When creating a campaign experiment,which is a best practice?

Isolating one variable per test is crucial for accurately determining cause and effect. If multiple variables are changed simultaneously, it becomes impossible to definitively attribute performance changes to a specific alteration. For example, if you change both your bidding strategy and your ad copy in a single experiment, you won't know which change (or combination) led to the observed results. This best practice ensures that your conclusions are valid and actionable. Choosing two or three metrics is a good practice, but the core principle of an experiment is isolating variables. Assessing performance over a period that includes ramp-up is also a good practice for accurate measurement, but again, it's secondary to isolating variables for a valid test.