-
The broken conversion goal explains most of it. Running maximise conversion value when Google had never seen a purchase event means it was optimizing toward nothing for weeks. Smart bidding on e-commerce needs real purchase data to know who buys. Without it, it’s spending blind. Those three weeks of budget taught Google the wrong signal. That historical data is still in the model even now you’ve fixed it. The existing campaigns have been shaped by broken inputs. Starting fresh with new campaigns might give Google a cleaner learning signal than trying to recover what’s there.
The other thing worth checking is that PMAX doesn’t break down which placements are driving clicks. A £0.13 CPC for your niche is very low. Shopping and Search clicks in furniture tend to cost more than that. If a large chunk of those 1.8k clicks is coming from Display or YouTube, that explains the zero conversions. Those audiences are browsing and not buying. There’s also a case for moving away from PMAX entirely for now and running a standard Shopping campaign or a search campaign targeting high-intent terms like “divan bed Ireland” or “Ottoman bed buy online.” PMAX is a black box. Without existing conversion data, it has no real signal to work from. Standard campaigns give you visibility into what’s triggering your ads and what’s converting. What does your product feed look like? Titles and descriptions specifically.
