Five truths (?) about Google Ads AI:
1. It is true that the Google Ads AI, with broad match, smart bidding, and PMax allows advertisers to get access to ad inventory that was previously hard to target in a profitable manner.
2. It is true that, with enough time and money, the Google Ads AI will be able to learn and predict the conversion value of the clicks from this additional inventory.
3. It is true that this learning requires data, and that data costs money. Our money. (We pay for the trial-and-error that is part of the machine learning.)
4. It is also true that this cost is hidden in the reported averages in the Google Ads interface.
5. It is also true that Google doesn’t want you to prevent this cost. It’s becoming harder and harder to limit our ad impressions to targets that have proven conversion value.
We are forced to target Google’s unsold inventory. Inventory that is bursting at the seems with low quality clicks. We are forced to pay for the machine to learn how to weed out these low quality clicks. We are forced to continue doing so, because that’s the area where Google’s ad revenue has the most growth potential.
Please tell me I am wrong.
– Nils