Back to the Basics 2008: Keyword Research

[Chris Boggs]

In today's SEM Crossfire, "SEO Basics: Effective Iterative Keyword Research in 2008," I describe many of the methods used by our team to perform keyword research. This Genesis process for all things search engine marketing related has evolved drastically in the last three years since primary dependence on free Overture data.

I hope that everyone chooses to share some additional methods not covered in the article, or comment on ways to enhance some already mentioned.

Keyword research can make or break any campaign: organic, paid, contextual, or display. Do you agree with this? Anyone have any horror stories of keyword research gone bad? Anyone have better stories of kw research gone wild?

[Chris Boggs]

Thanks Kevin for updating that. I look forward to some additonal ideas!

[Chris Boggs]

There have to be some more ideas around keyword research out there...don't be shy!

[jimbeetle]

I've been meaning to get back to this thread, been wanting to talk a bit about from the SEO aspect.

The value of any SEO effort lies in the keywords you choose.

Over the past few years I've changed my basic approach to keywords and SEO. Guess it started when I realized I was reading many, many "What keywords should I choose for my site?" type of threads. It nagged me for a time 'cause I really couldn't put my finger on what bothered me about it as targeting specific keywords is one of the basics of SEO.

It finally it me when a question popped through my mind: "Do we really choose keywords? Or do keywords choose themselves?" I went with that for a while and realized that yes, for the most part, once you decide to write a page on a topic your decisions about what keywords to use are made for you. Of course, you make sure that you write effective titles, descriptions, heads with the most relevant (for your site) phrases, but then just write. Throw in a couple of related terms, supporting terms, synonyms -- I can't find any reason today to specifically target different pages to different combinations. I think one very strong page that shouts "This is what I'm about" is much preferred to a handful of maybe weaker pages that break up the actual relevancy.

One one of my more heavily trafficked info sites no single keyword phrase usually accounts for more than about 1.5% of referrals (some seasonal spikes might see certain phrases as high as 4%). Looking at one well-trafficked page for June, the targeted key phrase accounted for about 20% of referrals. The variations among the 80% in some cases include one or two of the targeted keywords along with some synonyms and related terms; in some cases none of the targeted keywords appear, just the supporting terms.

I started to refer to the way I currently work with keywords as "key theming," but it basically just boils down to answering the question: What is the topic of this page?

I do have to stress that this is the way I work with informational sites. I think e-commerce sites need an approach more akin to PPC, though some "key themed" pages can quite nicely support sell pages.