Split Testing Strategies
To really improve your results, you have to know what to test
When determining what to test, keep in mind the fundamental rule of testing: "Test only one thing at a time. In order to get valid results, all variables except one must be held constant"
The only exception to this rule is if you are starting in a completely different direction, or do not have a control. See: Discovering the "Next Big Thing" or What is a control offer, and what if you don't have one?
It is essential that all variables in your offer are held constant, except for the one you wish to test. Test only things that are significant. Do not make the mistake of expecting minor variations in your creative offer to have a major impact on your results. Testing single words can make a difference, but it is rare. It is better to spend your time, effort, and money on things that could have a major impact.
"5 things to test that could have a major impact on your results:"
1. Testing Headlines: To test one headline against another, you would want to make certain that you use the same click source, and keep the creative offers identical in all ways except for the headline.
2. Testing Price: Determining what price will maximize sales of a given product should be a primary concern of management. Vertster makes this possible in a direct sales environment. Simply use two or more copies of your sales letter or sales web site with different price points. Next, create a testing group which rotates traffic between the two, and compare which generates higher profits!
3. Testing Click sources: To test one click source against another, you would want to make certain that you kept the creative offer identical for all click sources you test. This can be an interesting test to determine if Google Adwords or Overture delivers better quality traffic!
4. Testing Consultants: To test one consultant against another, you would keep the click source identical, and run a test between the two consultants creative offers.
Once you have decided what you would like to test, you should next determine how long the test will run for. For guidelines on this, see this article: [link] (How long should you run a test before you trust the numbers.)
