Update Saga. Part 5

[Brett_Tabke]

What say you?

Over and done with?

All done all through?

[reseller]

Dayo_UK

>>Huh? - so you are saying that it was never said - and just what you think might happen?<<

Yes.. I think so.

[Dayo_UK]

>>>Yes.. I think so.

I dont. :(

Google dont seem to be able to get to grips with the situation and as time goes on more and more sites get hit.

Not good.

[McMohan]

I am not sure if solving the Canonical problem is all that difficult. Will it not solve the problem if Google by default considers www.dmain.com as the canonical URL, and credit all links to the canonical?

What issues we may foresee in doing so?

[petehall]

I'm seeing home pages out rank more relevant inner pages quite frequently.

Personally I don't think this offers the user the best result, I'd prefer them landing directly on the correct page.

Not on all search phrases though. Seems to be a bit random.

Anyone else seeing this?

[Dayo_UK]

>>>Will it not solve the problem if Google by default considers www.dmain.com as the canonical URL, and credit all links to the canonical?

It is technically wrong to do this unless Google are 100% sure that content on the non-www mirrors the www.

Technically it is possible to have different content on:-

domain.com
www.domain.com
shop.domain.com
fish.domain.com
etc

There would be no way that Google will credit all links to the www by default.

The problem arises when Googles can not correctly identify when the non-www matches the content on the www.

Petehall,

Yes, obviously that is an important part of the algo that the page specific to the search outranks the page linking to it (eg the Homepage etc) - however for me it is the other way around - the homepage can not even rank for its own name, "www.domain.com" search etc... - beaten by internal pages.

I am not going to call it a Canonical url problem anymore - I am going to call it a problem when Google can not determine the root url - as this AFAIK is why rankings drop - I have a feeling that the problem determining the root page though is very much connected to the Canonical problem, 302 hijack and maybe supplementals.

[zikos]

"I'm seeing home pages out rank more relevant inner pages quite frequently"
Yes I do,though I should be happy because my site hits #1 results in some targeted fields but the "money" keywords are RIP ,my theory on that is
1:the whole index update will start after all DC's migrate to J3
2:the index will stay like that until the next PR update
3:Matt or GG said something like "for those that doing well (even in a small chunk of cheese)things are positive....with Jagger
4:If some of us got some top ranking in some fields does that means that within the coming flux we will see some light at the end of the tunnel?
Comments are always welcome.

[LegalAlien]

Morning everyone ;)

My local default is 66.102.9.104 today, and I was happy to see this -- until I checked our serps...

What happened last night? We're now out of the first 50 pages for all but 1 of our key phrases. I monitor 10 2-3 word phrases that have all been steadily progressing back to their pre-September top 10 positions, and were listed every day since their re-emergence on October 21st. As of yesterday, all phrases were in the top 30 results.

The site can't be under penalty, as one phrase is still #8 for a very competitive term (63 million results). How can this just be flux after 4 weeks of constant listings for us?

Someone please tell me not to worry and that you're seeing the same. Reseller, any chance a steaming hot cup of cappuccino to calm my nerves? :))

[McMohan]

Dayo,
Agree technically www and non-www can have different content. We are not talking about sub-domains anyway here.

I was to write infact in the previous post, make www the default canonical, if "www and non-www are identical and both return 200 OK status" which will solve the problem in more than 90% of the cases IMO. In cases where the content is different (my suspicion is, it is a very small %), then leave them as is, with each canonical powered by their own links. I guess it will be a very simple solution to a large problem, (Unless I am missiing any other tech possibilities here)

[Dayo_UK]

McMohan

But Googlebot does not visit the page at the same time - eg. the non-www might be visited on the 1st November and the www on the 15th November.

Lots of sites change there homepage daily, weekly, hourly - or even on every visit (different ads run from a database or something)

Sooo - unfortunately that is not the ideal solution.

and as far as we know 90% of the time Google do get it right now - so there will always be sites missing?

>>>with each canonical powered by their own links.

So the whole sites get indexed twice - hmmmz - as long as results from the domain.com/www.domain.com are restricted to 2 and it does not result in onsite duplicate content penalty then I wonder if that would work.

Not ideal - but better than the mess at the moment?

Zikos - I am very much hoping that once J3 has fully been implemented we will see a big crawl based on the J3 infastructure.

[edited by: Dayo_UK at 1:47 pm (utc) on Nov. 21, 2005]

[McMohan]

Dayo

But Googlebot does not visit the page at the same time

Nor does the algo work at the same time the bot makes a visit.
Obviously I don't have the stats, but I am fairly sure the instances of sites having different content on www and non-www is far less than the no. of sites suffering from this Canonical mess now.