Meta VO for start trial event. Have anyone try this?

Hi guys,

I came across an article by Lucas Moscon about assigning a value to the Start Trial event and using it for a Value Optimization (VO) campaign. Here is the link:

I’m curious about how the algorithm actually behaves in this setup. If we assign the same fixed value to every Start Trial event, wouldn’t Meta still treat all trial starts more or less equally?

Has anyone tested this approach before? Would love to hear how it performed compared with using a fixed value.

I am using this approach in Moloco and it's working well. I haven't use ROAS for trial value but I have used trial_qualifier events and there is a significant improvement in quality (although CPA can be too high to compensate sometimes)

I strongly believe sending an average is counterproductive:
you have all the drawbacks of VO (need more scale, more volatile...) and none of the benefit.

VO needs variance.
Actually Meta isn't even supposed to let you run VO without at least 5 different values (that can be tricked, but I don't think it should)

+note that as Meta isn't a huge fan of people injecting predicted non-real values as revenue, they have this beta campaign valled LTV vision or something, which is meant to do 2 bits: handle predictions, and accept a lag for update up to 7 days (VO can also be done up to 7 days but it sucks horribly after day0)