What casts doubt on PPC A/B tests?

Too many testing variables at once makes it impossible to isolate which specific change caused a performance difference, thus casting doubt on the validity of the A/B test results. For an A/B test to be effective, you should ideally change only one variable at a time to accurately attribute any observed changes in conversion rates or other metrics to that specific variable. The max CPC you set is a single variable that can be tested, not a factor that inherently casts doubt on a test. Device bid modifiers are also a single variable that can be tested independently.