Wednesday: A/B Testing
Civil engineers know that the best kind of earthquake testing for a building is this:
. A/B testing is a way to get that type of real-world information for
your SEO campaign.
Commonly used in the direct mail industry, A/B testing (also called an A/B split)
is a practice of sending out two different advertisement designs and comparing their
sales. On the Web, A/B testing can be used to compare the conversion rates for PPC
ads, for landing page designs, or even for two different “Buy Now!” buttons.
There are several ways to approach an A/B split. Read through the following
options and determine which one best fits your campaign.
• A/B Testing: Proper
• A/B Testing: Practical
• A/B Testing: Page-Based
A/B Testing: Proper
A scientifically robust A/B test would follow these steps:
• Create two identical PPC ads.
• Point both ads to the same landing page.
• Let both ads run for a while. The period of time (be it a day, a few days, a week,
or even longer) before they are both showing approximately the same click-
through or conversion rates is your testing period.
• Keep one ad the same throughout the test (this is called the A ad).
edit to the other ad (for example, change “Purchase low-cost dental
insurance” to “Purchase discount dental insurance”).
• After each edit, wait one testing period. Did the change increase or decrease the
ad’s performance (either click-throughs or conversions)? Don’t forget to keep
• Try again with another edit.
• Lather, rinse, and repeat. With each testing period you learn whether the edit
The scientific approach is great, but we know the real world rarely presents the
opportunity for ivory tower–style research. And that’s why we’re hoping you’ll look
into the approach described next.
A/B Testing: Practical
If you’re using Google AdWords, A/B testing for PPC ads is easy! Just write one or two
additional ads for each of your Ad Groups. As clicks come in, Google automatically
c09. 8:10 257