Last weekend my roommate and I went out downtown to the bars. As we were getting ready she said to me, ” I am going to see if more guys hit on me if I put on red lipstick tonight. Last time we went out I didn’t wear lipstick.” We went out, had fun, and when we got home she told me about all her adventures. Turns out the red lipstick was a hit. This would be an example of AB testing in the real world. Well there are several factors that make this little experiment inaccurate… but it helps you understand the gist of AB testing.
AB testing is when you have two versions of an element and you are testing both versions simultaneously to figure out which one is going to generate the best results. There are several things that you need to know about AB testing before you proceed:
- Make sure to isolate one variable. This is very important because if you only change and test one variable then the results will be attributed to that one variable and that variable only. If you change multiple variables then you won’t know which change in variable generated the best results.
- Use a large enough sample size. The larger the sample size the stronger the results will be. Remember that you should also have a timeline and having an extremely large sample size will only detour you from using the results of the AB test to incorporate the variable with the best results right away.
- Eliminate confounding variables. This means split testing. For example half your visitors see the blue button and half see the red button. Or in the case of emails sending the emails at the same time. In my red lipstick example, there were many extraneous variables that could of skewed the results. Such as there were more men out at bars the second night, or the fact that she was wearing her hair differently.
- Don’t give up. Many AB tests need to be ran in order for there to be a result. So do not give up.
- Know how to figure out the CORRECT conclusion. Don’t let your hypothesis sway the true results of the AB test. Many times it is the variable that is least expected to outperform that ends up as the best option. This means that you need to figure out if your result is statistically significant or not before thoroughly examining the results.
The next question is where can you use AB testing? There are plenty of applications where AB Testing is not only extremely helpful but may actually be necessary in order to help generate the results that you are looking for.
- Layout and page design
- Headlines, opening paragraphs, & photos
- The sales funnel
So lets go back to fifth grade science? You remember the scientific method for testing? Well of course you do… but lets review just in case.
- Pose the question. What are you testing?
- Create a hypothesis and a prediction
- Test your hypothesis. To do this you must design the AB Test.
- Analyze the data. Find out which version helps achieve your desired results.
These steps will help you conduct your AB test effectively.
So what do I think about AB testing? AB testing is going to help you reach your goals and is very effective at doing so. It has been called the most effective way to turn “clicks into customers.” So why wouldn’t you conduct AB testing? You wouldn’t. Oh geez did I just used a double negative. What I am trying to say if you have any brain at all then you will use AB testing and start turning “clicks into customers.”
A/B Testing is Key to Unlocking Your Websites True Potential. (n.d.). Business Bee. Retrieved October 21, 2013, from http://www.businessbee.com/resources/marketing/ab-testing-key-unlocking-websites-true-potential/
Gotcher, M. (n.d.). A/B Experimenting: Perform Tests That Cannot Be Ignored – Business 2 Community. Business 2 Community – Building Deeper Business Relationships Through Engaging Communities. Retrieved October 22, 2013, from http://www.business2community.com/online-marketing/ab-experimenting-perform-tests-cant-ignored-0634507
New Marketing Optimizer for WordPress Plugin Released . (n.d.). Press Release Services – News Release Distribution Services – PRWeb. Retrieved October 22, 2013, from http://www.prweb.com/releases/2013/9/prweb11131529.htm
Sherbin, M. (n.d.). Create a Content Optimization Advantage with Performance Testing – Business 2 Community. Business 2 Community – Building Deeper Business Relationships Through Engaging Communities. Retrieved October 22, 2013, from http://www.business2community.com/content-marketing/create-content-optimization-advantage-performance-testing-0624418
The Ultimate Guide to AB Testing. (n.d.). Smashing Magazine. Retrieved October 21, 2013, from markstaton.files.wordpress.com/2013/10/the-ultimate-guide-to-ab-testing-smashing-magazine.pdf
Your company needs a champion for A/B testing (not ab testing) | VentureBeat. (n.d.). VentureBeat | Tech. People. Money.. Retrieved October 22, 2013, from http://venturebeat.com/2013/09/10/your-company-needs-a-champion-for-ab-testing-not-ab-testing/