Forum Moderators: Robert Charlton & goodroi
Continued from:
[webmasterworld.com...]
The same down-and-on problem here in Turkey.
But j3 goes on and off. No steady results. At least three different sets:
216.239.63.104 (I think with additional tweaks)
64.233.161.104 (still J2)
64.233.179.104 (J3)
Having noted this, I would like to refer my personal observations on the general quality of Google SERPS: In the sector I monitor, I should say that Google falls short of being satisfactory in terms of organic results, although there seems some improvements. And top 10 is populated with some irrelevant results.
My general impression is that the more competitive a keyword or a keyrowd combination, the poorer the organic results. When you qualify your searches with additional (and seemingly less competitive) keywords, organic results get better. The same holds true for using verbs instead of nouns in search strings. For example, "translated" yields better results than "translation."
The implied equation that more competitive kw = poorer results reminds me the popular conspiracy theory that Google degrades organic results deliberately to derive more money. And this theory has a sound argument based on an actual dilemma faced by Google:
Assume that Google has perfected its algo and therefore searches always yield good organic results that perfectly fits to the user requirements. That is to say, the user (surfer) finds what he seeks for within the top 10 results. Then why should he need to click those ads?
Apparently, perfect organic results = no clicking on the adword links, and therefore no money for the google! This is a real dilemma for Google in the face of pressure exerted by the greedy investors, and IMHO google is striving to establish a balance which is very hard to achieve.
Got to take a few more english lessons.
it is based on a false idea to start with.
each searcher wonts different things form his search.
so on searcher does a search for widget - but is looking for the how to make a idget site.
another searches for widget but is looking too buy one urgently.
the premise that the same search term used by x number of people all wonted exactly the same thing is false. (sure there are exceptions, but thas what they are, exceptions).
The perfect search result does not exist. O maybe it does but is sure reaches its consume by date real fast. It is too ephimeral a concept.
If this were the case, and I cannot accept that it is, then why do such keywords as 'national lottery'put the UK national lottery at number one spot? They had a number one spot PPC advert for this keyword up until 2 weeks ago, and now it has been removed, BUT the SERP's still show them at number one position.
Surely, they would be pushed down the listings if Google wanted them to spend on AdWords for this keyword.
IMO the SERP's generally show an excellent set of unbiased results, so far.
Saying that, they havent pushed me off the organics yet but the information sites bias is definately there, I think partly why so many page 1 serps are dominated by bbc pages (and others) that are very loosely related to the query.
<edit>Tidied up as I was mumbling ;)
If those irrelevant sites are spamming, supposedly their serps will short-lived, as Google is 'apparently' making every effort to remove such sites. If they're not spamming, then there's something about those sites that makes Google's algo believe that they deserve to be there. Therefore, why don't you spend your time looking for similarities between those sites and the relevant sites? If you find this, then you can then share something other than conspiracy theories with the rest of us!
"Producing good organic results is good for Google and good for searchers"
Not exactly, producing good overall results is good for Google and good for searchers, I was saying that ranking commercial sites via adwords and information sites via algo often makes a better result for the searcher.
In many ways, information sites dominating the organics and commercial sites using the ads makes alot of sense and in the circumstances I described are better for the searcher. Its not a conspiracy theory, its a possible argument as to why google would be pursuing this approach and the benefits to google and its users.
"There are only 10 spots on that first page, and there are a hell of a lot more than 10 companies willing to pay to be on the first page".
Exactly. the ads do a much better job of ranking those 100s of commercial sites all chasing the same money terms. The organic serps for commercial terms are less helpful because they are less democratic and more open to manipulation (and the more money involved, the more manipulation).
Its a reasonable argument IMO, agree or disagree but please dont just put it in the category of wild conspiracy theories
So you're saying that the rules being applied to AdWords should be applied to organic results? Have I understood this correctly?
AdWord listings are targeted on accordance with key words/phrases provided by advertisers. It seems to me that organic results have moved on from this, specifically to stop people from manipulating serps. How would you apply that to organic listings without opening them up to manipulation? I'd guess you'd have to build a complicated algo to weigh many internal/external factors in an attempt to foil such manipulators -- isn't that exactly what Google is doing?
As for commercial/non-commercial terms, from what I see in our logs, searchers are getting pretty good at determining search terms accordingly. What I mean is that users generally know how to search for "digital camera reviews", or "digital photography guide" as opposed to just "digital cameras", and Google seems to be getting better at sorting such differences.
<edit>For the mods -- we have nothing to do with digital cameras -- this was just an example ;))
No, Im saying that the algo for adwords CANNOT be applied to the organic serps (because of no clickthrough stats and no payment, think about how the adwords algo works) therefore for money terms, the adwords algo is often better for the searcher and for that reason Google could well justify the bias towards information in the organic results (i dont think they will because then they would have to admit that their algo cant cope with all the Spam).
"AdWord listings are targeted on accordance with key words/phrases provided by advertisers"
Well, not really, ranking in adwords is based on (roughly)
a) how much you are willing to spend
b) how popular (relevant) your ad is
c) targeted keywords
So you can target what you like but if you dont convert into sales then thats gonna be pretty expensive. Likewise, those companies that are converting can afford to pay a higher cost per click. So the top adwords sites are likely to be the most successful companies or products.
Compare this to organic serps where the top sites are often simply the best at seo (and in some cases spamming) or have been around longer etc.
And now consider that aswell as providing relevant results, googles increases revenue (and why not), thats why I think we are heading this way.
Why should a national airline appear at the top of results for a ticket purchase term? Most likely, national airlines will not even be tarketing the search phrases used to find cheap tickets. Companies targeting such phrases will be companies selling such items. Isn't this how it should be?
Everyone searching for an airline company will know the name of the company. If I search for "lufthansa", this is #1, just as it should be. If I search for "cheap tickets", then this ambiguous term returns a whole bunch of sites selling exactly that. If I search for "cheap tickets germany" I get sites selling to/from Germany. As I'm searching google.com, then this is how it should be. If I search on google.de I get companies in Germany selling tickets. It seems to me that this is exactly as it should be.
Perhaps you need to be more specific with your search terms/regional choice?
mods -- I don't sell tickets; just an example.
Agreed. I just didn't want to write all that out.
>>> So the top adwords sites are likely to be the most successful companies or products. <<<
Yes, I agree with this. However, I believe AdWord results are secondary in importance to organic results. For example, when you search for something, how often do you click sponsored results? You know companies have paid for these listings, so you are naturally biased against them, just like everyone that has been around for longer than 5 minutes.
>>> I think there will always be a place in the organic serps for commercial sites <<<
I think organic results are extremely important to commercial sites; particularly for determining the trust value of a site. I know this term has been battered in here over recent months, but surely trust is everything. Seeing the same site for various search terms is great branding and goes a long way towards developing trust. I don’t believe this can be achieved with AdWord listings.
>>> ...but the bar has been raised, you'll have to prove your worth via the usual Seo <<<
Yes, but isn't this just the same as with every other major update?
My site is ranking on #1 for a two keyword phrase and is gone from these DCs. Those DCs seem to be alive in Germany. Doing a site: command for two other sites, Google is showing a much lower number of pages indexed.
This is scaring me. I thought it couldn't come worse. Hope this is some flux only and it will pass by.
>>> I am from country A and searcing for "flights to widget" <<<
The term is ambiguous, so if the site you mention sells flights to widget, then why shouldn't it be listed? If you're in Tasmania, searching for flights to Tibet, and you want a local company, then realistically, wouldn't you phrase your search more directly? Perhaps "tasmania flight ticket tibet" or "tasmania buy ticket tibet? These searches produce good results on google.com and google.com.au.
You can't expect targeted results for a general term, and it's not really fair to Google to be complaining about them.
Whitehatwizard>> Google AdWords rewards the number of clicks, by putting keywords higher up the listing, even though they may pay less than ones under them, if the clicks on that advert are higher than one which may be bidding more [hope you understandd that!]
Zykos: I didn't understand your point about the lottery. I also think that it is of no use whatsover, in fact quite the contrary, to blaspheme the USA, just becuase they are the leaders in the search engine field. I see no justification whatsoever in SERP's being biased towards the USA. The USA is of course the biggest market, and as such, has a higher input.
I don't understand. I think we'd better just leave this!
Google is giving you free customers. If you don't want them, some of us will take them and be happy.
Stop the complaining about how people search. The business is as the business is. Take it or leave it. Or adapt.
note to self: must stop checking DCs
>>> results seem to be back to 'normal' on the DCs mentioned above. Maybe somebody hit the button by accident. <<<
I've been seeing two sets of page counts on various DCs for the past few days -- one accurate and one way down. I think the lower count is newer data, as this doesn't match any previous page count I've seen. Not sure what this means, but it's certainly worrying ;(
I have seen the odd dc show vastly decreased number of results only to bounce back again. Been happening for a while.
I still feel that the Google index is going through lots of changes for *some* of us....as posted in another thread - I have seen some homepages comeback which were long lost - not ranking - but hopefully this will come.
Still get the impression that Google cant work out the root pages for some sites (I guess this might be primarly what they are trying to fix)
For website owners who have the correct root page determined in the Google index I am pretty much sure they are wondering why there is still an update thread going - as those sites seem to have stable results accross the serps.
For sites which seem to have the wrong root page determined then results are all over the shop on the different DCs.
>>> For website owners who have the correct root page determined in the Google index I am pretty much sure they are wondering why there is still an update thread going - as those sites seem to have stable results accross the serps. <<<
The www/non-www issue was resolved for us about 2 weeks ago, but our serps are still pretty volatile. We're not dropping serps, but are literally disappearing/reappearing every few days.
Volatile accross the DCs or are the DCs all the same and volatile over time?
Doing a site:www.domain.com check on the DCs - is your homepage top?
I am hoping the sign of the fix is when your homepage is top on all DCs - and then improvements will happen.
Pretty sure that MSN have reached that stage for www.msn.com - and then hopefully re-indexing for the domains will follow (which is why I am not to worried about losing pages at this stage):-
[google.com...]
That is a bit low for results returned for www.msn.com
So I am hoping for stages as follows:-
1 - Correctly crawl the sites homepage (eg correct canonical/hijack problem for homepages at least.)
2 - Correctly determine this is the root page (eg site:www.domain.com returns homepage top)
3 - Crawl the site based on the homepage being the root. Hopefully PR power etc will then run through the site correctly.
I still think some sites are between 1&2 - homepages are not appearing top for the site:www.domain.com search - often in positions 2-8 - however, most homepages now appear to be crawled recently for the sites I monitor.
Once you reach stage 2 on all DCs then results tend to be stable accross the DCs.
I still see varying results across DCs, but only for some search phrases. Other phrases are stable across the DCs. All are 50m+ terms).
For site:www.domain.com our homepage is top on all DCs. site:domain.com shows homepage top, but doesn't list anything without www, except a subdomain. This is listed after all www pages, just as it should be.
Up until 2 weeks ago, deeper pages were appearing above the home page. We also had about 50 non-www pages listed. That's all fixed on all DCs.
Your post grew since I replied ;)
I think your 1,2,3 is spot on. With varying data on different DCs due to divided PR, canonicals, etc, once this all comes into alignment things will hopefully settle down.
However, this new page count data that's on some DCs doesn't fit the pattern, and might indicate another tweak or twist to the saga?!?
Ooops - sorry - I do tend to do that sometimes when I think a bit more.
There also seems to be a stage between 2&3 where the homepage ranks for other searches within your site.
EG. Site:www.domain.com www.domain.com which is obviously a search for www.domain.com within the site - if the homepage appears top for this search then there seems to be increased chance that it is being correctly treated as the root.
I have had a site thats homepage has appeared top on the site:www.doman.com search and is now appearing top on site:www.domain.com www.domain.com searches on some DCs (although not the original J3s :( - which has me worried a bit as these are still different for me for sites I look at that have problems.)
I just still feel that a lot of the problems G is having are related to how they determine the root - and a lot of sites as I said before I think are still in between Stage 1 to 2.
