Online Marketing – Marketing Don

How I dramatically increased conversions with a simple A/B test

I love testing. Everytime I complete a test, I am amazed at the results. This is another example of how completing a simple test can significantly increase conversions.

The testing strategy. How I decided what to test:

In your Web Analytics program identify some of your top pages and which keywords are bringing traffic to those pages. Here is a way to do it in Google Analytics:

Under Content
Click “Top Landing Pages”
- Pick one of the higher traffic pages.
- Pick a page that you want someone to take an action, has an “add to cart,” or fill out form button on it.

Click on the page url in Google Analytics to expand it even further.

There will be 3 pull downs under the graph. The first one defaults to “Content Detail.” Click on that first pull down and change the selection to “Entrance Keywords.”

This will help you identify the keywords that are bringing traffic to that page.

If you have Goals with values set up in analytics, you can also use the $Index suggestions Google gives. I havent found this to be the most reliable, but it is pretty easy. You can find it by going to Content then Top Content.
The last column will be the $Index Column. Sort the list of pages by the $Index column by clicking on it. The $ Index value tells you how much each page on your site is worth. $Index is (Ecommerce revenue + Total Goal Value) divided by Pageviews for the page(s).

You now know which page to test and have an idea what your customers are looking for when they get to that page. This is unbelievably powerful information.

The test: Change H1 header text to the targeted audience

Test A: The H1 header text stays the same.  It is general in nature and was something like More For Less.

Test B: Change the H1 header text to a term that brought customers to the page.
*I am being vague on purpose.  I do not want you to copy my text. Each site is different. I am writing this post to help you pick a highly effective A B test for your site.

Here is an example of B: If you are selling phones and people are coming to that page on the term “easy to use phones.” Then you would change test B to something with “Easy to Use Phones” in it. 

Now when potential customers come to that page, half should see the original H1 “More For Less” and the other half should see “Easy to Use Phones.” 

Track the conversions of that page. If you are using Google web optimizer, it will tell you when the test is statistically significant and can be shut off.  If you are not using Google, you can use this free calculator http://mixpanel.com/labs/split-test-calculator

Now that you have read the entire article, I will tell you how much our conversion rates increased for one of our services when I did this simple A/B test. I saved it for last, because you would have probably not have believed me if I mentioned at the top and you may not have read the article. I did this test and the conversion rate increased 111%.  That is not a typo. Not 1.11%, not 11.1%, but one hundred and eleven percent.  I will guess you will not have those results, but I will also guess your conversions will increase. Even if the rates increase by a small percent, this simple test would be well worth it.

I would love to hear from people that tried it and post your results, good or bad.

Tags: a/b test, analytics, increase conversions, testing, web analytics, web testing