IBM Data Visualization Tool Turns Out to be Amazing for Keyword Research

Wow. IBM's Many Eyes, a shared data and discovery tool, turns out to be a powerful tool for keyword research for MarCom and Search Engine Marketing purposes (hat tip to Sam Lawrence for piquing my interest in Many Eyes). I was trying to figure out what keywords and messaging to use for an upcoming product launch and the associated online marketing campaigns, and wanted to see how marketers are framing the conversation around some niche topic areas.

I started by reading a ton of posts from a selection of popular marketing blogs, but quickly became overwhelmed by the immensity of the conversation (marketers tend to, you know, be ridiculously verbose). So, if you know me, I quickly turned to data tools. In this case, IBM Many Eyes offered a great way to analyze text patterns and keyword density. It has two different tools to leverage.

First, you can create tag clouds out of any text. So I started with the Forrester Research Marketing Blog, and copied the 20 latest blog posts into Many Eyes (minus the author bylines to parse out repetitive names and dates). I set the tag cloud to 1-word display. My goal was to see what words were being used repeatedly. Here's what I found:

It was pretty noisy, but gave a good idea of recurring words. Then, I wanted to dig a little deeper, so I switched to 2-word analysis. This showed some more long-tail frequency, perfect for SEO and PPC:

Now we're starting to see some interesting results! Here's what's in the forefront of some leading industry analysts' minds. So, we know "brand monitoring" is a popular topic. With my SEO/PPC cap on, I also want to know what variations of brand-related terms are being used. Next, I turn to Many Eyes' Word Tree tool, and enter "brand" as the root word. After sorting by Frequency Order for easier viewing, we end up with this:

This gives us a pretty deep dive into long-tail terms, as well as what's the hot topic in the space. It makes sense that "brand monitoring firms" are pretty high profile, as there are a slew of acquisitions going on (JD Power bought Umbria, etc). And, it's pretty related to our space, as we recently evaluated a number of vendors, as well as explored a few partnership ideas.

Next time around, I am going to throw a bunch more data into Many Eyes from a variety of sources (it can take up to 1MB of data at a time) and get a wider slice of the conversation. You can also upload 2 separate blogs, and compare the 2 side-by-side on a keyword basis. Let me know if you have some even better ideas to use this (or other tools) to dig into social media conversations.