I’ve been reading a lot lately about how Google is changing and the semantic web is coming and that links with anchor text are dying and that pretty soon you won’t need to get links and that you’ll just need to create content with a lot of words related to your topic in it and customers will come flocking.

I disagree.

Links are going nowhere, mainly because links are a fundamental building block of the web and quite frankly, a site that discusses a person, place, or thing without linking to it creates a poor user experience. So even if there is no ‘perceived’ SEO benefit, links will always exist on the web for usability, users will click them, search engines will find them, and therefore links will have SEO benefit.

Also, this whole concept of Google being able to judge the relevancy of a document simply based on the ‘co-occurence’ of other terms is a fantasy in my opinion. If that’s the case, then cue up the bulk LSI article creators on Fiverr! If Google doesn’t look at links, and just looks at co-occurence then there will be millions of LSI-optimized (lsioptimization.com is available …) horrible articles about everything under the sun and Google will have no way to tell which ones to rank.

I’m not ragging on co-occurence but writing using related terms isn’t new, and no matter how large a signal it is there will always be an even larger ‘popularity’ score involved based on how users can ‘vote’ for sites.

Before Facebook and Twitter the only way users could ‘vote’ on the web was by linking to an article from their website. Therefore, links were a proxy for popularity, and quantity/quality of links was a proxy for how popular something was.

Today Google has many other ways to tell how users ‘vote’:

1) What they click on in the search results
2) How long they stay on a site and what they do
3) How many Facebook likes, retweets, Google +’s, etc. a page has
4) How many links a site has
5) How many people search for a subject/person/brand/entity
6 – Infinity) Other stuff

They also have tons of other ways to judge links:

1) How long has it been on the web
2) What is the anchor text
3) How often is it clicked
4) Where is it on the page
5) You get the idea …

But keep in mind that users and search engines still arrive at pages via …. LINKS! Whether they click on it from Facebook, or Twitter, or from an email, people click links, and the quantity and quality of those links out there will always be a vital factor in how your page ranks in search engines. Sure over time certain attributes of the link will change in value (the authority of the page/site the link is on, the anchor text, the usage metrics of the URL the link is pointing to, a lot of stuff mentioned here) but links are here to stay.

For those that hate reading and want the 8 second summary I’m basically saying that as long as people use links to surf the web they’ll always be a very valuable signal for search engines to determine how to rank sites.

It should take you no more than 90 seconds to read the last four paragraphs of this interview with Debra Mastaler from Alliance Link. If you want to learn how to build links read this, then read it again. (Copied below for your reading pleasure)

You have a new product and want to use it to build links. Before going public, you do a soft launch to your customers with an invitation to link and incentive (sales.) You then run a national contest (promotions) and announce it via press release (publicity). You also send a release announcing the winner.

While the first part s is going on, you instruct your copywriters to create a humorous piece of link bait which is launched on Digg and a handful of other social media sites (promotions). The linkbait has a tie-in to the contest. (promotions) A blogger outreach program begins and offers free samples of the new product and an invite to review or send rate cards with advertising opportunities (promotions). Once the reviews come in, bundle them, and send release showcasing the successful launch. (publicity).

Based on the success of the launch, have sales and copywriting staff write a case study/white paper referencing the process, reviews and customer feedback. Offer white paper to key journalists before going public, (publicity) and then add the paper to any content source that will take it (promotions).

And so on. By the time this link cycle is complete, you will have touched on almost every facet of marketing without having to purchase a single link or having any one of the tactics stand out. Balance is key.

It’s disheartening to see high Page Rank blatant paid links helping sites rank in the SERPS. It seems right after you work so hard to build links ‘the right way’ that Google approves, you see this crap leap past you in the SERPS and give you the flaming finger as it settles down a few spots above you.

In an industry I’m competing in, a site has taken it’s place in the top 10 using primarily purchased links in sections labeled ‘sponsored links’ or ‘paid links’ that are off topic and clearly spammy. The links don’t have no follow and the target site’s neighborhood includes such esteemed industries like Russian Girls, Australian Online Pharmacies, and Discount Prescriptions.

It makes you wonder how far away Google is from being able to detect spam and when it will actually not be worth it for webmasters to buy links from off topic high pr sites. Because from what I can see, they’re not even close. As Google works a lot with probabilities and patterns, I really don’t think that they can decipher spammy links on highly trusted domains yet unless they’re manually reported. It seems since Google update Dewey that links that would not have counted in the past are being given credit in the document scoring process.

I could report them to Google and watch them be stricken from the index, but that does me no good long term and I’m sure the corresponding karma (I’m a big believer …) would come back to bite me in the back.

Has anybody else seen similar results, or am I totally missing the boat here?

November 6, 2007

Stubhub announces a deal with ESPN and gets highly targeted links from team specific pages throughout the ESPN.com site.

“The agreement integrates StubHub branding, links and promotions across ESPN.com, and creates a convenient destination for fans to search for tickets to sporting events nationwide. StubHub is now dynamically integrated into ESPN.com Team Pages, Scoreboards, Game Previews, Clubhouses, Schedule pages and Travel Indexes. A custom StubHub “TicketCenter” also appears on every major sport index page where users can buy tickets for any team.”

This is a hyper relevant link on one of the most authoritative sports sites around with targeted anchor text. Shockingly StubHub doesn’t dominate this market. Razorgator, Tickets Now, and a few other players put up formidable competition and fare quite well in many of the SERPs I have looked at.

As a link builder, one of the many things you can use to assess the level of competition in an industry is to check the backlinks of a few of the top competitors. Unfortunately we don’t get too much help from the search engines here, as they all report wildly different numbers and it’s impossible to tell which links count in the eyes of Google. I believe that the 80/20% rule applies to link building, where 80% of the link pop for a page or site is obtained from 20% of the links. I actually think it’s closer to 90/10, but as with most things in SEO, it’s unprovable and not information you’ll find in a backlink analysis.

Using Search Engine Land as an example, you can see the inconsistency between reporting:

Yahoo Site Explorer: 573,000
linkdomain command: 325,000
Google link command: 6,800
MSN link command: 0 (they blocked it again)
Alexa: 143

Two things to note here:

1) It’s common knowledge that Google’s backlinks are only a sliver of what is really being counted.
2) Alexa reports unique domains that they have data for linking to a site, not total links

What to do with the data

Regardless of the numbers reported by the engines, it’s not worth it to go over these with a fine tooth comb. I would recommend using Joost De Valk’s SEO link analysis firefox extension and trying to get a good idea of the different types of links that your competitors are getting. The extension will list the PageRank, anchor text, and whether or not it’s nofollowed. What I generally look for are types of links the competitors are getting. In order to do this I think the “linkdomain:site.com -site.com” works best as it provides you the most flexibility to filter the data.

For example, using retail giant Walmart as an example, to find how many directories a site may be listed in, I use a search in Yahoo like linkdomain:walmart.com intitle:directory. This shows all the pages that have the word ‘directory’ in the title that link to walmart.com. There is margin for error as some directories don’t have the word ‘directory’ in the title and some of the pages that do have it may not be directories, but I’ve found it to be reliable enough to give a quality snapshot.

You can also do the same thing to see how many links they have from blogs by doing a query like “linkdomain:walmart.com blog.” This shows you how many links they have from pages with the word “blog” in it. You could filter it a step further and search for only .edu blogs using a the query like “linkdomain:walmart.com blog site:.edu.”Again, the same margin for error exists for the reasons mentioned in the above paragraph, but for sites that don’t have blogs in Technorati (authority score in Technorati = links in the past 6 months) this is another reliable snapshot.

You can slice it and dice it any way you want which is why I really like utilizing Yahoo explorer more than any other tool. Depending on what you’re looking for, there are an infinite number of ways you can use this to help you in your quest to build links.

A common question these days is the issue of the effectiveness of reciprocal links. People have often made compelling arguments for both sides. Here’s my take.

Recips Work

I think reciprocal links never really stopped working, it’s just that Google got better at detecting relevancy. Before it seemed like Google used to count all recips regardless of whether or not the two sites were relevant, but now far fewer are passing juice. Recips from pages that have a high probability of being unrelated, and who reciprocate with plenty of other unrelated sites more than likely had their links discounted. But there were still some recips that counted between pages that were on the same topic. I think Google is too advanced to discount or give credit for all recips, and still give the same amount of credit for links between two sites in similar categories, regardless of whether or not they’re reciprocated.

This is often why you look at some competitors sites and conclude that all they really have going for them is links on other sites links pages, and they’re still ranking real well in the search engines. Although lots of those links may not count since they’re from old school unrelated sites like mentioned above, there are still likely more than a few that are from sites that only linked out to sites within the niche and who are passing the majority of the overall link juice.

How to Find the Juicy Sites

In order to make it easier to figure out which of the reciprocal links might be the few that are passing juice, you can run some queries in Yahoo utilizing the linkdomain command. The first query I would do is something like: linkdomain:competitorssite.com intitle:”key phrase”. If a site has the keyword phrase you’re targeting in the page title, it is a decent relevancy signal and the page is worth examining. If there are links only to related sites on that page, then it is a link worth getting, and you should have no qualms about linking out to that page from your own site.