Setting up Google Website Optimizer for a Multivariate Test

Posted on Thursday, June 17th, 2010 by Print This Post Print This Post

Categories - Analytics, Consumer Experience, Featured, SEM

I have come across several testing platforms, and from my experience, they all have their advantages and disadvantages. I get asked about Google Website Optimizer (GWO) quite a bit in my line of work, so I figured I’d check it out and let you know what I found. I will break this review out into two parts.  Part 1 will focus on the set up process.  Part 2 will focus on the reporting.  One major caveat: I am approaching this tool from a marketer’s perspective and not a technical perspective, so please bear with me. “1. Choose the page you would like to test. 2. Choose your page sections. 3. Review desired page sections with your technical team. 4. Identify your conversion / success / goal page.” Google leads with these instructions when you start your experiment.  This seemed pretty straightforward so I tried this out on my own.  I must admit that given my technical limitations, I was surprised at how easy it was to set up a simple multivariate test.  The instructions are easy to follow. I successfully tagged the sections that I wanted to test, e.g. the headline and the main hero image.   GWO guides you step-by-step and checks to see that you copy and paste the code correctly onto your html (see screenshot). I wanted to try something a bit more complex, such as swapping out an entire section.  I consulted our technical lead, and he was able to implement this quickly.  If you are a tech-challenged marketer (like me), I recommend that you work with a technical resource because you may run into some challenges based on the different variables that you are trying to test.  Based on our experience with GWO, we recommended it to one of our clients.  We executed the test based on a strategy that we developed.  I worked with our technical lead to set the test up quickly, and we launched the test without a glitch. Pros
  • FREE
  • Easy to use
  • Easy technical implementation
Cons
  • Cannot control traffic split between variations
  • Cannot reset data pool -- must stop the test then copy it as a new experiments (this at least allows you to keep a record of all the tests that you have run in the past with data integrity
  • Can only track one conversion point.  You can tag multiple conversions, but these will not be tracked separately.
Stay tuned for Part 2 where I will delve into the reporting features of Google Website Optimizer.

Tags | , ,

2 Responses to “Setting up Google Website Optimizer for a Multivariate Test”

  1. I want to suggest to try reedge.com as testing tool, its a nice alternative to GWO

Trackbacks/Pingbacks

  1. Weekly Search & Social News: 06/22/2010 | Search Engine Journal

Leave a Reply

Follow Us on Twitter

Authors