Hypothetically let’s say you’ve successfully set up your AdWords account. You built a campaign (with all the settings right) and a couple ad groups (with well-targeted keywords) with two ads so you could test which is better. The account has been running for awhile and one of ads is performing better than the other. But is the test valid or is it just chance?
Don’t let the big words scare you, statistical significance is just a fancy term that mathematicians use to say, in effect, the difference is legit (and not the random 1 in 1000 chance of a bad ad doing better than the good one); the mathematical way of validating what will probably look like an obvious answer. So how do you apply statistical significance to your small business PPC? Two ways.
- Rule of 30 – Make sure you have at least 30 responses (can be either clicks or conversions) and that the results are at least 10-20% different. For example, each ad ran 1000 times, one was clicked 25 times (2.5% CTR) and the other was clicked 35 times (3.5% CTR). You have 60 responses and the second ad received a CTR that was over 25% better. Replace the underperforming ad.
- Use a Statistical Significance Calculator – You can go out and learn the math, or you can use a simple online calculator like this one. For a CTR test you put the impressions in as the sample sizes and clicks as the responses. For a conversion rate test, clicks are the sample sizes and conversions are the responses. Note: Conversion rate tests will take much longer than CTR tests.
Running your tests through this simple step will ensure that you make the right call on winners and losers while testing your small business PPC.