How should you approach evaluating video action campaign performance to drive more conversions?
Compare your CPA to that of brand Search performance and allow three to seven days to achieve desired CPA performance before making adjustments.
Compare your CPA to that of non-brand Search performance and allow three to seven days to achieve desired CPA performance before making adjustments.
Compare your CPA to that of brand Search performance and allow one to three days to achieve desired CPA performance before making adjustments.
Compare your CPA to that of non-brand Search performance and allow one to three days to achieve desired CPA performance before making adjustments.
Explanation
Analysis of Correct Answer(s)
This approach combines a realistic performance benchmark with an appropriate timeframe for Google's machine learning to optimize effectively.
-
Benchmark (
non-brand Search): It is best practice to compare your Video Action campaign's Cost Per Acquisition (CPA) against your non-brand Search campaigns. Both campaign types target users who are still in the consideration or discovery phase and may not have a pre-existing preference for your brand. This creates a fair and relevant performance comparison. -
Timeframe (
three to seven days): A waiting period of three to seven days is essential before making major adjustments. This allows the campaign to exit the initial learning phase, where the algorithm gathers the data needed to optimize bidding. This duration also accounts for conversion lag—the natural delay between an ad interaction and a user completing a conversion—ensuring you evaluate performance based on more complete data.
Analysis of Incorrect Options
-
Comparing to
brand Search: This is an inappropriate benchmark. Brand Search campaigns target users already looking for your specific company, resulting in a significantly lower CPA. Using this as a target for a Video Action campaign, which often reaches new audiences, is unrealistic. -
Allowing
one to three days: This evaluation window is too short. Making decisions this early means you are acting on incomplete data before the system has had a chance to learn and stabilize, which can disrupt the optimization process.