How To Do Variable Testing Using AdWords Campaign Experiments

By Nicole Bandy | Sep 5, 2013
More Articles by Nicole

adwords campaign experiments bill nye

Just like we learned from our friend Bill Nye the Science Guy, every experiment must have an independent variable, a dependent variable and a control variable. So if you want to run experiments with your PPC ads, the AdWords Campaign Experiments tool can help you do this, and set your tests up properly.

The benefit of using the AdWords Campaign Experiments tool is that it creates the conditions of a real statistical experiment by comparing your changes against a control. This also lessens the risk of unfavorable results on your whole campaign by only applying changes to part of your audience.

What are AdWords Campaign Experiments?

AdWords Campaign Experiments (ACE) is a tool within Google AdWords allowing you to test changes in a campaign and monitor the effects.  Experiments are set on the Campaign level and can be found under the Settings tab of a particular campaign. As shown below, in order to set up an AdWords Campaign Experiment, you must specify a name, a control/experiment split percentage, and a start and end to the experiment. The control/experiment split percentage will determine the likelihood that Google will serve the control group (attributes that remained the same) or experiment groups (attributes you changed) in the search auction.

adwords campaign experiments settings

Choosing Variables to Test

With AdWords Campaign Experiments you can test several independent variables at the Ad group, ad and keyword level of the campaign, such as keyword match types, bids, landing pages, etc.

ACE limits the number of experimental changes you can make per campaign at 1,000. Additionally, you cannot test the effect of using various Campaign settings such as budget, ad delivery method or bid method.

Analyzing the Dependent Variable

The dependent variable is going to be the aspect of your account you are hoping to improve with the test. For a landing page test this might be conversions or conversion rate. You do not have to do anything in AdWords to focus on a particular performance indictor for your experiment, but you will want to be aware of what the goal of the experiment is.

You can analyze your data while the experiment is running under the Segments Tab located in the Ad group, ad and keyword levels. When you select experiments, Google AdWords provides up or down arrows next to each metric representing how likely it is that your results are due to chance.  The more arrows there are, the more likely it is that your results are not due to chance, meaning there is a relationship between the experimental changes you made and the results.  When Google shows up and down arrows together, the results were not affected by your experiment.

adwords campaign experiments results

Control Group:

In the past, AdWords users tested changes using basic before and after analysis, which neglects the outside variables that can affect your campaign performance at any time. These include things like seasonal traffic or interest, an outside marketing campaign or promotion, or changes from your competitors. The control group in Google AdWords Campaign Experiments serves to limit the influence of outside factors on the test, and make it statistically fair. When you set up the experiment, you need to decide how much of your audience you want to be the control versus the experiment.

Beware that there are some limitations to variable testing with AdWords experiments. You cannot run more than one experiment per campaign at a time, nor can you test campaign level settings. See a full list of limitations here.

Questions? What will you use AdWords Campaign Experiments to test? Comment below or tweet at me @NicoleBandy.

Share this article

Share on LinkedIn Share on Twitter

Receive Monthly Digital Marketing Tips

Subscribe to monthly updates from the Marketing Mojo Blog to get the latest digital marketing tips, best practices and insights - hot off the presses and straight to your inbox!


Blog Search