This article will show you how I would use The Last Keyword Tool to create a website blueprint. A poor man’s Krakken, if you will. There is, of course, much more manual work involved as nothing is done for you, however you can indeed do the keyword research necessary to build an entire blueprint.
Obviously, in this article I will be revealing the basic guts of the algorithms for how we select the keywords in our blueprint functions – though there are no synonymic sets in TLKT, so they won’t quite be the same 
The first idea we should look at is keyword relevance. I’ve covered this a bit before. The easiest way to do this is with “LARI” – an algorithm that combines the LSI and Bayes. Let me give you that formula here:
((((keyword_bayes_score – min_bayes_score) / (max_bayes_score – min_bayes_score)) + ((keyword_lsi_score – min_lsi_score)/ (max_lsi_score – min_lsi_score) * 1.3))/2.3) * 100
You can use this column in a user defined filter like this:
(((((keyword_bayes_score – min_bayes_score) / (max_bayes_score – min_bayes_score)) + ((keyword_lsi_score – min_lsi_score)/ (max_lsi_score – min_lsi_score) * 1.3))/2.3) * 100) > 50
Notice that I added an extra pair of parenthesis and then added the parameter ‘> 50′ so now this filter will return only those keywords with a LARI > 50.
The beauty of this is that you can tweak the number 50 up and down to your hearts content. You will find the optimum value for LARI will be slightly different with each market. Play with it until the results look like what you want.
Ok, so that covers relevance. Anything with a LARI > 50 will be pretty relevant.
Let’s look now at “scope”. I define scope as the number of competing pages. The number of competing pages gives you an idea of the extent of the conversation on the internet about a given topic. Generally speaking you want things with the broadest scope at the top level of your website and things with the smallest scope at the lowerest level of your website. In this way a website about dogs will have a silo about dog food and that silo will contain articles about the different brands of dog food or flavors of dog food or types of dog food. You get the idea.
This is a recommendation. It does not always hold true. One classic example is a friend of mine who created a website on “health and wellness” – she had a silo on “antioxidants” and in there she wanted to include an article on “acia”. Acia however had a scope that was bigger than antioxidants and actually even bigger than “health and wellness”!
How can this be? We did some back engineering in Krakken and found that “acia” actually falls into the market category of “diet” – it is touted as a weight loss product. And doing a quick Google search verified that this was indeed the case.
This did not negate the fact that ‘acia’ is an antioxidant, however, and my friend then had to decide if she was going to include some kind of a diet silo or stick with her original idea of antioxidant – because of her USP, we decided that antioxident was the better choice.
Now that you have an idea of the advantages and pitfalls of scope, let’s talk about how to create a filter that shows us scope.
Typically what we do in Krakken is we use the parent term as the starting point. What is of interest is to see the keywords which are the next “level” up and the next “level” down. You can define a “level” to be anything you want, but today we are going to look at “magnatudes” as our levels.
In this was a level would be defined as 10, 100, 1000 – such that the next level up is 10 x the number of competing pages of the parent term and the next level down is 1/10 x the number of competing pages of the parent term.
Our filter for the next level up looks like this:
keyword_google_competing_pages > seed_google_competing_pages * 10
and our filter for the next level down looks like this:
keyword_google_competing_pages < seed_google_competing_pages / 10
This works great, but I’m getting keywords that aren’t very relevant… so let’s combine this now with the LARI filter we had before:
keyword_google_competing_pages < seed_google_competing_pages / 10 and (((((keyword_bayes_score – min_bayes_score) / (max_bayes_score – min_bayes_score)) + ((keyword_lsi_score – min_lsi_score)/ (max_lsi_score – min_lsi_score) * 1.3))/2.3) * 100) > 50
I applied this to a “coffee” theme I had and got some great results, though down around a LARI of 50 the topics weren’t quite what I was hoping for so I increased the filter to 65 and that was much better:
keyword_google_competing_pages < seed_google_competing_pages / 10 and (((((keyword_bayes_score – min_bayes_score) / (max_bayes_score – min_bayes_score)) + ((keyword_lsi_score – min_lsi_score)/ (max_lsi_score – min_lsi_score) * 1.3))/2.3) * 100) > 65
But now I notice that I’m also getting keywords that are really niche – I have an upper bound on the number of competing pages, but I don’t have a lower bound, so the numbers of competing pages can be really small. Let’s fix that:
keyword_google_competing_pages < seed_google_competing_pages / 10 and keyword_google_competing_pages > seed_google_competing_pages / 100 and (((((keyword_bayes_score – min_bayes_score) / (max_bayes_score – min_bayes_score)) + ((keyword_lsi_score – min_lsi_score)/ (max_lsi_score – min_lsi_score) * 1.3))/2.3) * 100) > 65
You’ll see I added the second line:
“keyword_google_competing_pages > seed_google_competing_pages / 100 and”
to keep the keywords from having too small of a scope.
Now I have some really great candidates for silos themes for my coffee website. I’m off to drill into a few of these… Then I’ll use this same filter to find the articles for the next level down…
And during all of this, of COURSE, I’ll be drinking another cup of coffee!
Oh… if you’re not currently subscribed to The Last Keyword Tool – you might want to give it a spin.
To your success!
–Sue
Theme Zoom Krakken is an integrated application suite that radically combines Market Analysis, Competive Analysis, Keyword DNA Creation, Automated Silo Structured Blueprint and Website Development.
For more information visit Theme Zoom.
About SueSue Bell is the CEO, founder and mastermind behind Theme Zoom's 'Krakken' Search Engine Marketing Tool, The Last Keyword Tool and Domain Web Studio (DWS). Her expertise with system design and databases allowed her to back engineer popular search engines and discover methods for Search Market Analysis and Competitive Analysis that turn SEO and search engine rankings into a science. This resulted in the creation of tools which make the life of an SEO/SEM easier by automating both the difficult and mundane aspects of Market Research, Keyword Research, Website Site Structural Design and Implementation. Sue retired from her career supporting the Military in her thirties and recently came out of retirement to mastermind the programing behind our powerful and proprietary market analysis tools. She currently enjoys living in the arid Arizona desert, spends her spare time dominating online markets, and is available on a limited bases for online marketing consultation and training.
