I hate to revisit this topic again so soon, but it appears to require clarification for many folks. I've tried to lay out the process of how this theory works as clearly as possible, and provide a few examples.
In our example above, the W3C page (a highly valuable resource on web standards) linking to your page can have an impact on the rankings of all the pages on your site. This isn't just a replication of the classic PageRank flow model (PageRank in its original form is nearly dead, IMO), it's an indication of all the tasty factors that the W3C can pass to your site making the search engines take your domain's content a bit more seriously.
The general premise seems simple to me - search engines provide recognition to a domain based on the trust and editorial weight it has received from current and past links. Content on a well-linked to domain often outranks far more relevant or individually well-linked-to content on other, less linked-to domains. Although I get a lot of nodding heads when I explain this, the theory seems to break down in practice. People I've talked to hear this theory and still wonder if they should link build in to multiple domains and then link all those domains to their main site. They ask if they should host their blog on a separate domain or how to build content for a domain they just registered and I'm always left thinking - "hey, didn't we just go over this?"
If the above illustration wasn't enough to convince you, let me give another quick visual on a related note. In this diagram, it's important to think of the linking pages A-G as equal to one another (and on different domains than one another).
When I say "worse," I mean a lot worse. There's no scientific way for me to quantify how much less the example on the left is worth, but my spidey senses tell me it's about 1/9th of the value. That's not good - you should want 9/9ths of the value for your link building campaigns, particularly since attracting links is one of the hardest things search marketers (and website owners) have to do.
Here's a few practical applications of the "Rising Tide" rule:
- Don't launch your blog on a separate domain
- Don't create a series of junk domains that will link in to your primary site (spending that $7 each year on Ho-Hos is a better investment)
- Don't write great articles or content that goes on other sites just for the value of the single link back (there are other marketing and branding reasons to do this, but it doesn't make sense from a pure link-focused persepective)
- If you have completely separate businesses or passions that you want to dive into, see if there's a way you could relate them enough to justify building a single general or broad subject website, rather than creating many.
- If you're a big brand launching websites to go along with offline media campaigns, 301 that site back to a subfolder on your main domain.
- Unless you're in the business of building and selling domains, it almost certainly pays more to concentrate on one single domain and make that site the most authoritative it can be.
And now a few examples of this effect in action:
- In a search for my name, Wikipedia's page about me (which was just recently created) is ranking ahead of thousands of more valuable, targeted and well-linked-to pages, but since the Wikipedia domain is so strong, even new content with barely any links gets tremendous respect from Google.
- In a search for 24 season two, Amazon.com's page, with less than half the external links and lower direct relevance than Fox's 24 page is ranking #1. In fact, there's almost a dozen big-name domains with few external inbounds ranking ahead of the official page.
- A search for oatmeal reveals that SEOmoz is a significantly more powerful domain than many popular breakfast brands (and that Matt, our developer, has a wonderfully twisted sense of humor).
- Michael Martinez likes to point out that Matt Cutts' blog has pages in supplemental as a juxtaposition to my arguments above, yet when we look at the pages on Matt's site that are in supplemental, they rank first, even for topics and terms that Matt is rarely associated with, like this search for languages, this one for howto files and this one for firefox vs ie stats. Note that this is also a good example of why you shouldn't worry about having lots of pages in supplemental (neccessarily), as long as your rankings are kicking tail :)
I hope this blog post helps lay the issue to rest a bit. If you've got more examples to share, I'd be grateful - it's late and I'm tired after the drive back from Portland. More on the SEMPDX show tomorrow, BTW :)
Matt's sense of humor is great - you can't miss that title in SERPs! Whether or not you'll find what you expected is another story. But I still like the warm, fuzzy feeling I was left with after that comic relief from a SERP.
I know! That cracked me up!
I thought that title would increase click-through rates, which will make me ridiculously rich and powerful.
...but then you'd lose all the money to the powerful oatmeal lobby for defamation... :D
I've been dealing with this with a client recently who has several frankly vast sites, with LOTS of high PR deep pages, and all of them pointed not to the domains, on the home page links, but to...
...you guessed it, index.aspx files. Sorted this out on each site, and sorted the links between each site, and rankings leapt up overnight. Not insane, straight to number 1 from nowhere stuff, leaps of 15-25. But still, enough that it illustrates the point perfectly.
So yes, listen to the Rand. He speaketh the truth.
Pete,
I've never even thought of pointing links to the homepage at a domain level. On every page of my website I have the homepage link as /index.asp
Are you advocating that this should be changed to https://www.DOMAIN.com ?
Not as long as you're consistent.
Meaning you should only use one preferred url and stick to it and 301 all other options to your preferred url.
But since most people would link to the domain only, it probably is wise to use www.domain.com, but it really doesn't matter as long as you use one consistently and 301 all other options.
Case in point
Seomoz is currently optimized for ranking for "home" and "blog"
Rand (and other great commenters)
What about when you have a bunch of sites other than your main site that already have top ranking for a bunch of the relevant keywords? Would you 301 those "gateway" sites even though they give you current traffic through their top rankings? Are you suggesting that doing so gives you the full 9/9 benefits that you would have achieved if you had kept to the main/umbrella site in the first place?
This is a long-winded way of asking whether you think that the current position of the G site(s) on the SERPs matter...
I am actually considering doing EXACTLY that... I have a client in real estate that has 1 main site scoring awesome for its keywords and 15 micro-sites scoring fairly well for its keywords...
Since I just took on the project, I also realized he was doing reciprocal linking with all of its micro-sites...
Anybody, like Peter, Rand or Andy, or any other great soul around, would recommend that we take all the blows now, build a powerhouse main site with 100-150+ page of content instead of the mere 30 we have now, and redirect all links while losing some SERPs in the top10 from the micro-sites? Any help is more than appreciated!
I always tell people never to take anything I say for granted, and to do their own testing and take actions based on the results.
In the last few months I have consolidated a few sites into one, and also launched plenty of satilite sites that are targetted at particular niches, sometimes even if they could theoretically have been published on an existing domain
Rand suggests consolidation around the same time that Yahoo launches a whole load of niche sites.
How many people actively link to About.com as an authority source? I am sure they have something on SEO, have you ever read the pages or subscribed? The traffic there seems to be declining.
Should all the Weblogs Inc blogs be hosted on AOL.com, and their old domain names be used as redirects? Would such blogs be treated in the same way by Technorati or the search engines?
I highly respect the teachings of Leslie Rhode and Michael Campbell, yet most of my blogs would be looked on as massive ball linking.
My sites have an insane amount of what might be looked on as duplicate content, but in February 50%+ of my traffic came from Google search, and half of that entered on duplicate content pages.
What is more important to Google, Trustrank or LSI?
How comes when you search for Digital Camera in Google, most of the first page of results are niche sites?
I am not an SEO
I don't take anything for granted and spend a lot of time testing and tracking.
In your situation, you have lots of satilite sites - would it hurt to merge 20% with the hub and see if results improve?
Take another group of 20%, and turn one of them possibly into tagspace that is shared with the others. Test track and compare.
The problem is that it is hard to do these things if you are working with someone elses VRE.
Guillaume - you know me, I'd go for the small hit now in the hopes of gaining back those rankings in a few weeks (301's do take time to propogate through the engines) and possibly becoming a true leader in the field. The power of all the domains together and the value of having the content receive the ranking power of one large domain far outweighs the micro-site strategy.
The only time I really recommend against combining sites is if you want to sell one or more. It's nearly impossible to sell a "section" of a website, but if you have several small or mid-size properties on different domains, you can sometimes get more individually than you would for a large domain (depending on the niche).
Well, there is a 30pages home site on a very precise niche, and 16 other domains which are 1 pager on the vertical niche as well, just theme targeted pages. I think this is a whole mess, but I'm not sure that redirecting 1pager domains (especially since all of them are on different C IP blocks) to only 16 new content page on the homesite would really help.
I also have another issue: the pay per click campaign had like 16 adgroups, one of them going to EACH micro-site, and none to the home site...
I suspect the last person was trying to send traffic to those "1 pager site" to then get tons of referrals to the home site with great anchor text. Although, most of them just have PPC traffic, very little inbound links and are more on steroids than anything else.
My call would be to build one mighty site and send all Adwords traffic to the homesite as I'm doing the redirects, but I fear losing alot of the acquired SERPs...
What do you guys think, Rand, Andy and maybe Peter or any other? Have you ever had a situation like this? The traffic is rather low as it is a very precise niche...
Thanks for any feedback on this
I would go the Legal SEO route
1. Continue using those single page sites for PPC as a highly niched squeeze page, but maybe build them out to 4 or 5 pages for possibly better relevancy
2. Footers are wonderful things, and certainly governments in Europe are making things tough for webmasters insisting that you include contact details on every site by law. Including a link to the parent site from the footer couldn't ever be looked on as something naughty.
3. on the parent site create 16 new pages that relate to the content on the previous 16 seperate domains, and include a contact form and address details on each. Have the only link from the pages either to topic hub or homepage.
4. Avoid interlinking the 16 squeeze pages, but feel free to turn them into 16 minihubs gaining their own links.
5. There is always the option of creating additional sites that link to the minihubs
Hey Andy. Thanks for your reply.
Just a quick question: let's say those 16 1-pager
Don't you guys just HATE trying to explain to clients how page rank doesn't matter like it used too??? And then you still get emails from them saying - "We have a GREAT opportunity to get this link on a PR7 site!!!" - then you have to go back and explain it again....argh. Reminds me of a great line from Ausin Powers, "I eat cause I'm fat, and I'm fat cause I eat - it's a vicious cycle".
And I don't necessarily agree with you on not launching a blog on a subdomain...I think in some instances, it's beneficial :)
I often think of it as like lifting up a tarpaulin - the more lift points you have, the higher you can lift the entire sheet. The sheet obviously being the website inventory.
However, the imact of the lifting effect is dependent on the competitiveness of the marketplace - less competitive areas equals stronger lifting impact on your pages, but very competitive areas means it's more easily swamped by competitor activity.
Surprised you've had to clarify this, though, Rand - you must have a lot more padawans than I realised. :)
Rand this is a great and timely discussion. Thanks for sharing your input and explanation. I've recently been explaining this very concept to some clients, and you have done it very well. :)
Good article - makes a lot of sense.
One question... I'm optimising a real estate site at the moment. When the company started, they just focused on one geographical region. As they've grown, they've started to cover other regions and countries, each with their own domain. They've since added in an umbrella domain to cover the lot, but that site's got no unique content on it
If they were starting from fresh, they'd be best off focusing on the umbrella domain. But in this instance, however, should they leverage the trust in the first domain as a short-term solution, or focus on the long-term by developing the umbrella site and 301ing the others to it?
Keen to hear your thoughts! :)
Ok, here's how I'd do it...
1) Take all the content and put it on the umbrella domain
2) Set up each place as a subfolder, so the architecture breaks down like this:
https://www.sitename.com/location1/content/index.htm
https://www.sitename.com/location1/inages/nicepicture.jpg
etc...
3) 301 all the domains so that you get the weight coming in from them
4) Post-move housekeeping. Make sure everything works, any directory links are changed to the new URL, get in contact with partners to repoint their links...
But yes, I'd take the hit now, and go for the long term gain of being on the umbrella domain.
My ONE caveat with this would be if you are targeting primarily "local" searches and have field offices in every city/state that you're expanding into. In that case, it can actually have some weird value to split up the domains, because you can rank in the local results for each (as long as you register with all the local listing services).
Note how Google will show a map and 3 local businesses on a search for something like "real estate agent miami" - those are the kind of listings you could target with separate sites. Just be aware there's some benefit in weird cases to avoiding my advice above... Maybe that's just how the search world works - like the English language, lots of exceptions to the rules.
Rand,
You mention "local" searches may benefit from multiple domains - what about international searches ? What if a company targets multiple countries...? I would assume this justifies more than one domain. Although not a straight forward one as you have to be careful with things like dupe content and building up authority for multiple sites... What's your take?
Joe
Cheers Pete and Rand - good to hear your thoughts on this. Appreciate it! :)
Hey when can we get custom menu backgrounds for our member pages...
https://www.seomoz.org/users/view/30 - Oatmeal's got his own :(
oatmeal is alot cooler than us. it's high school all over again.
You guys should check out the one Matt made for Lisa.
Hey no fair. I want to play too.
Maybe a little customization of profile pages is something you could add as a benefit of premium membership.
To get your own menu background you have to mail me something incredibly shiny and expensive. Covering the package with flattering poetry about me wouldn't hurt, either.
I mention about.com the last time you posted on this subject. As someone else commented, they seem to be doing well with subdomains.
If your theory was true, wouldn't Geocites, Tripod, and all the other subdirectory big-name hosting domains be ranking a lot better? They have a ton of links pointing to them across the whole domain.
I do think the "authority factor" of your domain will have an impact on rankings of all pages... but maybe not as much as you think it does.
You need to study how well About.com interlinks between subdomains.
Wordpress.com in many ways does something similar with their integrated tagging system.
The most important factor for me is the ability to niche and interlink relevant content, whilst maintaining a good user experience.
If I remember correctly a few years back the user generated subdomains of Geocities and Tripod DID receive much higher rankings than they deserved and attracted a lot of spammers who threw up many subdomains in order to cash in.
After it was public knowledge it was corrected. Not sure how, manually or automatically, but Google found a way to find out that those user generated subdomains had nothing to do with the main domain and therefore received less link love/authority.
tbfpa - you took the words right out of my mouth; nice call. I think this holds true for many sites (excuding Wikipedia) where UGC or user-created sections are possible.
This doesn't take into account the internal linking structure both of the site doing the linking, and of the destination site.
Wikipedia has powerful interlinking that makes individula pages relevant, and that is sometimes true of blogs as well.
I will give you another example:-
I now rank above Matt Cutts for Toolbar Pagerank, and Matt has used that term in the title and URL of the page I am outranking, and I am sure received more backlinks.
https://andybeard.eu/2007/01/toolbar-pagerank-ball-linking.html
My whole site is in supplemental most of the time
Andy - I wonder if you've got personalized or geo-targeted results. I see Matt ranking in position 7 and I can't find your domain in the first 50 results...
Maybe it is your datacenter
I even ran it through your keyword difficulty tool (though there were some bugs in the report)
Top 10 Ranking Sites at Google URL Page Strength Report Date 1. https://toolbar.google.com/ 9.5 9.5 / 10 03/08/2007 View Report 2. https://forums.searchenginewatch.com/showthread.php?t... 5 5 / 10 03/08/2007 View Report 3. https://www.searchenginegenie.com/seo-blog/2004/09/to... 3.5 3.5 / 10 03/08/2007 View Report 4. https://www.google.com/support/toolbar/bin/static.py?... 6 6 / 10 03/08/2007 View Report 5. https://www.top25web.com/pagerank.php 3.5 3.5 / 10 03/08/2007 View Report 6. https://en.wikipedia.org/wiki/PageRank 7 7 / 10 03/08/2007 View Report 7. https://andybeard.eu/2007/01/toolbar-pagerank-ball-li... 3.5 3.5 / 10 03/08/2007 View Report 8. https://www.internetbasedmoms.com/seo/google-toolbar.... 2.5 2.5 / 10 03/08/2007 View Report 9. https://www.mattcutts.com/blog/new-toolbar-pageranks-... 5 5 / 10 03/08/2007 View Report 10. https://www.websitetrafficbuilders.com/google_page_ra... 2.5 2.5 / 10 03/08/2007 View Report
I meant to ask earlier, and I only skimmed the comments on similar questions, but does the term "junk domain" also apply to variant domains. I'm eventually creating site right now that's called "The Comic Book Reinvention Project". It just seemed intuitive to me to buy variant domains like www.reinventingthecomicbook.com, or the(dot)comicbookinvention.com, etc. I didn't want some hack to come along and divert traffic with a similiar URL.
Yeah - we bought dozens of variations on web2.0awards so we'd own all the potential domain traffic for the terms. However, the best thing you can do is just to 301 those back to the main site.
Want to point out another reason why it may be a good idea to have seperate websites, instead of one large one.
When I first started I had one big site, well it really was the only site I had... So if you want to test something, the only site you can test it on is the big one.
I made some mistakes and that site got punished in all SE's and made me no money at all anymore. Then I decided to throw up a lot of different niche sites, so if the rankings for one of those sites plummeted I at least still had good rankings for the other niche sites.
For me this worked out really well and I am happy I did it this way. It's just another thing you can do to make sure you don't put all eggs in one basket.
Mind you I don't work for anybody, they are all my own sites. It probably is different if you work on clients sites, as you can't do any testing on those anyway, but still, having several sites can minimize the risks sometimes.
That sounds like a good plan. Do you think that there is any type of website that is suitable for all of the eggs in one basket. What would be the characteristics of such a site?
I would say the topics have to be very well related to eachother.
For example if you have a site that sells all kind of shoes, like women, men, girls, boys, athletic, casual, sandals, boat and outdoor shoes etc., it would be better to have one big site.
If you have a site that sells all kinds of not related items such as shoes, t-shirts, camping equipment, baby clothes, cameras, jewelry, music, food, etc., it would be better to have different sites for each category as you would otherwise be competing with the truely authority sites such as Ebay and the like and you would have a very slight change to beat them in their own game.
So IMO the theme of the website is the most important subject in order to determine if you should go with one or more websites. That and the number and quality of the competitors and the profitability of the niche itself.
For example if the highly ranked competitors are all true authority sites with broad subjects, it would be better to start a niche site.
If the highly ranked competitors are authority sites with a single subject, it would also be better to start a niche site and hope you will gain enough authority to outrank them in the future.
If the highly ranked competitors are non-authority sites, you would have a good change to outrank them with a broader themed site. But you have to be carefull that another person won't start a highly themed website and overtakes you...
IMO the more competiton you have the tighter the theme of the websites should be. And because the competition will only become bigger, I feel that the theme should be as tight as possible in most/all circumstances.
Basically I think the time for websites with broad subjects are numbered and it would be better to start niche sites. Either you can't beat the already existing authority sites anyway and/or the SE's will give less weight to them in future in favour of dedicated niches, IMO.
One drawback is that they should bring in enough money to justify a site with just one subject. And I think if that is not the case, SE's will continue to rank the broader themed authority sites anyway, so it will be better to forget about that niche entirely.
That's funny about typing in oatmeal and how well he ranks. One of the panels I went and saw at SXSW was with Matt Mullenweg from Wordpress and one of the things he said was along the lines of, "If you want more info, just go to Google and type in matt and click I'm Feel Lucky." That kind of blew me away. He even beat out Matt Cutts.
Superbly put, as usual, Rand. I love your anecdotes about the name ranking. :)
I note that this was written in 2007,which may affect your answer, but...
I believe it was on this site where I first read how to build an internal linking system on a site. I note that it is the exact opposite of your "worst" external scenario! If you put a mirror below your "worst" diagram, it is exactly how I was taught to build internal site structure (assuming that what is shown is one pillar for one key phrase you want to rank for).
Is it true that external link building and internal link building are done in an inverse method?
Great post. Investing power in a single domain seems to be of increasing importance and that's the route I'm recommending to all my clients. Our parent company's site has tipped over recently into being a 'power-domain' (for want of a pithier descriptive term) and has started ranking for tonnes of theme irrelevant terms on the basis of various odd pages scattered through the site - mainly client portfolio sites. Of course, this pleases our clients no end when they see our portfolio page ranking well ahead of their own site...
In my experience creating sub domains using the content keywords as the sub domain title is very effective especially when there are full URL links to the actual domain's home page.
Generally sub domains are considered as 'not part of the main domain' and Google/Yahoo etc will index them as if they are seperate sites.
Sub domains are effective popularity building link options for SEOs
Rand,
That ranking for "oatmeal" is awesome! Especially for not doing it intentionally. After seeing that I know how you can monetize this blog... Post about some commercial products and slap two big squares of affiliate links or contextual ads in the top of the post... and change the title tags to something that seduces clicks.
Or, you can leave the "ground up pigeons" title tag in place and see how much the big oatmeal brands will pay you to stop stinking up their SERPs.
Thanks for this post -- we've just been discussing whether to go with https://www.jibjab.com/blog or https://blog.jibjab.com -- I had a hunch that it was better to go with the former, but you've helped to solidify that opinion. Please let me know if there's anything incredibly obvious that we should also consider when launching a new blog on a site that already has a fairly high page rank (7/10) and 1,000s of inbound links.
Thanks!
Dave
VP, Marketing (JibJab)
Dave, I'm such a huge fan of you guys. It's great to see you here on the blog and I'll certainly find my way over to the JibJab blog once that launches - send me an email ping on that if you get the chance :)
My general sense is that this has been going on for awhile now and to me it makes a lot sense that it would. I think most would agree that trust and authority have taken on an important part in how well a page ranks. Isn't that what the sandbox is all about.
When you're trusting a page it's hard to put to much emphasis on the single page. It's much more likely you would evaluate the trust based on the domain as a whole.
Right or wrong trust is often associated with the company you keep and in the case of a sngle web page, the pages around that page are the company.
As a human being if I encountered two similar articles one here on SEOmoz and one on Johhny's spam site I'm going to trust the article here a lot more simply because of the other articles I've already come to trust here. Why would a search engine see it any different?
Thanks for posting this info, Rand.
I found it very helpful, I also enjoyed the presentation yesterday in Portland. Great job!
Good explanation.
I have agree that the second diagram could be interpreted different ways. The worse side may also be illustrating how page G on some site, due to its own authority and inbound links, may pass more weight than a bunch of links from smaller, less authority pages/sites.
The points are good, just think the diagram may be open to various interpretation.
Thank you Rand, for illustrating this so beautifully! It's something I've struggled with explaining to potential clients in the past who are under the belief that "more is better" and "get them to the site and they'll find their way" are the best attitudes to have. :)
I've written a post and linked here, and will definitely be directing my clients here when they start to question my process! :)
I admire your simplicity with the pragmatics, great write. I am fully impressionable at this point so i jotted it down.
Thank you.
Understood and makes sense. Has been like this since I first began playing with SEO 3 years back. Than it was explained to me as the " Amazon Effect", but the logic behind it is still the same.
But... Isn't the picture Worse/Better a bit misleading?
One strong link is better than 6 weak ones, isn't it?
And in the Worse picture 6 sites link to "G" and "G" links to your site. Doesn't that provide more value than the Better picture, since that one receives 6 weaker links?
Following the same logic, isn't it a good idea to build links to already existing pages that link to your site in order to build up the trust rank of your site? To me, the Worse picture shows just that. So IMO the Worse picture is actually more valuable than the Better picture, but that just might be my screwed up way of reading it...
I think the poinit is that A-F are the same pages in both diagrams. Would you rather have them pointing at your domain or a page on another domain that points at your domain?
The question is whether there are cases where "worse" is better. For example, if G were on a really well trusted and ranked site with few outbound links and your domain were weak, then couldn't that be a better scenario? Not that you'd want to develop all your content that way, but high-quality links like that might be a good thing to have in the mix.
We're supposing that A-G are all equal with that diagram, sorry if that was unclear. I'll edit the post :)
Would you say all SE's interpret links the same right now, or are you just talking about Google? I thought you did a good job of explaining so users of all levels can begin understanding more about how these processes work. That's one thing I really like about this blog. It's basic, yet advanced, you know?
Looking forward to your updates on the conference in OR. I couldn't make that one.
My view on that is that, they all use (pretty much) the same system in how they look at links, the difference is in how they weight them, and whether there's a cutoff point (i.e. only pages up to two degrees of seperation away have an effect/five degrees, ten degrees etc...)
Also, one link that might carry a lot of weightin one engine mght carry less in another. It's all relative. You just learn over time (hours spent pouring over backlinks for sites) what's going to do what where. As Rand says, it's like a sixth sense, gained just through time spent trawling.
Thanks for the detailed explaination! Makes sense and it seems to be the case for the sites I run.
I'd agree with Pete, with the exception of MSN, who doesn't seeem to give as much weight to a powerful domain as Google and Yahoo!
Think that will change anytime soon? MSN always seems to be catching up on link related algo changes.
Great clarity Rand. This is an excellent guide for a topic that seems straight forward, but confuses many people.
Rand,
Love the post, but I have a couple questions. First off, I don't dispute your arguments, but could you give us better examples? I am not necessarily certain that your 24 season two example is a good one because a lot of sites outweight Fox for that keyword, and are likely much much more relevant.
Second, and more the point, your post made me wonder about something that I would love to get your feedback on. That question is, are there multiple layers of authority? More specifically, a page-level authority and a domain level authority. I think SEOs have a tendency to group them together, but after reading your post, I am starting to wonder whether or not we would be much better off differentiating the two.
Any thoughts?
phantom - when I checked, the Fox page had a few hundred inlinks, while the others had decidely fewer. However, those domains as a whole are far better linked-to. The content and keyword targeting is very much the same, so where's the difference? In the domain strength.
But, yeah, I'll update the post with a couple more examples.
Yes Randfish I agree with you, But PR Flow is important according to me :)
Interesting addition about pages in the supplemental index Rand.
Conceptually, domain weight makes a lot of sense.
If someone, some company or some website is viewed as authoritative and popular, then it is seems logical and justified that everything that person/company/website has to say would be given extra weight, automatically, before being judged on it's own merit...
It's how people think and really, isn't that what Google is ultimately going for?
(Don't believe me? This principal is exactly why bad sequels make a killing at the box office and Stephen King keeps selling novels! ;) )
As to whether the graphics are useful Rand... Well I did replicate this one this morning when this very topic came up. Very useful, thanks.
I agree that the original PR paper is just about irrelevant at this point, but can you really demonstrate that the link pop is being spread to other pages because they're on the same domain rather than because they're linked together as part of a domain? Are you saying the link love is spread throughout the site regardless of its navigational structure?
And isn't it odd that, at least when it comes to links, Reaganomics turned out to be right...
Yeah - my experience has been that it's not based on the "PR flow" but on the fact that as a domain as whole builds up link love, pages on that domain rank better, almost regardless of how far away from the well-linked-to page they are. The Wikipedia example above is a perfect illustration of that; it's a long path from the "Rand Fishkin" page back to a page that gets great link love.
Rand,
First off, I 100% agree with you. Mind if I borrow those graphics to show clients? As they say, a picture is worth a thousand words.
I would actually like to discuss an offset of this subjest. I would love to hear your thoughts on domain vs subdomain and how it affect the same attributes you mentioned above.
I have a few ecommerce clients that are forced to use subdomains due to platform restrictions. From what I have seen I get the feeling that links to subdomains help the main domain (Not as much as if they where in a subfolder though) but links to the main domain do not help the sub domain, or at least not very much.
That theory is based on a gut feeling and things I have seen but of course I don't have a large enough sampling to be able to actually quantify or prove it.
One large test case I can think of that of course isn't mine is About.com. They are almost all subdomains and have some decent rankings but considering how long they have been around, you would think they would be a little more dominant. If they didn't use subdomains, I think they could rival Wikipedia in rankings.
Your thoughts?
There is one situation in which you want to have content reside on a different site. This is the notion of writing an article for purposes of publishing it on someone else's site. Of course, you only do this with someone that has a very powerful, authoritative, site.
This can be quite an effective technique for building links directly to your site. Of course, as you outline above, it almost never makes sense to put your blog on a different domain, create domains, etc.
My point is focused on leveraging the trust of other people's already established domains.
Eric - Yeah, for example, contributing a great article to AListApart.com is a terrific way to spread the word about your authority and value. As I said in the post - there are good marketing reasons to do this, but for the pure link, it's rare.
Can you elaborate on that Rand? If the content is not duplicative of something on your own site and you otherwise wouldn't get a link from that high authority site, why would it not be a valuable link?
Brian - It's not that it wouldn't be a great link, it's just that if your site has the exposure to attract a lot of outside links, that's often preferrable. For example, let's say you write a post on YOUmoz, hoping it gets pushed to the main blog and seen by thousands of search folks. The single link back to your website from that page won't be nearly as valuable as all the links pointing to that post, so if you could have gotten the same attention and visibility posting it on your own site, that would be the preferred decision.
Man, my writing blows tonight. I hope that made sense.
Makes perfect sense, thanks.
I like the concept of guest-writing on a high authority site in order to point a link at a targeted internal page that I want to rank. For example, you have an important tutorial ranking for a keyword phrase at position 11 in Google, and you want to push it into the top 10 results.
What you are saying is that the high quality content that would be required for that guest submission should attract plenty of links on its own, which would boost the overall authority of the domain and perhaps inch up the target page that way. I agree, but as you've alluded to in your answer, some sites have trouble attracting general links, so sometimes you've got to do what you've got to do.
The key point then is how well your site would attract links to that content. The majority of sites aren't going to be able to pull those links so writing for YOUmoz would make sense since the one link could help build the authority that will later pull those links.
For Brian it wouldn't make sense to submit to YOUmoz. For me it would make a lot of sense.
I was thinking the same thing as Brian. Taking AListApart as an example, writing an article makes a lot of sense in spreading your brand and building your authority by being associated with their site.
But say you were to write an article for AListApart under a different name. I would still think that any links back to your site from the article would still be valiuable links.
I agree the authority and branding you would get from writing such an article is worth more than the links, but aren't the links still valuable in and of themselves?
Makes sense.