Once a PMax campaign is launched, we watch its performance closely. We monitor serving momentum, PPC metrics, and audiences to optimize the campaign. This optimization leads to the transition of the campaign from manual bidding to automated bidding. From then, we continue to manage the campaign from disapprovals to other critical issues that may arise.
PMax campaign optimization is different from other types of paid search campaigns. It’s more automated and more data is hidden from the advertiser. Some refer to it as “a black box”. So reporting beyond the basic metrics can be a challenge.
Most of PMax ongoing optimization requires an indirect approach. For example, by including the same Assets in several Asset Groups but varying the targeting, we can by process of elimination determine which targeting suggestion performs best. Building more structure into the campaign pays off as we eliminate weak channels thorugh the process of elimination. We can also do SKU-level optimization by comparing the performance of the same SKUs across differently-targeted Asset Group.
One of the tricky aspects of optimizing PMax occurs when PMax campaigns acquire brand conversions on the search channel, effectively swiping them from brand search campaigns. This inflates the performance of PMax leading to an overstatement of performance. Recently, the platforms have begun to allow negative brand keywords to be deployed in PMax, but there are issues with that. PMax continues to change and improve.