SES London: Link Building And Domain Name Issues

[Garrett]

The Search Engine Strategies London conference started a week ago and I'm glad that Alan Webb (http://webpronews.com/authors/alanwebb.html) attended and shared his observations over at SEOChat (http://forums.seochat.com/t11609/s.html). I've gone through his posts and lifted what I think are the most important SEO tips - be sure to check out his complete posts.

Webby (as he's called in SEOChat) attended two sessions on link building, and, as search engines pay so much attention to who's linking to you I thought that'd be a good place to start.

From the basic link building session he reiterated an old link building maxim he thought was the "general thrust" - "create quality content that is going to get linked to naturally from the same or similarly related themed websites."

Matt Cutts, a Google developer, said that "thematic incoming links from authority sites carry more weight than on-page optimization." Wow! That's something I've suspected for some time, especially after watching the Nigritude Ultramarine contest rife with wiki blasting, but to have Google say it is certainly reason to start beefing up on your inbound links.

Thomas Bindl of OPTOP (http://www.optop.de) suggested donating your services to a charity site. Especially if they give you a link, as charity sites are often both considered authorities and have high PageRank.

I found it interesting that, in addition to link farms and guest book entries, forum posts are "increasingly counting for nothing."

In the advanced link building session Mr. Cutts emphasized that the quality of links is more important than quality. But again, as demonstrated in the Nigritude Ultramarine SEO challenge, quantity clearly has an effect.

If you're not alternating your link text, especially if you're buying links, you may be harming your rankings. Identical link text, "looks like (and 9/10 times is) a sign of manipulation of its PageRank algorithm." Natural links, such as those appearing here in this article, use varying link text. Be sure to mix things up if you're linking for the express purpose of raising PageRank.

It was also suggested - by Google I assume - that you only buy links for the traffic they may provide. It's apparently easy for Google to tell when sites allow links for pay because the links go off to such unrelated sites. Not that this changes the fact that links to your site give your site better ranking. I guess if Google can tell they may work something into their algorithm, but link vendors will just spread out their links more, or create themed sites based around their clients.

The sandbox theory (http://www.webpronews.com/insiderreports/searchinsider/wpn-49-20040506GoogleSandBoxEffectRevealed.html) got some talk, though Matt Cutts denied it existed. On one panel it was suggested that sandboxing may only occur when "a site launches and all of a sudden a large number of links point to it with the same link text."

Alan brought up an issue he has with his own site that he had resolved - he's got both German and English versions of his site and has separate folders for them. Now it's considered "better" to have separate domains with "language/country specific top level domains for the different language content."

http://www.abakus-internet-marketing.de/ is his German language URL.

http://www.abakus-internet-marketing.de/en/ is his English language URL.

Here's what he learned he should do:

Find a web host in the US and/or UK to register my .co.uk / .com pages. Making sure there are no trademark violation issues first.

Move the english language content from my German tld /en/ directory to the uk/com domains. Change the .co.uk content to be more in line with the uk market (avoiding duplicate content and helping conversion by specific Geo targeting). If I have a .com then change the content to be more US focused.

Use 301 moved permanently in my .htaccess on the .de domain to make sure that there is search engine friendly redirection to the English language websites (.com or .co.uk) (Never use meta refresh or javascript redirection). (Yahoo has issues with 301s (http://www.webpronews.com/insiderreports/searchinsider/wpn-49-20040607YahooBustingUp301Redirects.html) though -- ed.)

Pursuade all those linking to my old /en/ folder to switch to the new .co.uk domains / .com domains.

Alan also included some very specific copywriting advice he heard from Charon Matthew of MediaCo. Charon recommended "four keyword phrases (NOT keywords was stressed) in 250-300 words or for longer pages 8-10 times in 500 words."

(If you need help in finding phrases consider checking out this keyword formula (http://www.webpronews.com/insiderreports/searchinsider/wpn-49-20040604FormulaForKeywordConnectivity.html).)

Ammon Johns, who spoke at the conference and was the inspiration for my no hats SEO article (http://www.google.com/url?sa=U&start=1&q=http://www.webpronews.com/insiderreports/searchinsider/wpn-49-20040420NoMoreHatsHighRiskvsLowRiskSEO.html&e=747), wrote about his experiences in the cre8asite forum (http://www.cre8asiteforums.com/viewtopic.php?t=10615). He also attended the WebMasterWorld Pub Con.

Thanks to Barry and the SEORoundtable (http://www.seoroundtable.com) for the tip. Be sure to check out his review of Alan's post (http://www.seroundtable.com/archives/000530.html).

Again, here's Alan's complete posts (http://forums.seochat.com/t11609/s.html), and special thanks to his Abakus Internet Marketing (http://www.abakus-internet-marketing.de/en/). (Alan, just let me know when you change your domain and we'll get those links pointed in the right direction.)

(Thanks to Martinibuster for correcting a misattribution - according to him and Alan Webb it was Thomas Bindl who suggested volunteering work for a non-profit site, not Mike Grehan.)

[cbp]

Good report, especially to hear this from the 'horses mouth':


thematic incoming links from authority sites carry more weight than on-page optimization

CBP

[Dave Hawley]

"thematic incoming links from authority sites carry more weight than on-page optimization."

Now, if we just get an authoritive definition of "authority sites" I suspect it's not just a high PR (whatever high is).


I found it interesting that, in addition to link farms and guest book entries, forum posts are "increasingly counting for nothing."

Hear that Mel and Fathom :) Something I have suspected for a while now. See: http://www.webproworld.com/viewtopic.php?t=20911


It was also suggested - by Google I assume - that you only buy links for the traffic they may provide. It's apparently easy for Google to tell when sites allow links for pay because the links go off to such unrelated sites

Hope you are reading all this Mel ;) I too have suspected Google would NOT have too many problems identifying and NOT crediting paid links. I'm VERY curious to know if this would also include paid directory listings. I would say it's very likley Google doesn't simply pass on PR based on the category page PR and number of links.

[achronister]

It was also suggested - by Google I assume - that you only buy links for the traffic they may provide. It's apparently easy for Google to tell when sites allow links for pay because the links go off to such unrelated sites

I am still having a hard time with this "unrelated sites" statement. I can see how Google could determine in a limited manner what sites are related, but this is putting a human element to the algorithm with I don't believe they can accomplish. Is a fabric site related to a car site? It could be if you wanted to upholster your seats. In my case, how about casino sites, hotels, limo services, weather info, etc. None of these are on topic to concert tickets, sports tickets, etc but all are services that my client base uses.

IMO nearly any site could be related to any other site in several different ways, and I don't see any way for a SE algorithm to determine that correlation.

This statement claims they can determine paid links "because the links go off to such unrelated sites" but I still don't see how they can penalize this when linking to "unrelated" sites could easily benefit the user.

[Dave Hawley]

but I still don't see how they can penalize this when linking to "unrelated" sites could easily benefit the user.


I doubt very much they "penalize", simply not credit PR etc. If this is consistent, which I would think it is, it would do more good than harm. Remember, it's not likley possible to reverse engineer anothers site.

However, I hear what you are saying. IMO Google would er on the side of caution (not pass PR) when/if they are to pass credit on for a link. Unfortunately their hand has most likely been forced, thanks to the shonky SEO business practices of telling customers to spam forums, guestbooks, FFI unrelated sites, etc as a way to gain PR.

[Andilinks]

IMO nearly any site could be related to any other site in several different ways, and I don't see any way for a SE algorithm to determine that correlation.


You are right about relationships between and among sites--these are things that even humans have trouble with.

But I can see how Google could allow a threshold of x number of "apparently" unrelated links before assuming that there is abuse.

Don't forget that Google really does have the "big picture" and can not only make comparisons in the aggregate but also has a history on a every sites attempts at abuse and can easily develop a profile for various SEO trends. These SEO trends can be identified as they spread and SEO practitioners (and SEO schools of thought) can be identified by their profiles independent of IP's, domains, geography etc.

I expect Google to be just a few steps ahead of SEO trends as they appear--that is their business after all and they seem to be good at it.

Andi

[cbp]

I doubt Google will ever penalise for unrelated links. They may just give more weight to related links (maybe not for PR, but in the relevance part of the ranking formula).

I agree, with how can they determine relevance:
eg anchronister eg above fabric = cars; there also real estate and house content moving businesses; etc.... as I can see no logical way that Google can determine these, that they just probably won't do it ...

So, all they can do is give more weight to links from sites that are on the same theme (which is different to related) - my best guess is that they determine the theme of a site by using the same technology (CIRCCA) that they use to determine relevancy for Adsense.

CBP

[cbp]

Just thought about what I wrote - I think they may penalize a site for unrelated outgoing links ... not a penalty as such, but part of the "quality signals" they talk about. eg links to link exchange websites could be seen as a 'poor quality signal'.

CBP

[Andilinks]

eg links to link exchange websites could be seen as a 'poor quality signal'.


Which would be a small part of the 'abuse profile' that I wrote about above. Who, after all, would know more about SEO than Google, looking through the other end of the telescope, as they do? I think the corollary to PageRank would be AbuseRank and every website probably has one, it's just not published on the toolbar.

Andi

[debtfree95]

Thanks very helpful info

[martinibuster]

From your newsletter:
"Mike Grehan of Site Interactive suggested donating your services to a charity site. Especially if they give you a link, as charity sites are often both considered authorities and have high PageRank."

That is incorrect. It was Thomas Bindl of optop.de who said it, at the link building forum at SES in London.

[bhartzer]

OK, so Thomas Bindl gets the credit. Whoever invented the idea, it's still a great one--getting a link from a charity site.

[WebRankInfo]

It's very interesting but I'd like to know precisely how to find authority sites for any topic?
I saw that Teoma gives "link collections from experts and enthusiasts". Maybe it can help?

Have you got any definition and a way to find authority sites?

Olivier

[Andilinks]

Have you got any definition and a way to find authority sites?

An authority site is one that is widely regarded as an accepted source of expert information or advice. Unless abuse is distorting the value of PageRank there should be a direct correspondence between PageRank and authority since by definition people link to authority sites because they are authorities.

The Google directory will list sites in any given category in order by PageRank. Authority sites will appear near the top.

Somehow though I get the feeling that this is not the answer you wanted.

Andi

[martinibuster]

Garrett's mistake (http://www.webpronews.com/insiderreports/searchinsider/wpn-49-20040608SESLondonLinkBuildingAndDomainNameIssues.h tml) about Mike Grehan's quote is still not corrected.

This quote:
"Mike Grehan of Site Interactive suggested donating your services to a charity site. Especially if they give you a link, as charity sites are often both considered authorities and have high PageRank."

Is wrongly attributed to Mike Grehan. It was Thomas Bindl who made it. In the interest of accuracy, it would be nice to see someone take thirty seconds of their time to correct this error. It makes the reporting look sloppy.

[edhan]

Does linking really promote high Pagerank? So far when I am searching for a particular product with Google, the top 10 sites appearing does not necessary have high Pagerank. So far I am happy with Google search term since I always get what I want within the first 2 pages. Hope this will be maintained down the road.

Edward

[Dave Hawley]

Does linking really promote high Pagerank?

Yep, as a general broad answer. However, PageRank is only one of MANY factors used to determine the relavency of a page in the SERP's.

[jackson992]

Good report, especially to hear this from the 'horses mouth':


thematic incoming links from authority sites carry more weight than on-page optimization

CBP


You mean bad report:)


This just confirms for me how illogical Google is

[fathom]

I found it interesting that, in addition to link farms and guest book entries, forum posts are "increasingly counting for nothing."

Why?

...and did this read "absolutely nothing". (note: #1)


In the advanced link building session Mr. Cutts emphasized that the quality of links is more important than quality.

Why?

...and how can you produce the former by switching the design model around... think "mainpage" and why it is usually the single most referenced page in SERPs for most websites? (noted: #2)


But again, as demonstrated in the Nigritude Ultramarine SEO challenge, quantity clearly has an effect.

Factor of Note: #1 - all links from pages indexed in Google that Googlebot can and has permission to crawl and are penalty free adds something.


If you're not alternating your link text, especially if you're buying links, you may be harming your rankings. Identical link text, "looks like (and 9/10 times is) a sign of manipulation of its PageRank algorithm." Natural links, such as those appearing here in this article, use varying link text. Be sure to mix things up if you're linking for the express purpose of raising PageRank.

Duplication of anything is easy to find and discard.

Factors of Note: #1 and Note: #2

How can you remove this "duplication" while retaining what little you get "per link" to produce "the quality of links" as apposed to "quality"? (while some might think that's a typo - it isn't).


It was also suggested - by Google I assume - that you only buy links for the traffic they may provide. It's apparently easy for Google to tell when sites allow links for pay because the links go off to such unrelated sites. Not that this changes the fact that links to your site give your site better ranking. I guess if Google can tell they may work something into their algorithm, but link vendors will just spread out their links more, or create themed sites based around their clients.

BINGO!

You really are debating the wrong things here...

Sharon & Roy have hit a key point - when things are taken out of context you miss important information.

Unfortuately, you "must engage in some research to truly gain a definitive competitive advantage".

[Sharon & Roy]

If you're not alternating your link text, especially if you're buying links, you may be harming your rankings. Identical link text, "looks like (and 9/10 times is) a sign of manipulation of its PageRank algorithm." Natural links, such as those appearing here in this article, use varying link text. Be sure to mix things up if you're linking for the express purpose of raising PageRank.


Hi Garrett,

So are you saying that those poor site owners who are providing images or text links with prewritten HTML code on their "Link To Us" pages as a courtesy for their visitors to copy and paste onto their pages if they would like to link back, that they will be harming their rankings if they ONLY provide 1 link text and/or 1 image for their visitors to use?

And if alternating your link text will keep these site owners from harming their rankings, is alternating 2 enough or should it be 3 or maybe more?

As an aside if there is anyone who can provide any proof that not alternating link text harms rankings, we'd appreciate it if we could have a peek.

Thank you!

Your Friends,

Sharon and Roy Montero

[Dave Hawley]

I don't buy the "alternating your link text" theory. Google themselves state it near impossible to reserve engineer anothers site.

[Brittany]

Hey everyone,

I just wanted to let you know that I've split this topic. All posts and discussion regarding Sharon and Roy's 'unrelated sites' post can now be found here (http://www.webproworld.com/viewtopic.php?t=22682&postdays=0&postorder=asc&start=0).

Have fun discussing!

Brittany

[cbp]

in another thread, rcmedia2004 wrote:

Hi all,

I just read this infomative article about the Google convention in England recently at >>>
http://www.webpronews.com/insiderreports/searchinsider/wpn-49-20040608SESLondonLinkBuildingAndDomainNameIssues.h tml

One part of the article concerns me greatly...
"If you're not alternating your link text, especially if you're buying links, you may be harming your rankings. Identical link text, "looks like (and 9/10 times is) a sign of manipulation of its PageRank algorithm." Natural links, such as those appearing here in this article, use varying link text. Be sure to mix things up if you're linking for the express purpose of raising PageRank."

We use the same anchor text exclusively... will this lead to problems for us?

Thanks in advance.
_________________
Rob.C


(I got an error message trying to merge the post to tis thread, so have just cut and paste it - sorry)

CBP

[Dave Hawley]

We use the same anchor text exclusively... will this lead to problems for us?

I highly doubt it. There is just too much evidence against this theory. Take big name companies like Microsoft etc. They have many thousands of links with the anchor text "Microsoft". They seemed to be ranking ok in Google last time I checked!

Google also states it is near impossible to reverse engineer anothers site. If using the same anchor text causes harm, all competitors need to do is link to each other with the same anchor text many times over.

Just do what is best for your site visitors and Google will look after you.

[fathom]

We use the same anchor text exclusively... will this lead to problems for us?

I highly doubt it. There is just too much evidence against this theory. Take big name companies like Microsoft etc. They have many thousands of links with the anchor text "Microsoft". They seemed to be ranking ok in Google last time I checked!

Google also states it is near impossible to reverse engineer anothers site. If using the same anchor text causes harm, all competitors need to do is link to each other with the same anchor text many times over.

Just do what is best for your site visitors and Google will look after you.

Unfortuately this isn't what was said.

1. 1000 page website

2. 1000 external links

3. links all go to same URL

4. Same anchor used

is easily detectable... thus can be made to be "worth less" when used in calculating ranking order.

This is also the suggestive idea by Matt Cutts that "sig links" are increasingly becoming worth nothing - the pattern is easily detectable.

[Dave Hawley]

1. 1000 page website

2. 1000 external links

3. links all go to same URL

4. Same anchor used

is easily detectable... thus can be made to be "worth less" when used in calculating ranking order.


What's the difference in that and what I stated?


This is also the suggestive idea by Matt Cutts that "sig links" are increasingly becoming worth nothing - the pattern is easily detectable.
I *proved* that to myself (as you know Fathom) long before Matt Cutts was *suggesting* it. At the time it was being proven, yourself, Mel, Ferret, Janeth and many others all blindly took the most popular opinion that forum signatures do help with with Google rankings.

Janeth was the only one big enough to admit she was wrong.

[fathom]

1. 1000 page website

2. 1000 external links

3. links all go to same URL

4. Same anchor used

is easily detectable... thus can be made to be "worth less" when used in calculating ranking order.


What's the difference in that and what I stated?

Here you say "I doubt it". (to same link anchors)



This is also the suggestive idea by Matt Cutts that "sig links" are increasingly becoming worth nothing - the pattern is easily detectable.
I *proved* that to myself (as you know Fathom) long before Matt Cutts was *suggesting* it. At the time it was being proven, yourself, Mel, Ferret, Janeth and many others all blindly took the most popular opinion that forum signatures do help with with Google rankings.

Janeth was the only one big enough to admit she was wrong.

Here you say "I proved that it will not work". (to same link anchor)

A sig is a link with a defined pattern > so is site wide links... if it works, it works, if it doesn't, it doesn't...

Your not entirely wrong - but your not enitely right either.

Change the pattern they work "exceptionally well".

[Dave Hawley]

Here you say "I doubt it". (to same link anchors)

I said "I highly doubt it" to same anchor text links harming Google rankings. You said "Unfortuately this isn't what was said." Then appear to proceed to agree with me.



Here you say "I proved that it will not work". (to same link anchor)
No, I say that forum signature links don't help with Google rankings.


A sig is a link with a defined pattern > so is site wide links... if it works, it works, if it doesn't, it doesn't...
If you mean by "works" they boost Google rankings, "it doesn't".


Your not entirely wrong - but your not enitely right either.
Which bit is not right?


Change the pattern they work "exceptionally well"
What does, forum signature links? If so show me an example.

[fathom]

Done!

Example created!

[Dave Hawley]

It's pity they don't prove your point though.

[fathom]

It's pity they don't prove your point though.

what... you haven't seen anything yet! :-)

[Will.Spencer]

I have a number of web sites which share footer links.

On 11 June I created and deployed a PERL script that creates pseudo-random English sentences as anchor text for each link in the footer.

Approximately one month later the test results clearly show that this has had no effect.

This causes me to believe, until I see evidence to the contrary, that changing or dynamic anchor text will not "fool" the Google algorithm.

You can examine the script in action at Fort Liberty (http://www.fortliberty.org).

[Dave Hawley]

you haven't seen anything yet! :-)


That's my point.