Sunday, August 19, 2007

Beautiful Visual Link Exploration Tool

Neato. I am not sure where I can across this but it is really cool to play with.

Just type in a URL and it will give you a visual map of 6 related links. Incoming links are green, outgoing links are blue. Just click on any of those links and it will map the 6 inbound/outbound links for that page. Any links on that page that link back to previous pages will be mapped too.

What a cool way to visualize how neighborhoods are linked together. You could spend hours on this.

You can change the number and ratio of links if you want. I think I'll continue to play with this for a white to see if any real useful process can come out of it. Maybe not, but it is fun anyways.

Saturday, August 18, 2007

Top Ranks for $10/Click Keywords Going to Doorway Pages

I just did a random search for a top three word phrase for a client to see where they stand organically - and keep an eye on the competition. My client currently bids around $8.00 to be #3 for this phrase in the sponsored results, and is ranked #7 organically. The top bid for top position in Adwords is well over $10.

The top organic result for this phrase is extremely keyword rich to say the least. The three word phrase is repeated twice as the first six words of the title, once at the beginning of the snippet, and once in the page url.

The landing page is rather ridiculously over-optimized for the phrase; it begins the headline, begins the first paragraph, is repeated again in the first paragraph, and is the start of the next five paragraphs.

No legitimate business would ever want to do this because it kills your credibility, would destroy your brand, and cannot possibly convert very well anyways. You've to hand it to them though - they do rank #1 - so if that is the extent of their short-sighted goal then they have certainly achieved it.

Still, I can't believe that it is as simple as over optimizing a single page to end up number one for a phrase that costs $10 - $15 a click for the equivalent in Adwords. This position is literally worth thousands of dollars per day.

So at the very least, this deserves a deeper look:

At the bottom of the page there are links to dozens of similar sounding pages, under the heading "other articles". These pages are comprised of several groups of extremely similar variations of related high value phrases. Only one of the three words in the phrase are different in most of these phrases.

Each of these pages leads you to a unique page of content that is super optimized for that phrase. They've created a matrix of very similar, but unique, super optimized pages. They've created one page for each variation of their targeted phrases, and each of the pages are linked from every other page using the appropriate link anchor text.

It's classic, beautiful search optimized content and architecture, but one problem I see is that all of these links are buried at the bottom of the page, and therefore shouldn't be considered to point to pages as important as the links in the main navigation on the top of the page. So why is all this working so well?

A quick check of the source code reveals that those links on the bottom are very near the beginning of the HTML - Only a tiny amount of code and and text content of the page is before them. These links are also placed in a way that they appear to the search engine to be embedded in the copy, which will make them appear to be very important links.

The most unusual thing about about all of this is how close to the beginning that main text area is. It starts right at the beginning of the body! Where is the header and the main navigation area? The main navigation has dozens of links and sub links that should come before this. These links aren't even on the page anywhere! A few searches of the source code for URL's and visible phrases in the main navigation turn up totally blank.

Hmmm. Where are these visible links, where is the header?

Aha, the entire header, navigation and links are all hidden in 3 javascript files included in the head tag. Cunningly, the entire main web site main structure (this section is only the tip of the iceberg) is hidden from Google on this page, and all of the super optimized content and links have become the new site to Google. The visible navigation, which takes visitors to the main site, is quite prominent on the visible page. Their strategy, it appears, is to get visitors to this page and hope that they'll click any main navigation item to get into the "real" site - never to find their way back to this over optimized page again.

A lot of work has gone into this, but I couldn't see how anyone could allow their whole web site to look like this. It's a step above the old auto-generated doorway pages from the nineties, because each of these pages had to be written by a human, but I still couldn't see them as a serious business so far.

So then I go to the home page, and I see that the copy is written quite well, and they seem to have a lot going on. This company has actually been in business since the 80's selling packaged software. Clicking around I noticed that all of the very similar links found on the original Google optimized page were nowhere to be found. I looked at the site map, thinking that surely they would use this page to link to their optimized page section, but I can't even find the links there.

Going back to the optimized page reveals a clue, the breadcrumbs show that I am in a section titled "articles". Looking at the footer on the home page reveals a link to "articles" that is not even in the site map.

Except for the over optimized content, it's quite cleverly done. If they were a little less aggressive with the keyword stuffing in the content they could easily argue that these pages were all legitimate articles about key phrases that their customers were searching for. They would be "legal search landing pages" instead of the much feared "illegal doorway pages". Yes, it is a fine line, and one that they still walk very carefully even if they are stumbling over the edge.

Each of their optimized pages truly are a unique article written by a real human being. It may not be auto generated, but any sane human being would agree that they must be written for something other than intelligent humans. The search engines wouldn't know this, of course, because the content won't be picked up by a filter that is searching for extremely similar pages, for too many similar phrases, or even a filter to detect randomly generated content - so they get away with it.

I wouldn't recommend some of these tactics, because they could easily land you with an over optimization penalty and get you kicked out out of the the index all together for a while. I do , however, think it is a very good illustration how aggressive you have to be to rank well these days.

Saturday, August 04, 2007

NY Times Explains the Google Search Engine

If you are not in the Internet marketing industry, and want to get a feel for what exactly is all the fuss about search engine marketing, there is a good article on Google in New York Times last month that I just found, ironically, while searching for references to the Google search quality team: New York Times on Google