To start off, what is A/B testing? It refers to testing out a control group versus a variation of a web page. The A refers to the control group while the B refers to the variation. Control groups in A/B testing are the original web pages that you have in place. Variations are the modified web pages where you made a change from the original page.
So, what is the point of doing A/B testing? The reason behind doing A/B testing is to see which page is better optimized. In other words, it is to see which page draws in more traffic, gets more clicks, has more sales, or signups. The exact metric you are looking to improve, and measure will vary depending on the site you have and what you are trying to do. So, to summarize, the reason for doing A/B testing is to be able to measure and find out which page is more effective at getting conversions which are clicks, signups or sales.
Getting an A/B test done is pretty simple. You just create a modified version of the page you want to test. This is the B page or variable page. The page can be completely different or altered slightly. For example, you can just change the headline or graphic or you can build an entirely new page. Whatever you decide to do, the next step is to make the page visible to some of your visitors.
A/B testing, which is also known as split testing involves showing half your customers your original page and then showing the other half of your visitors or customers the new modified page. You then gather data about each of the pages to determine which one is better. For example, you can gather data such as the number of sales, sign ups, number of time spent on the site etc. After comparing the data from the A/B test, you should be able to clearly see which site is more effective at conversions.
After A/B testing is complete, the person testing the site will usually take the website that had better metrics such as more sales and make it permanent. The results of A/B testing can vary however. Sometimes one site will be better at conversions. Other times, however, there will be no significant difference, so it may make sense to keep the original web page instead of the modified or B version.