the value of PR and popularity only and not for the users of each site.

The result is we have a lot of curious posts saying "Google has broken me", "My family will now starve". And suggesting Google shoudn't change things without informing everyone beforehand. Google cannot be expected to be the free money trough for commercial sites, and should be expected to do things to reduce spam. Notifiying what these things are, even if it is a good idea, would be impossible given the complexity of features.

Im seeing a shakeout that should in time convince totally commercial, revenue generating sites to go elsewhere for promotion. No search engine can index the whole web reliably, and the hype about Google has made people feel that it can. It can't.

So though we have 5 sites, 4 information ones and one comemrcial one, we target google for the information non-revenue generating sites, and overture and adwords, and many other avenues for the commercial one.

Thats why i dont have any problems with the searches i do, and some of my work requires professional research including finding info online.

Whether Google is trying to "lead" people to Adwords in competitive "money" areas is a separate issue, but it may well be that Adwords provides better results as its hard for competitive commercial sites to gain "authority". Sure it brings the staus quo moneyed corporates to the top, and it reduces the ability of mom and pops, and "poor" or "family" companies to compete, which was a great thing about the web before. But now big biz has found the web, and its a serious business. Now there is just so much competitiveness and spam in competitive commercial areas that google can no longer keep up. In the end money talks. Its sad, but was inevitable as the Web grew to become mainstream.

To me the key is to always keep changing strategy. Those who think that a small company can keep up in big money areas in competitive areas are living in the 1900's. However you can still niche and compete in niche areas. it takes more time and more brains, but to me that is a sound strategy, and it is the very basis of the Web as first conceived.

The Web was not invented to sell things, but to share information. Its now that the first real realities of that are kicking in. Keep that in mind in web strategy.

All which leads to my opinion that in my very small world and important keyword areas, it is one of the best Google updates for a while for relevance and finding good info.

[steveb]

"All which leads to my opinion that in my very small world and important keyword areas, it is one of the best Google updates for a while for relevance and finding good info."

I certainly agree. This update is better than the previous ones, but of course there will be imperfections. If somebody searches for "widget" it certainly makes sense that widget.com rank highly. There is just too much emphasis on URL though. Google should tone that down a bit. Still, considering the buying pagerank and other issues, this update was a pretty darn good response from Google.

Linking, content and titling

...need all be valued. We moved from being perhaps a little too far on the linking extreme to perhaps being a little too far on the titling extreme. I expect we'll move a little bit back next update now that the dips who paid for pagerank see what a waste that was.

[subway]

chiyo I think you're right, backward links to information based sites are much more genuine compared to backward links to commercial sites that in actual fact would not willingly link to another competitor, hence back links to information based sites should and I think do hold more weight.

[jacon4]

chyio, i generally agree that google is an informational type search engine used for research mainly i think. it has been dead in my commerical field for over a year and it appears that other commerical areas are getting hammered now as well. i simply found other ways to market my site as i believe others here will do once they get over the shock. the thing is though, if google continues to rank sites exclusively with machines ( puters ), as the web grows wont the same thing happen to informational type sites eventually? it seems to me that google has a design flaw by ranking sites with puters ONLY.

[glengara]

For me, the 'fly in the ointment' of the argument that this update favours informational pages is some of the dire results on 3/4 word queries.

[subway]

I think the title of this thread should be "Has Google Shot Itself in the Head?"

Even 2 word searches are returning dire results. The whole alltheweb thing is looking very attractive now.

[Liane]

Don't get me wrong, I have no personal complaints about the last update. I'm quite happy with the overall results and the update had very little affect on any of my sites. But as someone suggested here, I did a search for "place name accommodation" and the first result was a site map page for a yacht charter company?

True, the company had a link to accommodations in that place ... but I can tell you that it is definitely NOT the "most relevant" result Google could have or should have served up.

I did a few other searches for similar things such as "place name resorts and although the first result was definitely a winner, the second result was not.

I have to conclude that there is definitely something amiss with the algo, despite the fact that results in my particular area are virtually unchanged and remain very relevant.

[xbase234]

Generally speaking, I do not see major changes in my areas of search. I still see a lot of sites missing, but I am still able to find relevant info.

[Ready To Roll]

Since I normally get lousy rankings from Google for my commercial sites (even though I'm constantly having my pages spidered recently), this update didn't really affect me, although one of my new non-commercial sites is in the top ten for it's main keyword out of 350,000 :)
However, yesterday I was searching for a local production company that could produce a low-budget show for me. I entered keywords such as 'los angeles', 'video production', & 'low cost'. Nothing usuable was returned. The results were either companies not in los angeles, not low cost, or not video production companies. Or they were government or Yahoo directory results.
Went to good ol' AltaVista, entered the same keywords, and BAM, found a great company on the first page (near the top!) that we might be going with.
I'll always try to get my sites ranked well on Google, but as far as using a search engine for relevant results, well.....

[Paully]

Immediatly after reading Ready_To_Roll's post. I hopped over to altavista to run a few queries.

IMO, Altavista is "caca". Very limited, and non-relevant results for my searches. I think I will stick with Google for now.

[SlyOldDog]

I did a few searches today to find some pretty specific stuff. I only needed 1 search to find what I wanted.

Google is definately not broken! .... sadly.

I see Alltheweb has an Alexa rating of below 200. That means that even if Google really was broken, nobody would notice Alltheweb was better anyway since they must have a tiny fraction of Google's searches.

[Ready To Roll]

Trust me, I'm no fan of AltaVista; I'm just saying in this case, it worked for me a lot better than did Google. Tried Alltheweb, and it seems to be returning some even juicier results!
So, in this particular, unbiased, unscientific instance only, the rankings of the search engines used, in terms of relevance, are:
1. Alltheweb
2. A/V
3. Google
Not that it means a darn thing (and I'm quite sure my old statistics professor would have something to say about the validity of the experiment), but it does feel good for ME to actully put GOOGLE at the bottom of the rankings for a change, instead of the other way around ;)

[creep]

I have a number of 100-300 page sites that normally rank well for the topics these sites cover. This update my sites are all on the back pages as the direct result of people buying unrelated high PR domains and putting 1-3 page sites on them and linking them all together with anchor text. Cant google's algo see that these are all 1 page sites jammed with the same words? Maybe they could factor in a size of site filter that would kill these things. I checked tons of other topics and it's happening everywhere. PR and relevancy should be tempered with site size to filter this spamming out. It has gotten worse every update...But this time it is out of control. If this isnt stopped i see alot of normal SEO's will have no choice but to join the club and where will thius leave the quality of the index?

[David_1cog]

I've been reading all the contending theorys over the last few days and still a little uncertain as to what has happened. I've found sites that disprove the 'DMOZ is King' theory, and one of my own sites shoots the 'incoming links are paramount' in the cajones.

Purely based on a couple of my sites, I think Google has become far more sensitive to keyword repetition. My own site took a minor nose dive, and the only thing I could pick up between it and my client sites was keyword repetition:

'1element widget design : widget design : widget accommodation : widget optimisation : atlantas europa'

Too many 'widgets' is my theory. I've done little investigation so please shoot me down if you've got evidence to the contrary.

[stuntdubl]

This is actually the best assessment I've came up with as well. I seem to be able to find evidence that discredits nearly every other theory, but I think this is what really nailed my site this update as well......my question is will stick, and should I start stripping words.

[deejay]

It's the one thing I have been able to settle on.

On one site I have a number of 'catalog' type pages each for a different product set... the only real difference between them is the keyword density on the page, according to the number of products in each set. No changes in links, PR, anything else - I basically took a month off.

The one pattern has been a rise in the lower density pages, and a fall in the higher density pages. Those at 8-12% have risen (including a couple of number ones - yippee!)... those over 18% have fallen... in one case from position 32 to 61.

Unfortunately I haven't really got any that fall between 12 and 18% to find the balance point.

[jaytierney]

Do you mind if I ask which tool you use for determining keyword density?

[stuntdubl]

Wow...thanks for the percentages deejay....that's the best concrete evidence I've heard yet this month.....to bad it'll change next month though:)

[David_1cog]

...should I start stripping words

I already have and it worked. :)

My (simplistic) view of this update is that Google moves closer to its idealogical goal:
http://www.google.com/webmasters/dos.html
... relevant and appropriate content, acknowledgement from your peer (theme) group, penalising 'spam' (defined above).

This update is not perfect (one of my sites went to #1 from nowhere - undeserved IMO) - but I'm happy if the simple SEO techniques get clipped - <title> , <description> , <h1> , <alt> , etc. etc. because it leads to manipulation by the relative few (?) who know about it.

That's my 2 Euros for tonight.

[deejay]

jaytierney - I use a couple, but those figures are from the keyword density analyser at searchengineworld - I run it with titles included, but meta keywords and descriptions EXcluded for Google.

<add> should have mentioned - the percentages quoted are for a two word phrase. If you're targetting single words I expect the figures to be lower, or higher for three word phrases</add>

stundubl - it never ends :) I'm sure we're all closet m@sochists.