Currently, Stompernet has two new videos by Leslie Rohde that have stirred debate about what LSI, siloing and theming are, and their role in ranking in Google. First of all, don’t waist your time or money with Stompernet. They charge too much, and DO NOT listen to the first video and abandon LSI, siloing and theming. Let’s get on with it.
Stompernet Claims that Google does not use LSI in its ranking algorithm and they can prove it. In truth, this is a false claim, and I’ll show you why. Their whole video series is just a strawman, Strawman is a rhetorical technique (also classified as a logical fallacy) based on misrepresentation of an opponent’s positions. Their whole LSI debunking argument is based on a misrepresentation of a rival’s (Charles Heflin) position.
Stompernet is unethical. They create the illusion of having refuted a proposition by substituting a superficially similar proposition (i.e. their mathematical definition of LSI), and their proof is simply a search for a word in singular and plural. Leslie suggests that if Google used LSI, the results would be identical in terms of search results and number of results returned. The fact is, they are not. I would argue that this would only be true if LSI were the only ranking factor – something I don’t think anyone believes. In this case by claiming to refute the original proposition without ever having actually refuted it. In doing so, they misrepresent the idea of LSI, and simplify it to a mathematical equation.
Secondly, they claim to have debunked a rival’s theory, when in fact, he has never put forth such an idea. When they claim that Charles Heflin has stated that LSI is what google uses to rank websites, this is simply a false claim.
Finally, their proof just does not make sense. They simply do a search for a word in the singular and plural, and suggest that if Google used LSI, the number of results returned would be identical. When the number of results returned are not identical, they claim to have made their point.
Leslie simplifies LSI to just an algorithm. When webmasters refer to LSI, they are most often referring to theming or siloing. Often, a word takes on more than one meaning, i.e. Band-Aid is a brand, but it also refers to a bandage. In the first video, Leslie Rohde states that Google does not use LSI in its ranking algorithm, and says he can prove it. The proof is simply a search for a word in the singular and plural. Leslie suggests that if Google used LSI, the results would be identical in terms of search results and number of results returned, and they are not. I would argue that this is only true if LSI was the only ranking factor – something I don’t think anyone believes. For me, this creates a major problem with the first video, which means that whatever else Leslie says is simply not relevant.
This post is about defending Theming or Siloing, LSI (or better phrased Relational Semantics), Co-Occurrence and Expert Verbiage to outrank websites. You see, IMO, Leslie seems to be trying to convince people that something they believe to be true, isn’t, when in fact its not really something they believed in the first place. You see, when Leslie uses the term LSI, he does not use the term in the same sense that most people do. To Leslie, LSI is the mathematical equation laid out in a research paper.
Of course, to give a little background, this LSI (and talk of Siloing) was created by Bruce Clay way back when, which he called theming. Here’s a quote…
“Silos and theming have been around for a long time, and back in 2007 you would have heard the newest buzz words on popular webmaster forums: such as “silos”, “themed websites” and “expert verbiage”. But IN REALITY professional SEMs have been using these very same strategies for years on end for consistent, high SE rankings for their clients. The proper application of silos and themed websites had been a closely guarded secret of these professional SEO firms… up until the advent of Latent Semantic Indexing.”
However, Stompernet doesn’t seem to want to educate its viewers about this. For those who would like to be in the know, this is what happened in 2006: (when the tables really started turning in Google’s favor). Seo2020.com explains what was going on at Google:
On December 28th 2006 Google filed a new patent application titled “Detecting spam documents in a phrase based information retrieval system”. Google engineer, Anna Lynn Patterson is the inventor or this patent. Here is a quote directly from the patent:
“An information retrieval system uses phrases to index, retrieve, organize and describe documents. Phrases are identified that predict the presence of other phrases in documents. Documents are the indexed according to their included phrases. A spam document is identified based on the number of related phrases included in a document.”
So Google has filed a patent for an algorithm that will index and rate the “relevance” of web pages to determine if there is also an occurrence of phrases related to the subject matter of a web page.
To give y’all some more background, I came along in 2006 and started doing SEO based on Relational Semantics, Co-Occurrence, and Siloing, guided by The Master Plan, written by Charles Heflin. I am still a member of Charles Heflin’s SEO2020.com membership site. The principles he teaches, Siloiong, LSI and Social Web, have been taught by the greats such as Bruce Clay, and these techniques have proliferated. Although I am no longer a member of Themezoom, their keyword research tool not only utilizes LSI but also the theory of Co-Occurence. It also has the ability to organize a site that, done manually, would take a long time. As Dan Thies puts is, “What I do know is that LSI techniques can be applied very effectively for Keyword Discovery, and as far as I know that’s the main thrust of what Russell has done with LSI for Themezoom.
So while both Google and ThemeZoom stake their work on LSI, Stompernet’s video is taking an extreme position (and creating a lot of controversy in the process); simply stated, the intent of the video is not to create clarity in SEO but to attract attention and sell their products.
But that alone will not negate their well made ‘social proof’ video.
What they are failing to tell the viewer is this: while LSI is only one aspect of ranking (there are other factors), in no way has it been debunked.
I’m a member of SEO2020 and bought it upon first release. That’s when I first learned of LSI and Siloing. Google came out with their patent on LSI in 2006. Here’s how you can save a whole bunch of time: read what Charles Heflin has to say when it comes how to ranking in Google (hint: you have to add links to make it grow). Anyway here’s his quote.
Couple “social buzz” with relevant and semantically relevant words in your documents and the categorical structure of your site and BAM . … you’re at the top… simple, repeatable, no mystery … and it certainly is NOT LSI but like Leslie said …much more intelligent than using a theoretical LSI engine.” Charles Heflin
So I guess keyword density is the answer for your content, right? I’m sorry, excuse the sarcasm, but if you knew anything about LSI, you’d understand LSA, Latent Semantic Analysis then you’d understand the concept of theming–which you actually talk about in video 2. And I’m not gonna front: Stompernet has made some great videos before. It’s tough times right now, so they need money. I understand. So I don’t blame them for trying to create a marketing frenzy for their next SEO product, and every other video of theirs. Their tests of Ecommerce sites are fantastic examples of SEO applied theory, and it made sense. That’s what SEO is by the way…applied theory.
But this video is about ranking in Google, and they state that you DO NOT NEED strategies involving LSI, because Google does not base their rankings on LSI. And Leslie, the author of the video, further states YOU DO NOT NEED TO HAVE PROPER SITE STRUCTURE. Hmm, so this premise is that Siloing or Organzing your site is not needed at all to rank. It is true that there’s more than one way to skin a cat, but I prefer the most effective way of ranking in Google, the way that Google values with its own patents. And that is LSI.
This is a tricky value proposition and confuses the shit out of people. You can get a site to rank if you have very strong links, but what the video fails to tell you is that Charles Heflin and company never state how LSI content writing and keyword research is now combined with social web and expert verbiage.
Their simplification of the matter is intriguing but more likely insulting to people like Bruce Clay, Russel Wright, Charles Heflin and myself, reducing our applied theories to meaningless case studies. They are attempting to mislead for their own gain, and they are actually deleting Charles Heflin’s replies to their blog–that’s right, DELETING COMMENTS. That’s like beyond lame. I also work in Online Reputation Management and this is about as spammy as my worst clients–old CEO’s who don’t get that the internet is about dialogue not monologue. You would think an Internet Marketing and SEO company like Stompernet would embrace this. Absolutely unbelievable.
So how do you rank in Google, Stompernet? Could it be related to themeing as you refer to in the second video?
Google doesn’t rely solely on LSI to rank, but it’s crucial; they rely on many factors to determine rankings, and now that the social web has arrived, they also rely on that as well. LSI is simply related keywords: an idea that is contrary to the once dominant idea of keyword density–say a word enough times on your page and the search engines will rank your page higher.
Of course, search engines sucked and then Google came along and developed a new information retrieval system that uses phrases to index, retrieve, organize and describe documents. That is why they have a patent for LSI or Phrase identification in an information retrieval system.
I’d be happy to discuss information retrieval any day of the week with Stompernet members.
The patent states:
“The system is further adapted to identify phrases that are related to each other, based on a phrase’s ability to predict the presence of other phrases in a document. More specifically, a prediction measure is used that relates the actual co-occurrence rate of two phrases to an expected co-occurrence rate of the two phrases. Information gain, as the ratio of actual co-occurrence rate to expected co-occurrence rate, is one such prediction measure. Two phrases are related where the prediction measure exceeds a predetermined threshold. In that case, the second phrase has significant information gain with respect to the first phrase. Semantically, related phrases will be those that are commonly used to discuss or describe a given topic or concept, such as “President of the United States” and “White House.” For a given phrase, the related phrases can be ordered according to their relevance or significance based on their respective prediction measures.”
Fact is, there are tools out there that are getting just as good as Themezoom, namely Nichebot Wordstream KewordSpy, SEMRUSH and Market Samurai, but Theme Zoom does integrate so many of these functions under one system it can be beneficial for some. One thing I will say is they need to work on their layout. That patent is the BASIS for Google relevancy, starting in 2007, as well as their patent for Co Occurence.
Yes, they have a patent on Co Occurence, which basically means they are looking for clusters of semantically related words that show a mastery of a topic, so they can safely send a searcher to your website and be sure she will be satisfied.
So this why some documents can outrank authority news sites like the www.nytimes.com or www.webmd.com. I’m going to give you three great examples. And as an aside, though Dan Thies doesn’t believe in Silos and Themes, he certainly believes that by using LSI and Co Occurence to conduct research, he can outrank his competitors (without these techniques it would probably take a backlink from ungodly sources…)
Case Study 1: http://www.heart-valve-surgery.com
My first example involves the site about heart valve surgery. At 33, a guy had to have this surgery and wrote a book to help people go through this life changing experience. This topic of heart surgery is complicated, and so you’ll notice only .edu and very authoritative sites ranking for such keywords about heart surgery.
Yet his site www.heart-valve-surgery.com ranks Number 1 in Google for over 60 keywords. He ranks in the top 20 of Google for over 1,400 keywords!http://www.semrush.com/info/heart-valve-surgery.com
This is a perfect example of expert verbiage and LSI (or Google determined synonyms) in effect. Google is ranking his site ABOVE www.webmd.com, www.merck.com and www.clevelandclinic.org, the most respected heart institute in the world for 100’s of keywords, which I’ll show right below.
The website started a couple years ago, www.heart-valve-surgery.com, contains the expert verbiage and related words that Google is looking for, a topic that is difficult to outsource to writers because its a complex subject matter. Which is why it’s a great example for my case study. And yes, he gets many comments and feedback which adds to the social aspect.
A Google search for “valve replacement aortic” which only contains one keyword in his root domain, outranks authority sites named above including .edu sites and others with less page rank.
The expert verbiage and relational semantics Google is looking for is used in http://www.heart-valve-surgery.com/heart-surgery-blog/2008/02/14/ventilator-tube-intensive-care-unit/ which results in these google results
For further showing of this site, go to http://www.quantcast.com/www.heart-valve-surgery.com
I can’t give further rankings of his keywords, but if you are a professional SEO, then you’ll be able to utilize the proper keyword tools to reverse engineer his rankings, proving a simple applied theory: That Google awards sites that employ expert words, semantically related, and can scan the text and understand using Co Occurence algorithm.
Thus, this site outranks the most trusted, authoritative websites on the internet. So now you understand that what Stompernet is stating is not true, and of course keyword density is not important. It’s the various related words, the theme density and semantically related images and links on a website that matter.
You could argue that the heart valve site does use a proper url structure, but that’s not enough of a retort to explain how it outranks the sites it does.
Case Study 2
When I have a website that needs top ranking for clients, I employed Siloing and got incredible results. Most variables stayed constant and it was quite obvious what happened.
A website about hemorrhoids, called http://www.crhcenter.com, wanted to rank on Google Page 1 for “hemorrhoids.” Of course that’s not the best keyword to rank for since the intent of the searcher is informational, but you try convincing a bunch of doctors and professionals who want to be THE SOURCE, you do the long tail and give them what they pay for.
I employed the tactics of Siloing, using a virtual directory style of no-following, sprinkled with one wikipedia link (only variable, it was to a resource page), and their rankings catapulted to the first page.
Another example: acne
Google ranks websites that use images, videos, and words that only experts and well-versed writers would offer: an authoritative website about acne wouldn’t rank very highly by just repeating that word, or by organizing the site in a way that wasn’t relevant for a user.
The acne site would use related words such as eczema, rosacea, daily face wash, neutrogena, etc. It’s what I call ‘expert verbiage.’ My father was a dermatologist so I bring up this as an example.
In other words, if you intend to rank for the phrase acne then you’d better discuss things related to acne, which you can easily find using the keyword tool KWBROWSE
More Proof: Google Books, and uh like everyone’s info
Think about all those books Google has indexed in Google Books? And all of those scholarly journals as well? Well those words are being filtered into the algorithm. Google rewards your page as relevant because you know what you’re talking about and it wants to give the most relevant results to users.
Take a look at my resources below: here is a dissertation on LSI from last year. I was used as an expert witness in one and there’s also another lighter one here that is in my opinion the ultimate resource for LSI in academia.
The reason Google can do this (search this way) is because people are actually contributing to the web. And to give another definition of Co-Occurence, here’s what seo2020.com states that it is
“the percentage of websites that contain both the main theme keyword (or keyphrase) and a secondary keyword (synonym) as well” and this ties in with keyword relevancy and your site having properly organized themes or silos. And themed links.
Anyway, back to some examples.
Google’s Recent Algorithm Change
Now to add to the mix of LSI, and Relevancy, and Siloing, and ranking in Google, which is what this is about–btw I named my SEO company Highly Relevant just because I believe so strongly in Relevancy–Google has just announced an improvement to their ability to index and retrieve content based on long tail searches, ie search phrases that contain several characters.
That’s right. Google has gotten so good with their relevancy engine that they recently updated their search algorithm to give better, long tail results See: http://googleblog.blogspot.com/2009/03/two-new-improvements-to-google-results.html
MY EXPERT WITNESS EXPERIENCE – WHO IS CREDIBLE?
I have been in the industry a few years, and SEO is about applied theory, and I have the rankings and credentials to back it up. Putting theories into practice…. So what that means is you can’t claim something to work in SEO unless you have proof, because there are no rulebooks. The leading publication in SEO is SEOmoz.org and they survey the leading SEO’s and publish the most agreed upon factors for ranking in Google every year. Take a look here: http://www.seomoz.org/blog/ranking-factors-version-2-released
I know for a fact that it’s difficult to get credible sources–I was an expert witness in an internet case and had to prove the value of a website that ranks for two, three and four word phrases, whereas the defense countered that the sites were unimportant and unprofitable since they didn’t rank for general one and two word terms.
CONTROVERSY EXISTS IN SEO
Time Magazine recently referred to it as magic in order to explain Huffington Post’s incredible web traffic.
So Stompernet is taking advantage of SEO controversy, and I know that they are doing so to sell a product. But it’s one thing to be the Shamwow guy showing you how easy it is to clean up your spilled wine in your rented hotel room in Florida… it’s an infomercial so you take it with a grain of salt.
But to me, so is Stompernet’s video: it contains characteristics of an informerial–attention, controversy, interest and desire. They have done a damn good job in the video: it is clearly explained and convincing. But let me be clear, they are using a logical fallacy by basing their argument on an a false assertion– that LSI is the main way to rank in Google.
And here’s the problem. All you can do with SEO is demonstrate applied theories since the algorithms are not public. AT ALL. So in this field that we find ourselves in, I encounter differing theories on a daily basis, and I am not one to argue that one works over another, although certain techniques are more effective than others, and we learn this as time goes on through testing, and whenever Matt Cutts (Google Engineer) divulges information.
In videos prior to this, Stompernet wasn’t attacking any well known, applied theories as it is doing here. In this case, they are taking an extremist, absolutist position that allows for little debate. It’s a problem that continues to exist. I learned when living in Thailand for years that in Buddhism, and Hinduism, and even Aristotle, that two extremes serve no purpose in finding an answer.
Russel Targ says it best and it applies to SEO: “The exclusion of a middle ground between the poles of logic is the source of confusions.” Exactly Russel.
Stompernet’s extreme position does nothing but ABATE SEO. Is it a good thing or bad thing? Well that’s not necessarily the question to ask, there are both positives and negatives of this Stompernet LSI position.
The real question is what is their intent with this video? So moving along, it’s probably becoming clear that this video is extremist. But why? To help clarify or demystify? To trigger interest for their next product? Or allow for a healthy debate on their website? I’d give them the benefit of the doubt, but since they enjoy deleting comments of people who support the theories they ‘attack’ in their video, I unfortunately come to a conclusion of ill intent.
Intentions are important in my own ethics. It’s a fine line in marketing, since incredible promises are made all the time without evidence but of course with disclaimers.
So you have to ask the INTENT behind this video…is their intent to solve SEO myths once and for all or create confusion, attention and interest in their product? IT IS THE LATTER. So even though Andy Jenkins can quote Lesli Rohde, a SEO Engineer, and by the way Stompernet Faculty Member, and claim Leslie is the first guy to reverse engineer link reputation, it doesn’t necessarily support the argument of their video–you might as well have inserted a Paris Hilton quote saying she is patenting the words “That’s Hot” because she was the first person to use the words in every sentence she spoke.
So, in point of fact, I’m going to include resources that I’ve used to write this entry, unlike Stompernet.
Siloing and Internal Linking
Another resource you may find useful is one I wrote on Internal Linking and Siloing: The password is rex and is located at
Academic Papers on LSI and LSA
30 Resources on Siloing
I have collected 30 resources on SILOING on my highlyrelevant blog, which covers the topics of LSI, Co Occurence and everything dealing with Siloing.
Usually I’d end my essays with resources, but I’m not done yet. I’d still like to Pwn Stompernet. And include Mr. Heflin’s comment Stompernet deleted.
Not that I already haven’t. I’ve spent my Saturday reviewing LSI and social buzz, and I could go on and on, but let me make one last point.
Do a search in Google for mobile phones or swimming pools and take the top three results, and put them into Google Adwords keywords tool. They rank sites higher that are well organized by relevant themes. You will start seeing common themes as Google groups the sites keywords.
You will notice that Google reveals its understandings of this site in proper keyword groupings. This is also called Theming, based on Co-Occurrence and is one of Google’s central way of understanding text. Google’s loves an organized site. That’s what Siloing is about. Easy navigation for a user, or a search spider, with well organized content, containing expert verbiage, resources to other websites…. That’s what Google wants.
Now, that doesn’t mean a site can’t rank without proper siloing and LSI/synonyms. It can rank because powerful links can get a shitty site ranking, but not for long, and it’s more effective for ranking to stat with the end in mind.
Stompernet, you need to rethink what you’re saying here, because I am calling your bluff. And so is Charles. Your marketing intentions. Your sales hype. Because you are attacking my field of SEO in the wrong way. Why are there are no resource links provided about LSI and Co Occurence? And quite frankly Leslie, I’ve never met you but guess what, if you think that people are cashing in on LSI, you are actually using marketing scare tactics to cash in on your product. This is what Charles Heflin has to say about the whole matter and I think it finishes the post nicely.
Who ever said that Google is using an LSI algorithm? It is computationally impossible to reach a pure form of LSI. Google initially looked at this as a way to artificially determine “human relevance” so the best (most relevant) results will come up from a human query. Now they use social proof (social buzz if you will) to determine the relevance of your documents coupled with the use of keywords within the document. If you ever want to see what keywords Google finds related to a word or phrase simply use the “find synonyms” feature on their free AdWords keyword tool… Sprinkle those keywords in your document in a way that makes HUMAN sense. Couple “social buzz” with relevant and semantically relevant words in your documents and the categorical structure of your site and BAM … you’re at the top… simple, repeatable, no mystery … and it certainly is NOT LSI but like Leslie said …much more intelligent than using a theoretical LSI engine. The cool thing about coupling social buzz with semantically related words and synonyms on a document is the fact that you can rank with FAR, FAR FEWER inbound links regardless of pagerank or site age… Call me a liar… I have case study, after case study, after case study. I believe this video to be an ATTACK on the integrity of the teachings of Russell Wright, Bruce Clay (and others) and I can vouch for the fact that neither has EVER stated that Google uses an LSI algorithm. They simply teach their clients how to use semantics to determine market value and to plan the overall design (engineering) of a web presence to appeal to human visitors and appease the social calling of Google. I know for a 100% fact that this video is the beginning of a marketing ploy. A herding of the sheep… I here a bunch of bhaaa, bhaaa, bhaaa in here. Just beware of the wolf in sheep clothing who is after the contents of your wallet using terrible evidence to support a claim that is not founded in reality. Instead it is founded on assumption of the teachings of others when that assumption is a mirage designed to sway your opinion. Confuse the public, show (poor) evidence using nice videos and suddenly you have a herd of sheep with whom you can shear dollars from their hide. I encourage everyone to do their research before falling victim to mass media. We all should have learned this lesson already… We’re in a recession because of it. Nice strawman technique by the way. Just kidding, you didn’t even know you were employing it did you?
Well said Charles. Well Said.
I haven’t listened to this, but Charles is really into social media, and here is a free download.
Related Blogs on lsi
- Finally SEOs are getting the LSI Myth! « IR Thoughts
- Vector Space, Probabilistic LSI, and LDA « IR Thoughts
- Bottom Dropping Out on LSI? | Munsey’s Technosnarl
- LSI – Latent Semantic Indexing | Tonetek Blog
Related Blogs on siloing
- The Importance of Using Silos in Your SEO Strategy | The …
- Linktausch: direkte Verlinkung lohnt mehr von abgeteilten …
- Advanced SEO Tips & Techniques
- Advanced Page Rank Manipulation Techniques Part II –Virtual Siloing
- Can someone explain Siloing?
Related Blogs on stompernet
- Nearly 600 Entrepreneurs Gathered at InfusionCon 2009 for …
- Chicagoist’s “Beer of the Week”: Jolly Pumpkin’s “Luciernaga …
Related Blogs on themezoom