How to Conduct an A/B Split Test
How do you know which words are the right words to use to convert your customers and increase their lifetime value? Through testing!
You test your headlines, your subheads, your body copy, your guarantees, your offers, your prices, your order forms, your background color, your font style and every other element that contributes to the entire online selling process.
Writing one sales letter and running with it—even if you think it is the most brilliant sales letter in the world—is unwise. Because that leaves you with no way of knowing what might convert better.
And here’s the thing: testing is simple. Most people are just too lazy to do it!
The most simple type of test to conduct is an A/B split test where only one element is tested at a time, like price.
For example, let’s say you are currently selling a product for $39. And you want to know how raising your price would affect your conversion rates. So you run an A/B test.
You keep your control page active at yourwebpage.com then set up a new url, perhaps something like yourwebpage.com/test. On this new page, you duplicate everything on your control page and change only one thing: the price. Instead of $39, you charge $49.
Now you set up your traffic generation sources to direct half your traffic to your control page and half your traffic to the test page for a period of two weeks.
At the end of the two weeks, you compare the results: your control had a 1% conversion rate, but your test page had a 1.5% conversion rate. That’s a 50% improvement! Your customers are actually willing to spend more money on your product than you thought.
But you would have missed out on all that money if you had never conducted the test. Now you know, though, so your $49 price point becomes your control, and you test again. This time you change the headline to see which one is more effective.
And you keep testing different elements to figure out how to optimize your sales page. Or your squeeze page. You can test different headlines and “bribes” to determine which brings in the highest number of subscribers.
Running a sequential test is also an option. With a sequential test, you keep the same url and track the results for a certain time period. Then you change the element, like the price, and track the results for the same length of time.The results of a sequential test may not be as accurate, especially if a holiday or other special event occurred during either test period, but a sequential test is certainly better than not testing at all.
Glen Hopkins is an internationally renowned Internet marketer and
Glen specializes in helping online business owners build large,