Leveraging AdWords Campaign Experiments
Introduction to Experiments and Its Advantages
Campaign Experiments is a testing tool that allows you to better optimize your campaigns by helping you to answer some of the following questions:
- How changes to your keywords affect the overall campaign performance (e.g. changes in impression, conversion, CTR, etc).
- What changes you will notice if you increase or decrease the bid amounts on some of the keywords?
- How your campaign will perform if you make changes to your ad groups?
There are total four steps in setting up an Experiment, explained as follows:
Step 1: Setting up an Experiment
- Navigate to the Campaign menu and then click on the Settings tab.
- Under the Advanced Settings, Select Experiment.
- Click the + Experiment settings Settings button.
- Provide a name for your experiment as well as a start and an end date. You can also manually start or stop an experiment.
- Select the Control/Experiment Split (50% option is commonly selected). It means 50% of overall activities in your campaign will be conducted using experimental values.
- Click on Save.
Step 2: Defining Parameters
You can carry out experiments on bid amounts, keywords, ad groups and placement (for content network only). Conduct your experiments with business context in mind. Use your experiments to investigate effects on conversions, average CPC, clicks, and CTR. In the following example of experimenting the effects of adding new keywords:
- Navigate to the Campaign menu and then click on the Keywords tab.
- Click the + KEYWORDS button.
- Select the ad group in which you would like to add new keywords. Enter new keywords as you normally would (e.g. instructor led training, online training, etc).
- Click the check box that says Add as experiment only keywords.
- Click on Save. Your experimental keywords will have a different status icon which indicates that they live in an experimental world.
Step 3: Viewing and Changing Experiment Parameters
To view and make changes in your Experiment parameters make sure you have enabled the experiment segmented view. Here is how you enable it:
- For the above example (keywords experiment), you will first navigate to the Keywords tab and then click on the Segment.
- Use the drop down menu and select Experiment. You will be able to see the segmented view which clearly separates your experimental parameters from your default items.
- Notice that your experiment row has a different status icon than the default items. Within the experiment row of keyword, make the changes you want (e.g. set the bid amount, etc).
- You can also take the keywords in and out from an experiment by clicking on the status icon and selecting either Control Only or Experiment Only.
Step 4: How to Start and Stop Your Experiments
If you provided a start date and an end date in step 1, then your experiments will start and stop automatically. Here is how you manually start and stop an experiment.
- Navigate to the Campaigns tab and click on the Settings tab.
- Under the Advanced Settings, select Experiments.
- Click the Start Running Experiment button to activate your experiment.
- You can stop your experiment by clicking on Stop Experiment button.
- Once you stop your experiment you can click on the 'Apply: Launch changes fully' button to incorporate all the changes made during this experiment OR click on 'Delete: Remove changes' to discard the changes and revert back to the original settings.
Step 5: How to Interpret Experiment Data
Google recommends running any Experiment for at least two weeks. However, the bottom line is that you need to run the experiment till you collect enough data to have statistically accurate data. Once your experiment produces statically significant data, Google will indicate that with an upward or downward arrow in that particular field.
In the above screenshot you can see that the experiment did not make any significant difference in the number of Clicks (refer to the Clicks column in the screenshot). However, Impressions have significantly changed, which is indicated by an arrow symbol.