WordPress SEO – Siloing vs Massive Ball Linking With Tags

 

This is not the definitive guide to WordPress SEO, and I highly doubt I am the right person to write one, in fact I am not sure who would be. That being said this partial guide on a couple of aspects of WordPress SEO might offer a broader perspective on what is possible than I have seen previously published.

I just test things out and track, and I expect my readers to do the same.

I am not an SEO Consultant, in fact if I was I would probably set all my public blogs up in some uniform manner that I would preach to my clients because it would conform to the accepted norms that the rest of the SEO community adhere to – if I create a site that meets accepted norms for my clients, no one could point the finger and say I did a bad job with on-page optimization.
Other SEO Consultants also like to link through to posts that suggest to their clients that everything they have been saying for years is 100% true, so anything that agrees with accepted SEO principles gets more links, and gets read by more professional SEOs.

For you gamblers out there

1+6=7
3+4=7 as well

You get to the same or similar result by taking a different path.

Vicious circle, or benevolent circle… take your pick.

I am not a SEO Geek – I am technical, and extremely creative in thinking up solutions and linking structures, but ultimately I don’t know my way around PHP and .HTAccess enough to be truly one of the “geeks” able to take on any SEO task.

Then again, most of the best Copywriters that I know believe that a lack of a formal education gives them an advantage and can mean enhanced creativity. I can be extremely creative when applying SEO knowledge.

Warning

Apologies in advance, this is going to get fairly “geeky” and I encourage you to bookmark it for future reference, or ask questions in the comments. I will include lots of references where appropriate, and even if this stuff is beyond your current level of understanding, it is a good reference for the future.

I should also warn that this post is quite long and “meandering” – there are 100, 200, maybe more ranking factors and somehow other SEOs are able to place them in a tabular form.
I don’t think tabular, I think in 3D – what could represent some kind of negative factor could also provide positive factors.

Testing

I use multiple techniques, and I test things gradually.

I also haven’t done all the optimization on Andybeard.eu that I would like, partially because Google was reporting bad data that was extremely hard to work with regarding supplemental results – that bad data is also currently still being shown by the Google Toolbar, and doesn’t seem to have been universally fixed for all sites.
I like to write about the changes I make, and like to have some conclusive results to demonstrate both why I might make a change, and also a way of demonstrating the effect.

The aim of this blog was a “work-in-progress” – as I make changes I write about them, or write about the results from testing at a later date.

If you make lots of changes at once, it is very hard to track which one was significant.

Unlike a Sales page, where you can just throw traffic at it using PPC to test conversion rates, with SEO it can take time for things to change.

Questions about SEO?

John Reese has some questions about WordPress SEO

This ties in fairly well with some material I have written in the past, and also the results of a lot of the experimentation I have been doing on this site that I haven’t yet discussed.

I also receive recurring emails on the same subjects, especially when I have hinted at solutions but not yet written about them extensively on this blog. Some of the information in this post I am pulling almost totally from replies I have sent to readers.

SEO Ranking Factors

SEOmoz has a great compiled list of ranking factors, based on the opinions of 37 SEO Experts.

I largely agree with the final conclusions… exceptthe idea that you should optimize for all of these factors, and avoid all the potential penalties in every situation

Top 10 SEO Ranking Factors

All good stuff…

Factor Ranking
Keyword Use in Title Tag 4.9
Global Link Popularity of Site 4.4
Anchor Text of Inbound Link 4.4
Age of Site 4.1
Link Popularity within the Site’s Internal Link Structure 4
Topical Relevance of Inbound Links to Site 3.9
Link Popularity of Site in Topical Community 3.9
Keyword Use in Body Text 3.7
Global Link Popularity of Linking Site 3.6
Topical Relationship of Linking Page 3.5

Top 5 Negative SEO Ranking Factors

It should be noted that there was a huge amount of dispute over some of the following, and only an “average” level of acceptance for the highest items in the list.

Factor Ranking
Server is Often Inaccessible to Bots 3.8
Content Very Similar or Duplicate of Existing Content in the Index 3.6
External Links to Low Quality/Spam Sites 3.6
Duplicate Title/Meta Tags on Many Pages 4.1
Overuse of Targeted Keywords (Stuffing/Spamming) 3.3

A Handful of SEO Questions

John asked a handful of SEO questions and I am going to try to address them

I am not going to claim any of my answers are conclusive, though I have more experience in answering some of them than others.

1. Permalink Post Structure

/name-of-the-post/ or /name-of-the-post.html

I have absolutely no idea. I haven’t tested it, either for click-through or SEO performance.

I chose to use .html on this blog, and in fact on a few more, when I moved content over from Blogspot. At the time blogspot wasn’t providing 301 Redirects so I was using a meta and javascript refresh.

With there only being a change in domain name, and not site structure, I believe there might have been better signals to the search and ranking bots that the meta and javascript redirects were legitimate.

I have seen many experts suggest that making multiple changes in permalink structure is not a good idea.

What I might do sometime in the future purely as a suicidal experiment is switch off the redirects. Most of the links going to my previous domains came from article marketing, so it would drastically reduce my number of incoming links, but conversely change the whole link profile into lots more authoritive links in balance.
One negative aspect of doing that would be the benefit I might be gaining from domain age, and I am sure it would affect specific keywords.

2. Permalink Post Structure Directory ‘Depth’

www.domain.com/name-of-the-post/(or .html) VS. www.domain.com/archives/some/other/folders/name-of-the-post/(or. html)

This gets a little bit complex, because there are lots of issues involved

Using a date in a URL can be an indication that content is “dated” so if you are creating a niche website you might not want to give that visual indicator – conversely, it could be looked on as a positive indicator of how fresh content is – URLs with dates are longer (there are positive and negative issues) – I think this might come down to personal choice.

Even if you don’t use dates in URLs, you will still have date based archives unless you don’t link to them, and block them anyway with robots.txt or a plugin that adds noindex to the pages.

Some SEO experts such as Graywolf would suggest only having an article appear in one category, and only use excerpts in those categories. This is effectively creating a classic tree like structure.

Date based archives are one way a search engine bot can identify your site as being a blog or news site – I am not sure whether that has a positive, negative or neutral ranking benefit.
Article Directories also have RSS feeds, but generally don’t have date based navigation.

If you are publishing a multiple-author blog, there is a chance you might be included eventually in Google News or on Yahoo. I have read (though I can’t remember where) that to be included on Google News there is a requirement to have a unique 4 digit numerical reference within each URL.

3. Tags: Hype Or Reality?

John is looking to know if tag pages on your site can really make a difference.

I have written a lot about tagging with UTW (Ultimate Tag Warrior) – yes that is a tag page.

Michael Arrington from Techcrunch was recently writing negatively about tagging, speculating that Google might block Technorati from search results, because the content they display is in effect a search result based on a particular keyword.
It should also be noted that until extremely recently, Techcrunch was a heavy user of Ultimate Tag Warrior. They have now switched to having personal database pages on a db.techcrunch.com subdomain.
Their tag pages however still exist, and are called their company index.

Rand Fishkin was asking Matt Cutts about search results being indexed within Google search results at SMX and from what I gather, the primary criteria is whether the tagging is useful for users.

Techcrunch obviously has a lot of link juice to throw around, and typically writes more than one post about each company. They do however sometimes talk about multiple companies in a single post, thus there is occasionally some duplicate content appearing on different tag pages.

Techcruch really don’t need tags for SEO purposes, though using tagging might help a little with relevance – they are effectively using tagging as their primary HTML sitemap and show full content.

Thus I would think Techcrunch’s use of tagging is a benefit to readers and highly valuable.

I should also note that there are a lot of blogs in the Technorati Top100 that use internal tagging, with the pages being indexed.
Wordpress.com uses tagging extensively, with all tags being indexed. I am not sure if this is of benefit to searchers.

I am not sure about Technorati’s tagging being of benefit. When a Technorati tag page appears in search results, I personally don’t click on it, but then I would most likely not click on Wikipedia entries either, unless I was specifically looking for a Wikipedia entry.

Youtube might have blocked off tags, but then Google have to be careful of favouritism these days with lots of hungry Bloggers, and lawyers watching their every move and acquisition.
Youtube seems to already be doing well with Universal search. How do Google detect largely duplicate videos?

Whether my own use of tagging is as highly valuable to readers is certainly questionable, but when you look at the limitations of current search engines, I think using tag pages is actually fair game, as long as you don’t abuse things too much by using tags that are not relevant to the content.

The Biggest Benefit of Tagging

Dubious amounts of traffic from Technorati themselves?
Having useful pages for people to link to, not only from their blog posts, but also Wikipedia?

No…

The number one ranking factor was TITLE TAG, and to a lesser extent additional headings

I have been able to gain some additional traffic simply by tagging with and without spaces.

I could also add to the mix various forms of interlinking, the fact that every time someone picks up my content on a splog I gain around 15 links (though low quality generally, some of these sites are PR4+), and how easy it is to spread juice around if you need to.

Tag Pages & Duplicate Content

For the last few months, Google has had what I would regard as “The Google Yoyo

Here is a chain of events

  1. Google had only a few of my pages listed using “site:” though I was still receiving 40% Google search traffic
  2. I reported myself for paid links, and a short while later, I had 4000 pages indexed, and almost no supplemental
  3. I added a translation plugin – it creates translated cached pages for the whole site, but it only linked from my single pages. (more on that for another blog post)
  4. I started gaining more pages indexed, and still no supplemental
  5. I started seeing hour by hour different results, sometimes I would have some supplemental results, and they would disappear 10 minutes later.
  6. Indexed pages and supplemental results seem to have stabilised

I currently have around 11,000 pages in the main index, and 9000 pages listed as supplemental.

My translated pages gain less juice than any of my English pages, so are more prone to be supplemental.

Tag pages which only contain a single post are generally supplemental, even if they are linked from a piece of content that gained some good linkage both internally and externally.

Translated versions of tag pages that were supplemental, are also supplemental.

So far I haven’t done any optimization on my tag pages to try to make them more “unique”

4.9 > 3.6 ? (Titles > Duplicate Content)

One of the most common things to do with PPC advertising is to use a version of the keywords with and without spaces, especially with only 2 words, or with a website name.

I have done that with tagging for a number of products, and sometimes it has resulted in a good deal more traffic, or traffic when I wouldn’t have seen any traffic at all. On some promotions I am talking an additional 1000+ visitors.

I have mentioned in the past that I rank for WordPress Training with a tag page, actually so does WordPress.com – it isn’t a competitive term, but it only cost me typing one additional tag on a couple of pages.

Double results in the SERPs are often coveted, and these frequently appear when using tag pages, so one listing will be the most specific page on your site, and the second result might be a tag page to all related content.
If I was someone searching, I have a feeling a tag page might get more clicks than if you had a single listing plus a “more results” link – I don’t have any eye-tracking data to prove that.

Tagging and Google Blogsearch

I have seen some strong results in Google Blogsearch for my content, although it is hard to prove whether that is because of tagging specifically, or other ranking factors.

4. PageRank ‘Aiming’

I have blogs that use siloing and minimal crosslinking, and in fact I can just copy what I sent someone in an email a couple of days ago.
The person asking wanted to create WordPress structures similar to the structures described so brilliantly by Michael Campbell, and then in greater detail by Leslie Rohde in Revenge of the Mininet and the Dynamic Linking ebook.

********** email about WordPress Siloing ******************

You can do it with hacking the core files, or creating modified functions in functions.php

The easiest way however is to add nofollow to links generated by various functions

http://guff.szub.net/2005/01/27/add-link-attribute/

and possibly

http://www.sochi-travel.info/articles/wp-nofollow-from-home/

You could also use tagging to channel pagerank around a little, though that ends up slightly different to Leslie’s teachings.

You might also find this useful

http://www.memwg.com/seo-siloing-wordpress-plugin/

You might also want to use one of the many plugins that stick noindex follow or noindex nofollow on all the duplicate content pages.

I actually forgot to mention custom query string in that reply.

If I was using siloing on this blog, and no translation plugin, I would have less than 500 indexed pages.
With tags and translation, currently 11,000 pages in the main index, and 9000 supplemental.

Can a bigger net catch more fish?

Now Google seem to be reporting the supplemental results correctly, I can work on moving more of those pages into the main index.

Milk Bottles & Duplicate Content

For me, one of the most important concepts is that if you are going to have duplicate content pages, they shouldn’t leak Google Juice excessively.

Google themselves say that the reason for supplemental results isn’t duplicate content, but a lack of pagerank. Rank flows from one page to another, it doesn’t dissipate.

The same is also true if you use Dofollow, you should try to maximise your internal linking as much as possible with related links, tags etc to shift a little more of the incoming juice to other pages that might need it.

I see people suggest noindex, follow as being good for tag pages and other duplicate content. If you have external links on those pages, maybe noindex, nofollow would be better – I can’t see any benefit in the follow.

More Google Yoyos

Link attribution for syndicated content is really broken this week, maybe due to recent algorithm changes. As an example I used to have a very solid 3rd and 4th place ranking for “dofollow” – fairly understandable with the amount of links I have on my list of dofollow and nofollow plugins.

For the last week or so 2 articles that link back to my originals syndicated on Webpronews.com have been ranking, and those articles didn’t receive lots of links.

 

Liked this post? Follow this blog to get more. Follow

Comments

  1. says

    Hey Andy,

    Great post.

    Say you have a blog with one hundred entries and out of those, only 2 comprise the target of over 90% of your incoming links. Might it be detrimental to the rest of your pages if you aren’t distributing the page rank?

    Precision: of course it would benefit them to get the PR but are they penalized because of their seeming weakness in comparison to your strong pages?

    • says

      Alex, SEOmoz have published case studies in the past that even unconnected pages seem to benefit from linkbait.

      That is the global link popularity of a site, but I wouldn’t for instance know how that compares between a smaller and larger site. Does a larger site exponentially need more links for the popularity of site to be more significant?

      So I don’t think there is a penalty, and the global popularity benefit would make it impossible to spot.

  2. says

    Very informative. I’ve read it twice. I think I will visit more to read and understand what you have posted. Thanks bro. Nice to meet you.

  3. says

    Excellent overview and discussion relating to Reese’s query to the SEO “experts”. I think you probably do more than most site owners with respect to testing using the iterative process: change, check, change, check, change, check. That’s the same way I code a page. Change something, check the output correct the errors if any (and there usually are!) move on.

    Now all of this SEO jockeying and positioning of knowledge and speculation really seems to me to be falling short of the one true factor that should never, ever be overlooked. And that would be quality content, that is on point to your category, tag, site, niche, etc. Not sure completely how to do testing for that sort of thing; perhaps creating a page following all of your site’s other SEO conventions but being totally and completely off topic for the site’s focus and linking to nothing related, etc. That would be an interesting little experiment, Mr. Beard. How about it ;)

    • says

      James, quality content gets links (if linkerati see it), and topical relevance both on the site, and the site giving a link is already listed.

    • says

      Supplemental results are lack of pagerank, not duplicate content necessarily.

      In my case for instance, the pages Google look on as being supplemental are the tag pages that have only been used once r twice. It could be looked on that they are supplemental because other than the title, they are almost identical to the original article.
      Alternatively, they could be looked on as being supplemental because they only receive a link from that one single page, and possibly from my main tag page which has a lot of tags on it, thus any pagerank is split.
      My most popular tags gain lots of links internally, and a few from external sources.

      Customizable post listings would allow you to create category archives that don’t run as deeply. Some of your categories run 8 pages deep.
      You could also experiment showing just snippets for them as well, but that might make them less useful to visitors. Duplicate content is a lot to do with choice of landing page, but also relevance.

      There are lots of plugin solutions for blocking off what is being indexed, either as a plugin, a snippet of php in the header, or a robots.txt file.

      You have to ensure if you do block off pages, that Google Juice can still flow to your deeper pages.

      Also important is you have to compensate with more internal links. If you block off 20 categories from your sidebar, more juice will flow out of your site through external links in posts and comments and trackbacks (if you are using Dofollow)

      As I am answering questions, I spotted a few links I intended to include but forgot – one being a link for my reference to milk bottles. External leaks on duplicate content pages have to be avoided.
      http://andybeard.eu/2007/03/multiple-reasons-why-i-loathe-top-commenters-plugins.html

  4. says

    I reported myself for paid links, and a short while later, I had 4000 pages indexed, and almost no supplemental

    Do you mean that you notified Google about your paid links and now Google is showing all (or a lot more) of your paid links?

    • says

      What I did was report myself for having written paid reviews, with confidence that the quality of the reviews I write would mean that Google couldn’t class the reviews as having been bought for SEO purposes.

      He is the writeup where I first wrote about reporting myself for paid links.

      A week or so later I wrote about how I had pulled myself out of supplemental results, or no results showing.

      In hindsight it is hard to say whether that really caused a human review to fix things, or a human to trigger some kind of reindex to fix a bug.

      I have also noticed some very expert analysis of my tag pages from someone in New Zealand in my stats, not only specifically searching to see which of my tag pages were indexed, but also which were in supplemental results.
      This was around the time (by my records within a few hours) that I started to see “hour by hour” changes, as if someone was tampering with algorithms or that datacenters were reporting different results for a few days and for some reason I only saw a fleeting shadow of the changes coming.

      I do know that Google must generally like this domain, new content usually appears in the main index within a few hours at most.

  5. says

    Andy…Nice detailed post. I plan to mention it on our Friday “Who Said That?” so that our readers can benefit from it. The SEOmoz list is nice. I remember the first one Rand put out. I like the way you put parts of it easy to read for a quick snapshot. I’ll have to have Fred look at some of the more GEEKY part of the post since we use Word Press. On another note, I’m glad that I found your bumpzee follow community.

  6. says

    Great post – I love your insight into the subject. Clearly you have been doing a lot of research on the subject. I see your translation plugin has the same problem that mine did when I tried to use it and I have been extremely unsuccessful in using it.

    Keep up the great work.

    • says

      I am in the middle of writing something about the translation plugin, but I seem to get either Altavista or Google errors on a frequent basis.

      Yesterday I switched to Google because I was showing a yahoo error page – today I end up switching back to Altavista translation.

      The biggest problem is that you can’t see that things are broken unless you check daily/hourly.

      Now I have ended up switching it off, because I am getting Yahoo errors.

      It is impossible to gague how effective something is if people end up landing on broken pages half the time.

  7. says

    Hi Andy, I don’t understand this point, can you explain it?

    “I reported myself for paid links, and a short while later, I had 4000 pages indexed, and almost no supplemental”

    Thank you.

  8. Jim Harrison says

    The information you shared was excellent. I plan on using many of the things you’ve suggested. Thanks for all you do here. It’s really appreciated.

    Jim

  9. says

    Very useful information u’ve shared here;can you put some light on duplicate content also…itz jus a request…pls do share your view on the topic with us if possible.

    • says

      Duplicate content on blogs caused by various CMS issue or features – Google seems to handle it quite well as this article and those linked described.
      Duplicate content from syndication – Google’s algorithms are currently a total mess and they are not handling link attribution correctly.

  10. Mark says

    I am sorry if this is the wrong place for this question but I got intrigued by your related post item. How did you do that. I am also using the New Blogger.

    • says

      I have various ways to hanle it automatically with WordPress.

      With Blogger, you have to do it manually, or host on your own domain and use some very custom PHP

  11. says

    Very informative article. I am an experienced webmaster and marketer but I am fairly new to blogging and advanced SEO so some of the post went over my head a bit but that just gives me more reason to learn about it. I have a couple wordpress blogs and a few other sites and I have been looking for wordpress SEO info so I will definately be rereading your info and using some of your suggestions. Thanks for the great info, keep it up.

  12. says

    Impressive work, Andy. I became acquainted with your site through PPP, and I have started reading regularly. I appreciate the trial-and-error method that you use to tinker with the SEO capabilities of your site, and I have used quite a few of your ideas to improve my site’s readership and PR.

  13. says

    Andy,
    It is true that SEM types like to speak in absolutes. The reality is that SEM is the ultimate combination of left and right brain thinking-pragmatic, scientific, creative, artistic, and statistical thinking all at once. It’s sort of like recording music you’ve written.

    I like the humility in this post.
    Marty

    • says

      3-4 weeks, and I am not even sure it is stable now, though it has been fairly stable for 2 weeks now.

      It must help that I have PR6, 7… sometimes even PR8 links coming in from topical authorities – the links do eventually get buried off the front page, but eventually Google should be receiving the right signals.

  14. says

    Did you ever drop off completely during those times or just yoyo. I only have a pr4 but it has dropped off for 2 or 3 days completely. Albeit, a better SERP resulted when i came back.

    • says

      The only thing that was disappearing were the supplemental results listed, not the number of pages in the main index.

      I am not looking at the ranking for specific search results, other than my comment that they seem to have a problem with link attribution on syndicated content.

  15. says

    Andy,

    one edifying article you have there andy,
    /name-of-the-post/ or /name-of-the-post.html there are just quite the same. As long as you dont use “?”,such characters. Permalinks helps the search engine crawl to your site faster, so if you have categorize your permalinks, there a good change the search engine would spiders your website.

  16. says

    Good article Andy…

    I would still come back to the question of tags which is an interesting one.

    1. If you include tags and make dup content, what is the purpose.

    2. Based on the above you could set up tag pages to be somewhat different from your normal pages.

    3. Which pages would you prefer to be ranked? normal or tagged?

    4. If you close off tag pages to collect PR, you will be wasting PR on dup content on which could be used on other new, unique pages… especially if your blog is dynamically linked (Leslie Rhodes)

    Someone asked about minimizing PR, there are several ways…

    1. You can put in place a java navigation menu, here is the one I’m using for wordpress – http://www.silpstream.com/blog/wp-dtree/

    2. Call the sidebar as a page rather than from the wp script. I could find the article on that one but definitely sounds doable.

    3. Make sure all your outgoing plugins, feed url’s have the nofollow tag. Some do require editing of the script.

    4. Disallow in robots text all folders not wanting followed including wp-, category, archives, etc.

    5. Loop your PR. Set up a sitemap on your index page only which goes to your invidual posts and back to your index page.

    6. Minimize all outgoing links with nofollow unless paid links. Be carefull with creating an exact closed loop, the SE’s arent big on this so always have some outgoing links.

    • says

      Hi Chris

      I get some great quality comments on this blog from lots of people, and I include yours among them, and thus I use the dofollow plugin to share a little linklove.

      Most blogs also typically have sitewide blogroll links, top commentator plugins, text-link-ads etc, and it all adds up to lots of external links on every single page.

      I would class myself as a devotee of the teachings of Michael Campbell and Leslie Rhode, but at the same time an active community blog is a whole different scenario to a mininet of carefully engineered niche sites.

      I initially addressed this situation by coming to the conclusion that my best strategy would be 2fold

      1. Minimise the amount of external leaks on duplicate content pages
      2. Massive ball linking, ensuring I have far more internal links on every single page, than external links.

      I wouldn’t class this as hording pagerank, but cultivating it. PR would still leak from your site through whatever external links you have, but the pages wouldn’t be penalized by multiple iterations of the pagerank calculation.

      As a followup to this post, I went into this in more depth, and I think you will find my sandcastles approach might be ideal for a WordPress blog that is community friendly, but doesn’t wish to sacrifice itself.

      http://andybeard.eu/2007/06/wordpress-seo-masterclass-for-competitive-niches.html

    • says

      I am still experimenting with this one
      http://blog.taragana.com/index.php/archive/angsumans-translator-plugin-pro-for-wordpress-blogs-released/

      I purchased the gold version, not sure if that is still available for $100, as it was meant to be limited to only 100 users

      Now I have moved servers I intend to start playing with it again. There is a negative aspect to it that it bulks out the database files for the cache, and moving WordPress and backups is enough of a pain without 100MB+ databases.

      Also for some reason whilst the free plugin I was using was being indexed extensively by Google, when I switched, Google decided that it didn’t want to index the new ones (different URL structure) – that might be a problem in my own configuration.

      I also didn’t manage to get a lot of the additional languages working correctly, but my new server has more IPs to play around with to proxy the requests which might have been the problem.

      The translation is good enough to get search traffic, I wouldn’t class it as good enough to read easily, the same as if you were translating a foreign document with babelfish – it can be enough if you are used to that kind of translation.

  17. jrliem says

    Hi, nice post.

    Btw, have you heard about DupeMagic? It's a WordPress plugin for creating high quality unique content (not garbage). You can check the result using copyscape. If your site is not listed in the result, your page is unique.

    You can read the review here :
    http://internetmarketingtool.co.cc/seo-tools/du

    thanks,
    Samuel

Trackbacks