Cutting to the chase as always. The “Why” you would want to do A/B Testing will be covered in a separate post but once you have the “Why”, here’s the “How”:
[0. Make sure you have Google Analytics installed on your site]
1. Create a separate version of your website with the variation you want to produce on a separate url e.g. if your homepage is www.example.com, you might upload a new page to your site called www.example.com/?abtest=1 or www.example.com/index2.html. The naming convention is completely up to you.
2. Decide what goal you want your user to complete on the site, place an order? fill in a contact form? connect to live chat? etc. You can track the goal completion by doing one of the following: 1. Have the user access a particular page (e.g. thankyou.html or ordersuccess.html etc); 2. Have the user stay on the page for a specific duration; 3. Have the user access x pages per visit; 4. Have the user perform a particular event e.g. play a video, click a button etc. For options 2 and 3, you can let Google handle this. For option 1, you need to have your developer setup the page with Google Analytics on. For option 4, you need to use Google Events, here are the instructions
2. Go to http://www.google.com/analytics, login and find the data for your site.
3. On the left menu click Behaviour –> Experiments
4. Click Create Experiment and go through the setup wizard. (NB: if your goal is not already setup in Google Analytics, you will need to set that up by clicking “Create new objective” in the setup wizard)
5. Insert to your page (just after the opening <head> tag) the code the Create Experiment wizard gives you.
6. That’s it, you’re done. You should see data coming in within 1 – 2 days.