However, despite having written so much on the Internet, we have not found much useful information on how to optimize Maximum Performance campaigns with the twitter database resources and options available so far . That is why, at alGenio we wanted to make a small compilation of recommendations when it comes to optimizing this type of campaign, based on our own experience.
How to optimize Performance Max campaigns - alGenio

The first thing to do is to decide on the type of campaign setup. By default, whether you move your existing Shopping campaigns or create a new Performance Max campaign, a “resource group” will be created that will include a final URL, various image and video creatives hosted on YouTube, titles and various descriptions, so that the artificial intelligence can test and decide which of the options works best. If you create a resource group, the publication of your ads on the Display, YouTube and Google Discovery networks will be activated by default, which will use mainly graphic resources. In addition, by default, the “Final URL expansion” option will be activated, which will allow the URLs of the products in the feed to be replaced by landing pages and titles that Google Ads considers more relevant according to the custom audiences based on the intention that we have provided during the campaign configuration.
For those of us who have been managing campaigns for a while, this mix of options is a major internal debate, as it does not allow us to analyse afterwards which combinations of titles and descriptions, placements, networks and search partners , products and search queries have worked best. So far, we can only see the volume of impressions on placements, conversions by product listings and a report on which combinations of titles and descriptions have performed best according to Google, but without specifying what it uses to evaluate it. This inability to cross-reference data to analyse is what prevents us from using advertising investment to obtain relevant information that helps us improve our businesses from within, and react to our users' queries by adapting our copies .