AI Marketing Automation ConsultantAI Marketing Automation Consultant

Most people who do A/B testing are never truly giving their testing ad a full chance at achieving its best possible results. Furthermore, without testing the ad properly, these individuals or companies may ultimately lose money, time and potentially reject a successful ad that just never had the chance to prove itself successful due to the lack of owner knowledge.

In addition to the explanation of failed ads, Thies goes further into his article and explains the process of creating a successful ad testingcampaign. The technique he uses explains that by keeping the control ad running throughout the testing of the new ad, the user can save money and also “eliminate performance history as a variable (Thies, 2008)” in the test ad. By creating multiple copies of the control ad, the user can “avoid giving 50% of impressions to an unproven ad which may fail (Thies, 2008).”

Creating multiple copies of the control ad, the user does not lose as much profit due to the control ad always running. Instead of the user losing 50% of ad impressions to the testing ad, that percentage shrinks to 20% because of the additional control ad copies.

In addition to this type of testing saving profits and potentially accelerating the testing process, it also allows for the user to create a valid testing environment by “comparing the performance of the test ad against the copies of the control only (Thies, 2008).”


Thies, D. (2008). Split testing adwords: you’re doing it wrong. Retrieved June 12, 2010, from

Dr. Elijah Clark
Latest posts by Dr. Elijah Clark (see all)
Cite this article:
Dr. Elijah Clark (May 8, 2022). Google Ad – A/B testing [Web log post]. Retrieved from
Copyright 2024 - Dr. Elijah Clark Enterprises - Sitemap - Policy - Fees