Keyword Tools: Segmenting your keywords to optimize conversion

James Standen Analytics Canvas Tool, Keyword Analysis 5 Comments

Getting organic search to your site is important. There are lots of keyword tools available to help you pick which keywords to target, and who is searching for what, but what is actually happening on your site- which keywords are converting?

Getting lots of traffic to your site will give your web server a workout. Getting lots of converting traffic will drive revenue.

Are you wasting time on keywords that don’t pay?

You have to grab your keywords by the long tail.

The most important keyword question: Do they convert?

You use the first type of keyword tool to decide which keywords to target, and to generate content, but you have to look at your web traffic to know which ones are converting.

I am going to show you how to segment your keywords from Google Analytics (or whatever you use) and make sense of all those phrases, and find whats really important- not just traffic- conversions.

Tracking keyword segments over time- Visits OK, Revenue better.

Lets say you do segment your keywords- one thing you could look at was the number of visitors you get from each segment. Maybe it looks like this:

Sure, you can understand which type of keywords deliver the most visitors to your site. If traffic is your end goal, then this would be fine. But even more interesting, you can see which keyword segments contribute to your revenue- this lets you understand what your most valuable customers are interested in finding and seeing- and lets you optimize your content generation to create the kind of high quality, timely content that attracts your best customers. If you are running an ecommerce site, Visits aren’t usually the end goal- you are in business, and revenue is the better metric. A wide spread of conversion rates, and transaction sizes might make the segments that drive the most visitors be completely different than the segments that drive the most revenue.

Look at these two graphs- because of different underlying conversion rates and transactions sizes, you’d be chasing the wrong keywords if you were looking just at visits.

Repeat after me- “I don’t want just traffic. I want converting traffic.”

The tail is long, but there is structure in there

Even though there are thousands and thousands of different keyword phrases, obviously many of them are related. The first key step is to group keyword phrases together.

In the Analytics Canvas tool, you can do this by defining segmentation rules.

How they work:

  • Each rule answers TRUE or FALSE when given a keyword phrase
  • The rules are put in order of precedence. The first one to solve to TRUE defines which group the keyword phrase is in
  • The name of the rule is then placed in a new column for each row in the result data set

This means that with the right set of rules, tens of thousands of keywords will be grouped into a handful of groups, and then we can understand the trends and performance of those keyword groups based on our web traffic.

The order of precedence is important- since no matter what rules you define, its likely that there are keyword phrases that will result in a “TRUE” for more than one. Analytics Canvas makes it easy to experiment with rules, and their order, to create meaningful keyword segmentation strategies.

See its done using the Analytics Canvas Tool

If you want to see how its done, check out this Tutorial video that in less than 2 minutes, shows you how easy it is to define keyword segments, step by step.

Evolution of keywords over time

One of the key capabilities that Analytics Canvas gives you is not only the ability to segment your keywords, but to see how your keywords, (and their segments) evolve over time. Because you can pull huge amounts of data from Google Analytics easily, you can include ALL your keywords, and calculate segmentation across the enter data set, with a day by day breakdown, if you like.

Signup for the free trial now, and discover all the things you don’t know about the keywords people are using to find your site, and how they are evolving.

While the examples we used here are with Google Analytics, we’ve got Analytics Canvas users that are connecting to data from other systems (Web Trends, for example via the ODBC interface) keywords are keywords, data is data- Analytics Canvas is ready to take yours on.

Comments 5

  1. Hi James,

    I also find segmenting to be a fantastic way to see which types of keywords are converting well, and which are converting poorly. One technique I often use to segment and assess the performance of different types of keywords is to tag ad group names based on the amount of qualification its keywords contain:

    You can then quickly and easily compare different types of keyword segments, and adjust bids and budgets accordingly. I often find it very useful. Have you tried something like this before?


    1. Post

      Hi Alan,

      Certainly segmenting and grouping is key- whats interesting about keywords is that while rules can be defined, there is inevitably lots of human intervention and design involved- natural language processing is not yet up to the task 🙂 (Although now perhaps the natural language camp will descend on us and argue that techniques are getting more powerful every day- which they are.)

      The bottom line is, keywords are key, both for non-paid and paid search. Google Analytics exposes lots of great adwords information through the API, so all that is also available in Analytics Canvas and the opportunities to do great analysis are many- thanks for your comment!

  2. Pingback: PPC Weekly Roundup – Mid April, Social PPC Edition

  3. Hi James,

    I totally agree with the approach but it can actually all be done directly in Google Analytics using profile filters. I wrote a couple of blog posts about this a couple of months ago. You can take advantage of unused campaign parameters such as Campaign for Organic Search and populate these with your search term category names using regular expressions to define the rules. An alternative to search term categories is the number of keywords in each search term, also valuable information.



    1. Post

      Definitely Google Analytics can different types of segmentation, and often is a good place to do it. As you mention in your blog there are various approaches.

      I think you’ll find if you want to create more sophisticated segmentation and analysis, particularly including data from external sources, pulling the data out and Google Analytics is a powerful technique, and Analytics Canvas gives you some pretty high end tools to do it.

      Thanks for your comment!

Leave a Reply

Your email address will not be published. Required fields are marked *