You need to know to what headlines, call to actions and sales copy work best for your AdWords ads. You can only learn what combination of factors results in the best ad through testing. Split testing is the most common way to test your AdWords ads.

Split testing is a simple principal. You have an ad running for your website or product. This ad has an established number of impressions and clicks. But, you want to know of a different headline or sales copy will result in a higher CTR. So you create a identical ad to the original, or baseline ad. Then you change one thing, or variable about the second ad. That variable can be anything you want to test.

• A different Ad Headline
• New sales copy
• A new call to action
• Synonyms for certain words
• Use of capital letters
• Word or sentence order and structure
• The landing page

Anything you want to test can be variable. Just make sure you come up with distinct variables to test. For example, don't call a headline change and a landing page change a single variable. If you change both these items and obtain a higher CTR, you won't know if the new headline improved CTR, or if the new landing page scored a higher quality score which then improved CTR. Instead, treat these variables a two distinct variables and run a different test for each. For example, test the headline first. Then once that test is complete, test the landing page.

You now have the original baseline ad, and a new test ad. Add them both to your ad group and monitor the impressions and clicks. The trick to split testing is knowing how to compare the baseline ad with a high number of impressions (because it's been running) against a new ad with a low number of impressions. One may have a higher CTR, but how to you know if that trend will continue as the test ad gains more impressions?

A split test relies on statistics to determine when a test is complete. You can use the chi-squared calculation to determine if there is enough sample size to draw a statistically significant conclusion and declare a winner. The chi-squared calculation will return a confidence level percentage revealing how certain you can be that an observed trend will continue as impressions accumulate. For the chi-squared calculation, a percentage above 95% is considered probably significant. This is the confidence level you need to declare a winner. 99% is consider significant and 99.9% is high significant.

The chi-squared calculation and confidence level calculations are complicated formulas. Fortunately, AdWords-marketing-tool.com offers a free calculator to do the work for you. The AdWords-marketing-tool.com calculator gives you a chance to enter your impressions and clicks for both ads and returns a confidence level. If your confidence level is 95% or above, you can declare a winning ad. If below 95%, you must continue to allow both ads to run and try again when you have more impressions. THe closer the CTRs of the two ads are, the more impressions it will take to reach a conclusion.

You simply check the CTR of both ads when the AdWords-marketing-tool.com calculator says you reached a confidence level of 95%. The calculator will display the CTRs and the confidence level. If the baseline is the winner, discard your changes and keep the original ad. If the test ad is the winner, discard the original ad and use the test ad as your new baseline.

Continue to split test as many changes as you want. You will improving your CTR with every test.