Monday, May 2, 2011

5 Keys to A/B Testing



Through A/B testing you will be able to tweak and test different variations of websites, landing pages, and other online marketing materials. By A/B testing you are allowing the consumer and the data to prove which idea is the best rather than just speculation. I have 5 quick tips from the years I have spent doing A/B testing.

1. Analytics - it is key to have a good analytics system in place prior to testing. There are some great programs for A/B testing like Sitespect, Omniture, and Google. My company, Neutron Interactive, uses a proprietary in-house system for testing. Whichever analytics system you choose, make sure it is robust enough to allow for scalable changes and splits for testing. You will also want to determine which stats are most important to you during each test (ROI, Click-Through, Bounce Rate, etc...)

2. Target Market Understanding - whether you use panel, syndicated, or other data, it is important to know your audience for what you are testing. A company I do consulting for has a very distinct target audience, but within that audience there are many different personality types (analytical, personal, direct, etc. ). Because of these, I created various landing pages to speak to each audience separately and then to each audience as a whole. The results of the A/B testing variants showed that the analytical page produced the best results. This A/B test not only showed me which page to run, but it also gave me an even better understanding of my audience for that specific promotion.

3. Design - Once you have your analytics setup and you have a basic understanding of your audience, you will want to work on various designs for what you believe will be successful. Try different graphics, text, buttons, colors. The key is to Always be Testing.

4. Analyze Results - Once a large enough sample has been gathered, you will want to determine which test produced the winner result. This may change depending upon which stats you weigh the most. Once you find which test is the best, that page or design will become the new base or control for the next test.

5. Continue Testing - the goal of A/B testing is to continually increase the effectiveness of your marketing, and with each test you want to exceed the control page. Some tests won't be as effective as others. With continual testing you will learn more about your audience, keep a fresh eye on the evolving marketplace, and increase overall brand awareness and reach.

Always Be Testing and Know Your Analytics. These 2 things will ensure continued success in marketing practices.

6 comments:

  1. Ray Lund Via LInkedIn: Testing is one of the most important things a company can do to be successful in their direct marketing. Too often new marketers will jump from one marketing channel to the next, without testing and never get the true benefit from any of them.

    I have some great testing A/B and A/B/C/D testing plans: Here's the link: http://www.thelistwarehouse.com/T3/effectiveness-of-email-marketing-plan-testing.html

    I hope this is helpful.

    To Our Success!

    ReplyDelete
  2. Jason Sokul Via LinkedIn: Great post. Thanks for sharing.

    ReplyDelete
  3. Emma Hartsfiedl Via LinkedIn: Interesting article. When you do A/B testing for a client do you test every page on the website? For example, right now I am doing some A/B testing to see if a button on the top right of the page gets more people to fill out our contact form. I am testing two different buttons: "Contact Us for a Quote" and "Contact Us for a Demo". Right now I have the experiment running on 3 different pages (some of the highest traffic pages) if you were doing the testing would you test all the pages on your website?

    ReplyDelete
  4. Emma, I will send you an email. I do various testing depending on the pages and the audience.

    ReplyDelete
  5. Andy Hedrick Via LinkedIn: Thank you for contributing the article Joe. There are some great tips here.

    ReplyDelete
  6. Tom Sullivan Via LinkedIn - Nice A/B Overview

    ReplyDelete