Using a Google Experiment to monitor the impact of Lunio

The best way to judge the impact of Lunio on campaign performance is using the ‘Experiment’ feature in Google Ads.

This allows you to effectively split a campaign into two halves – one with protection enabled, and one without.

Using the Experiment method gives much clearer results than testing different campaigns against each other, as it means there are no other factors separately influencing the results of the campaigns.

To set up an experiment, in Google:

• Sign in to your Google Ads account.
• On the left hand menu, go to Experiments -> All Experiments.
• Click + to create a new experiment, and select ‘Custom Experiment’
• Once you’ve reached the ‘Set up Experiment’ page, enter the name of your experiment, for example ‘Lunio Test’.
• Select the ‘base campaign’ you’d like to test (effectively the campaign we will ‘split in half’).
• Hit ‘Save and Continue’.
• You can now choose to ‘Schedule’ your experiment, as below:
  
• When selecting the goal you’d like to judge results on, we’d recommend using Cost per conversion:
  
• By default, the experiment will use 50% of your base campaigns budget, and run indefinitely.
  You can leave this with default settings and end the experiment when you are ready, or you can amend this to your preference.
• Hit ‘create experiment’ to complete the set up!

Once the experiment has been set up, in Lunio:

• Under Google Overview, hit ‘Refresh Campaign Data’ to pull the experiment into Lunio.
• Once the Refresh is complete (which may take a few minutes, depending on the size of your account), find your campaign and experiment under ‘Google Ads Campaigns’ in Lunio.

For the first 1-2 weeks of the experiment, ensure that both your base campaign and experiment are set to ‘monitored’. Check the data is the same across both campaigns, which will give a good baseline to work from:

Once you have a baseline of data, switch your experiment to ‘protected’, leaving the base campaign as 'monitored' - using these two campaigns, you can then judge the impact of protection: