|September, 2002 Google Update Discussion - Part 1|
Discussing the major changes that took place
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..
|GoogleGuy recently dropped a hint that they are watching the numbers of domains owned by someone to try and tell the difference between SEOs and mom & pop shops. |
If this is the case, what will they do when domains are registered by proxy and there isn't any personal information to be found?
Also, with the price of domains being so low, why can't mom & pop shops own more than one shop? Haven't we all learned, in one form or another, not to put all our eggs in one basket? Are people being punished for diversifying?
GoogleGuy was refering to peopel who churn through domains. He was talking about 30+, and there is no penalty just for having those domains.
I personally believe that they are much more sophisicated than that in identifying those that are involved in SEO. I can easily see them implementing a points system, and if one "owner" seems to be pushing the line in too many ways they will ding all the sites.
Yes you can get by it with proxy ownership, but that is additional hassle.
I'm new to the forum but have read for a long time. I have a question regarding the new google update for September.
Vitaplease - I ran the test you spoke of:
"Please check for yourself using the function: "allinanchor: keyphrase" and comparing those results with a normal search for "keyphrase" in both www and www2"
The allinanchor results for my keyword phrase had my two different sites (same keywords) at the number 2 and no. 4 spots out of 2.4 million results. However, I went to www, www2, and www3 and one of my sites is no. 43 and the other is no where to be found....at least in the first 250 results. What the heck? I don't understand the correlation here at all. My results with the site that is now no. 43 fell from the no. 11 spot it was in for over a year in this last update I did lose an odp listing but it still shows up in the links results.
Another one of my sites that was no. 12 for about ten months fell to no. 45 and I haven't made any significant changes to the site.
I link pretty heavily internally using my keywords linking back to the main page and only have a few external links using my keywords.
Any recommendations or input from anyone would be great!
One of my domains which links to my other sites and does not have any links back to it, and it is not in dmoz has a PR of 5. But that could be due to the subdirectory sites within it that have many people linking to it or subdomain sites but many of the subdomain sites have lower PR respectively.
I got into this business to make a little money to supplement a rather small retirement check. My site is modest with about 500
pages and the pages have good content. I was fortunate to have
top 5 rankings on several keywords and was getting commissions
at the level that was ok with me - nothing to write home about however.
I have made no changes to my pages and feel really bad at Googles
Page views have decreased 45%
Clicks have decreased 35%
Commissions have effectively stopped.
Why Google? Why?
|Let's find out what happened - Sept. Update... |
Well in my quest to do this I have been looking at the SERPs and the sites that are now somehow finding themselves at the top of them.
I have come across a technique, a very unsavoury one at that, which without a shadow of a doubt would be classed as SPAM in a big way, and I'm talking 3-4 year old techniques. This technique is getting a particular site to #1-3 for over 50 competitive keywords. So I find myself in a moral dilemma
Do I report this to Google and get ZIP, NADA, NOTHING in return or do I change my whole outlook on SEO and apply the technique to my own site, forget the whole genuine well meaning optimising techniques and live it up at the top of the SERPS undeservingly?
This begs the question.... have Google hit the "Perform Rollback to 1998" button or did Google never actually resolve the problem of spam in the first place, did they simply perfect the art of getting so many quality sites to the top of the SERPs that you never even noticed SPAM was a problem because you rarely searched onto page 6-10?
Now they've changed their algo, they've exposed all the SPAMMY sites again.
Apart from everything that's discussed here - and maybe my question is stupid - is there any hint that Google is trying to reduce the influence of the Yahoo directory? I mean, it's 30 September and I haven't seen anything yet about what is going to happen with the Yahoo-Google relation. It is obvious that this month's update is somewhat different from the ones we have seen since quite a while, so could it be part of a bigger change that a separation of Google and Yahoo would certainly be?
I agree that some spammy sites have returned, but I wouldn't say the index is full of spam. I can't imagine any algo that could prevent all spam unless the spam filters become clairvoyant ;).
If you think you found a particularily nasty form of spam, why not share the tecnhique - without being site specific. Perhaps it has really been around a bit and has just surfaced within the terms you are using.
I think I remember GG saying they don't respond to those email spam reports or manually remove sites but instead try to build filters to automatically catch the spam.
Nothing stupid about that question, what happens tomorrow could explain an awful lot.
My site has seen some very interesting changes lately. The overall result is more or less the same though.
Imagine that I previously had top ranking for the search 'fuzzy widgets'. Well, since the update I am on page 8 (nowhere!).
However, if you search for 'fuzzy' only, I am number 10 (bizzare).
If you search for 'blue fuzzy widgets' I am #1 (reasonable).
The most interesting thing is that I have no external links (that I know of) and the pages in question are PR2.
In other words, my ranking for these searches has been seriously affected by the changes even though I have no external links affecting it. Use of the 'widgets' keyword apparently damages my results.
Question: How can 'fuzzy' rank so much higher than 'fuzzy widgets' when 'fuzzy' on its own has much less relevance to my content.
<footnote>I just read this back and I'm confused.</footnote>
I want to apologize for the previous post in which I was was happy about the new index because many of the spammy sites have gone away. I realize my comments offended some people whose sites were dropped. The sites I was referring to were the spam health sites where they have instant herbal or supplement cures for any and every health condition, making a lot of promises to people that aren't backed up by any reasearch. I really don't like these types of pages, especially when they rank higher than some of my pages. :) I was very happy to see them gone.
I realize after doing a bit more surfing and reading more posts today that besides the spam a lot of legitimate small sites that were not spammy also got caught in the last algo, and I'm sorry to hear that these sights were dropped. I have had sites dropped in the past and I know its a very depressing.
thanks vitaplease...you've just clarified exactly waht I need to do to compensate for our drop in ranking
unfortunately it means getting people to alter the text they use to link to us...fortunately this is also going to be a requirement of our current rebranding anyway
however it means we are currently getting terrific positions for searches on an organissation name that ahrdly anyone would ever search for and which ceased to exist as an active entity this morning :(
not Google's fault
I just want to say that my results have improved. It seems to me that the updated content certainly has weight. We have new content a few times per month.
As well I would like to also note that before the update from Google, we had alot of sites showing up under search terms that did not belong.
Here is what I see for my few hundred Google battered sites.
1. Backwards link anchor text is lower in value unless the PR of the source site is over 5. PR 4 on a site sending a backward link used to be of value....not anymore (at least in a competitive environment).
2. Anchor text for sites with a high PR for the category linking to another site makes that site relevant to the search term. This is a complete nonsense. A site that links to another and uses the anchor text of "blue widgets" now gets above the other site just because it has a higher PR. For regional sites this is a disaster (listen up GG) a site that sells blue widgets in Orlando Florida with a PR of 6 links to a site that sells blue widgets in New York with a PR of 4, now the Orlando site appears above the New York Site for a search of "New York Blue Widgets".....total madness. The only text related to the term is in the link anchor text?
3. Spam is back and is good. Want to sell Blue Widgets in Los Angeles simply put "Los Angeles Blue Widgets" in hidden text 50 times on the home page and you will? Even if the site has a PR 2 or 3 lower than the competition. Do I recommend this? Only on a domain you are willing to waste as I am sure Google has to see this is a nonsense sometime soon.
4. PR is king over consecutive KW's. A site with 3 keywords randomly scattered across the page will beat a site with consecutive keywords just on PR. Try "Burt Lake real estate" and what do you get? Not a real estate site in Burt Lake Michigan, but a site in Illinois that happens to mention "Lake" county and as it is a newspaper and it happens to have an editor who's first name is "Burt"....call this a relevant search....I call it nonsense.
It makes the top spot purely on PR, no regard for relevance at all...Google look at the positioning of the words relative to each other and get off the PR hobby horse to the current extent.
This doesn't only happen on this search term it happens on thousands. Should the user type the search in quotation marks if they want relevance? Yes, they should, but they don't as we all know, and now the results are nothing short of junk for regional web sites.
5. Dmoz....ohhhh gawd. I'm an editor, take it from me many of the folks who edit the ODP are mostly a dozey bunch....now Google gives an even higher rating to the first bit of the description they allow to go through. The Dmoz rules are clear...."don't duplicate the category or title in the description"....but look at how many editors allow that to happen....no way should Google allow this to weigh heavily on the search results. Plea to google, show the description but don't count it as relevant. I don't allow my categories to be manipualted, but many editors do. It is just another form of spam.
6. Alt text on images....seems to have been thrown out of the window.
7. Corrupt index. This is an intersting one. Can anyone explain why the same URL is appearing multiple times on different pages of SERPS? I have numerous searches that show the same domain at positions 7, 17, 27....all of which are root index URL's? Beats me!
8. The update now seems to be totally dynamic. Submit, submit, submit. I have changed the same site 4 times since Sept 26, each time it has been reindexed and repositioned based upon the changes. New cache copy...whole eight (not a mistake) yards. For all those that are paranoid that Google is out to get SEO's....I don't think so. Allowing this quick re-index of sites is playing into the hands of SEO's, not doing battle with them. Unfortunately the PR is not being recalculated, but view getting PR higher as a monthly job and SEO is far from dead.
9. Internal links and site size. This seems to depend on how competitive the search term is. Is the Adwords algo at play here...I don't think so. Google seems to have taken certain keywords and applied a differnt set of rules to them, other very closely related terms are not affected. For example the search term "<location> real estate" produces a set SERPS based primarily upon PR, but "<location> homes for sale" totally ignores PR, or at least diminishes it to a tiny extent.
The overall affect seems to be that the KW's Google has picked as very competitive terms now produce some very odd results (based purely on PR), but sites in those categories can still do well with lower PR's using the majority of keywords actually used.
This brings us back to content. If the site has a lot of content it is likely to continue to do okay, even if it has a lower PR and doesn't capture the primary search terms anymore.
Size does seem to play an important role for only the very competitive keywords. I have to conclude there are multiple layers of rules being used. On a search using the most popular keyword one set of rules apply, while on searches using slightly less popular keywords a totally different set of rules apply.
For a conclusion here is what I plan to do.
1. Get the PR up as much as possible and ditch all the worthless links. I cross link my sites to some extent, this historically was of great value but now I simply find the wrong regional sites appearing in the wrong places because of PR (yes I have unintentionally helped to junk the results). If the PR of the source link is not higher than the object site it is of no value. For those that don't have the luxury of choosing how to link sites together the only contribution I can make is that you should beg from relevant sites that have a good PR for a link (not easy).
2. If Google doesn't kill the spam soon then I think the only choice is to join the party. Don't do this with a primary domain. Get secondary and subsequent copies and put the spam in those. One day I am sure they will get killed, but so will the competition that is currently ranking high with spam and the status quo will be returned. If anyone from Google is listening...please kill the spam now....this is not a measure anyone will enjoy, but there is no choice. I have hundreds of non spammy sites that are getting killed by the spammers...it is costing everyone money....I can't fix it, only Google can.
3. Don't do anything that damages your primary domains on other engines. Anyone else notice the recent increase in traffic from MSN/Inktomi? I have, significant over the last few weeks. Inktomi has for the last 2 months been ranking sites in a very similar way to the old Google Algo (once the Looksmart sites are ignoed) the trafiic that has been lost from Google has been made up by MSN...or is it that users are already looking elsewhere for quality results?
I have never optimized for Inktomi, AllTheWeb/Fast/Lycos or Altavista, I only ever optimized for Google...ain't it odd all my sites now do great on the others but a huge percentage have taken a hit on Google.
Bottom line....Google needs to get back to basics, it had a great search algo...others have replicated to some extent, but that is not a good reason for destroying a live environment, something that worked, in the name of improvement.
My suggestion...Google go back to your July Algo, work on the improvements in a closed environment and when it truely works release it to the world. Using live data as a test ground will drive users elsewhere and Google will fall as fast as it rose.
This is not a comment from someone who has sufferered (thanks to Inktomi).....it is a comment from someone that now sees Google as one of the least relevant regional search engines on the net.
Welcome to WebmasterWorld Percentages,
I cannot judge for your search queries, but if you have general quality concerns, check this message from Googleguy:
There have been several messages with complaints on two word search queries where too much ranking emphasis is given on one of the two words.
I would agree that there is a lot of inconsistancy with DMOZ title and descriptions. But what to do? I have suggested a general pardon and a release of sufferance by subtracting one or two Pagerank digits throughout the whole directory ;). All distress and discusions on the importance of Dmoz would end..
> 1. ...for my few hundred Google battered sites..I cross link my sites to some extent..
Thats a lot of sites to interlink.
Could that be one of the reasons for ranking lower now?
Check the search function on top of the page, you'll find many threads on the perils of crosslinking.
Welcome to WebmasterWorld, percentages. I get the feeling that you've been reading here for a while?
My gut feeling now would be that anchor text is lower in value, but not just for low PR links. This fits with Google's presumed desire to deal with 'Google bombing'. This approach could lead to more reliance on PR and on-page factors.
Less weight on alt text has been predicted for a while. Anyone else notice a change there?
As for 'totally dynamic', Eric Schmidt hinted at more 'fresh' listings a few months ago. It's pretty clear that Google are putting a lot of effort there.
I worry a little about your plan. "Get the PR up as much as possible" is at odds with throwaway spam domains, IMO. In most cases I'd ignore the latter and focus on the former.
I continue to like the idea of publishing good content that people will want to link to, and that Google should want to list. This takes time, but the long terms rewards can be good listings with reduced likelihood of getting penalised. Quick-buck SEOs may not like this message, but sometimes the tortoise does beat the hare...
vitaplease, I think that subtracting one or two PageRank digits from dmoz.org and directory.google.com is overkill. Sure, some proportion of the ODP is corrupt. With over three million listings in 400,000 categories, 51,507 editors aren't all going to be altruistic. In most categories I'd say that the ODP is less corrupt than other places.
|I don't really understand any of this! I have two sites of my own and two sites I have written for friends.|
One site I did for a friend is a hyphenated (keyword) domain which I chose because my friend's company name domain was stolen out from under him the day his ownership lapsed.
All four sites are doing very well and receive great traffic from Google. All appear on page one in the SERPS and mostly in the top 3 positions for targeted keyword phrases. None have been updated at all recently. My own (money making) site has even had a PR boost this month from 5 to 6.
For the past several months, I watched as keyword-phrase domains surpassed my money making site with my company name domain, in the SERPS, only to be picked off one by one for Google rule infringements. I did nothing to counter their rise in the SERPS, choosing instead to see if the Google algo would pick up their tactics. And it did! Their penalties had nothing to do with their domain name though. (AT least I don't believe so. Otherwise, my friends' site should "theoretically" have suffered the same penalty ... and it hasn't.)
My own sites have a tremendous amount of valuable and pertinent information in regards to the theme of each site. They deserve their positioning in the SERPS as well as the respectable PR they both enjoy. They are clean, and contain no tricks used by a great many of my competitors.
Yes, I optimized my sites ... but only after I believed I had found the proper keyword density using my money making site as the test vehicle. That took a year and a half of testing. Each month, I would change a word here or there (and I mean a word here or there ... not whole paragraphs) and then I would wait to see, what, if any, changes it would make in Google's SERPS.
I only link to sites that I believe my customers will find informative. I actively search for "quality" sites to link to me about once a month in the hopes of increasing my traffic and not for PR purposes. If they insist on a reciprocal link, I read their entire site before making a decision. Often, I will just let it go or will only offer to link to pages that are relevant for my customers.
I have taken the advice from many of the senior members and mods as well as GoogleGuys's input and have not strayed from what they told me in the beginning:
1) Supply the information you want your customers to have.
2) Design the pages so that they are easily navigable.
3) Don't use any tricks that are specifically listed at:[url]http://www.google.com/webmasters/dos.html[/url]
4) Don't participate in link farms.
5) Check out your competitor's sites and build a better one with more info.
The only thing I have done on my own was testing for keyword density. It has worked for me. The only penalties I have suffered from Google (which had me scratching my head for a while) were related to having orphan or duplicate pages on my site (twice) ... and that was my own fault. I forgot to take the pages down when I was testing pages for www3 compatability.
I certainly feel for all those who are being financially impacted by this and other updates, but I am completely in the dark as to why these things are happening to you if you have done nothing outside of the rules Google provides.
Spending time trying to deconstruct or analyze Google's algorythm is pointless IMHO. Put your energies into building a really good, content rich site and stand back from the cash register! Chi Ching.
|My gut feeling now would be that anchor text is lower in value, but not just for low PR links. This fits with Google's presumed desire to deal with 'Google bombing'. This approach could lead to more reliance on PR and on-page factors. |
I tend to disagree with this in that I am hearing arguments from both sides. Last week I posted how easy it would be to use PR as a modifier on the anchor text. There aren't many > PR6 sites that would be involved in the "go to hell" campaign, so let's set the magic number at 5. We will change the portion of the calculation from
(link_text * (linking_page_PR / PR5))
Using the linear model this gives a PR5 link the same value as it always had, and a PR 2 site is less than halef the old value. If we go with the values of PR as 6^PR# then anchor text from a PR8 site would be able to make your site, while PR3 anchor text is almost worthless.
I don't actually believe this is the exact method that they used to do this. But I also don't think that they totally threw out link text.
Devaluing low PR anchor text would surely get rid of the google bomb, without getting rid of what can actually be a quite useful feature.
It sounds like you are doing it the way I am planning to do it. I'm glad to hear that it is working for you.
I will continue to play guessing games with the Google algorithm, as I and a programmer and a mathhead. I just cannot resist the urge to figure out a puzzle. I'm actually the sort of person that Google would hire (if they moved out of the hell that is the SF bay area)
Will I ever figure it out exactly? No. Does it matter? No.
I will make some minor changes to make GoogleBot happy, but I will not mess with anything as far as trying to trick GoogleSearch. (and I wish people would make the distinction)
Actually, I wouldn't draw too many conclusions (even if I could ;)), and even less I will touch my sites now and try to optimize them according to Google's "new algo". Google as we know it listens very carefully to whatever comments and complaints there may be, although less so from SEO's, and the news that their September update was a bit over the top is going around (I'm sure that GoogleGuy is here as well, even though he doesn't comment on anything since he would be flooded with 150 questions within 10 minutes). I would rather wait for another month and I'm sure that a good part of the changes made in September will be taken back end of October. Or, as a publication that I trust a lot, puts it:
|It should be noted that these changes may not be permanent. If you've had a significant devaluation in your page rankings, don't over react (yet) to this change. Google may very well do some tweaking to their formulas over the course of this month and your results may improve. |
Allthough the serps look a bit spammy for some words these days, i wouldn't call every wrong result spam.
I have a 100% proof that google has a bug (or simply new, strange algo) with redirects. I'll try to explain with a example:
- a search for "birds" returns "www.mice.com"
The word "birds" neither is in the content nor within any backlinks nor within the URL. Yes i know, google only shows the PR x+ backlinks, but "www.mice.com" is promoted only for "mice" - only - more than 30 backlinks with "mice" in anchor text - none with "birds"!
- then, when you try to go to "www.mice.com" you'll be redirected to "www.birds.com" ... ahaa!
- and the URL "www.birds.com" can't be found with google.
I'm 100% sure, google takes the keyword from the birds.com-Domain ("birds") and adds it to the keyword index of the landing domain ("www.mice.com"). Google then returns "www.mice.com" when you search for "birds", since only one domain should be listed for such mutiple redirecting domains.
Unfortunately i can't give you the real world keyword/domain example since it's in my area. But i have the 100% proof of that behaviour!
Can someone check this in his own area, please?
... no, i'm not crazy after this update! :)
You may be right - certainly something appears to be wrong somewhere.
One of my keyphrases (an industrial product with c.900,000 results) has lost almost all previous top ten.
The dropped sites were all very relevant.
The No.4 result now is a page from a museum about an artist - its not targeting the phrase and only mentions it in 'broken' form. It is almost impossible to accept that incoming links could mention the phrase.
So I say 'wait it out' - I think this will be fixed on the next update.
Anyone else seeing the good old <no frames> tag getting a bit of a boost this time?
Only got a couple of framed ones - both have seen a major boost
All optimized sites this month made monumental gains, dynamic and static. Anchor text was big factor in this update.
A few sites of mine that held for 3 years at #1 on competitive terms were bumped down a notch by higher PR sites. I know what I have to do to get them back to #1, though.
[edited by: xbase234 at 4:21 pm (utc) on Oct. 2, 2002]
I think they might be tweaking their algo already to put more weight bck on relevance instead of PR.
Example: My real name is pretty common. Not as common as Jane Doe, but pretty common. There's an artist with the same name as mine whose site always came up before mine in Google. She has her name in the title of a page. One of my sites would always come up next, even thought my site has higher PR, because I just mention my name in a few places and not in any title page. I'd been meaning to make a page with my name on it so I could try to be first in Google for my own name, which would make it easier for my friends to find my site instead of having to remember my domain name.
After the last update, my site came up first for my name and the artist was second. Great, I thought, now I don't have to make a page with my name on it. But a few days after the latest Google update, the artist's site came up first again. I'm not sure if anyone else is seeing examples of this, but I thought it might signal a tilt back to more weight on relevance instead of PR.
I'm still trying to make sense of this update.
Backward links, PR are important factors but are no substitutes for content and during this update I do not see site content receiving its due.
This I think is a major lapse.
OK... we poured our guts out in the previous posts about how good/bad the last update treated us... and what we think may have changed.
I've just checked some regional stuff and it's as if I walked into a demolition area. To say there has been some changes is a masterful understatement. It's not that the changes have happened, after all, Google is pefectly entitled to make whatever changes they deem necessary. It's the results arising from the changes that seem to impact on Google's reputation.
I'm seeing cases where every site sppearing on page 1 is off-topic and a good site (not mine) has gone from page 2 to page 26 for pete's sake! That's one almighty reassessment of a site's value.
From the perspective of Joe Public doing some searching I'd be wondering why the world's best search engine is not offering me what I asked for. Isn't that what their reputation is based on?
Leaving aside the impact on your own sites (come on, be objective!) is anyone else seeing inferior search results in their pet areas... in other words is Jow Public still being served the best sites available?
> is anyone else seeing inferior search results in their pet areas
I see absolutely no change in the results of the types of searches that I do, and must say that I do not have a clue what everyone is moaning about.
I see drastically improved results on one search term. And that includes pushing my site back into the 100+ place in the serps for that term!
I have seen a definite change in results, with the folks who had been first page now being a few pages deeper ... but the folks that are now at in the first page are not spam - my personal opinion is they're not as good as last months' first-pagers though, in terms of quality and PR.
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.