Optimising for qualified trials vs purchases with short trials vs purchases with paid trials

[maroon]

We have a 7-day trial on our yearly subscription and have used qualified trials as our optimisation event for the past few years, where we wait a few hours after the trial has started and then predict if it is likely to convert based on a few signals. The significant majority of our trial starts tend to be on D0.

We have a good trial conversion rate but are considering testing alternative optimisation events in order to get closer to actual purchases. A few we are considering:

• Purchases after a 3 day or 5 day trial - we’ve already tested this on the product side and seen no impact on conversion
• Purchases via a paid trial e.g. $0.99 for 1 week

I wondered if anyone has experience or views on whether qualified trials on D0 work better or worse as an optimisation event for Meta vs these other options (or any others)?

Cheers

 

[davamosk]

I'd definitely go with a paid trial as that action happens on D0 too. My experience with +d1 actions has been always negative as the MLs don't learn properly in real time so I always recommend targeting d0 actions if possible

 

[roasd365]

Based on my experience, paid trials (first week at $0.99) have lower ROAS compared to free trials. To be more specific, we observed around 30% lower install-to-trial conversion, but slightly better renewal rates with the paid trial.
Of course, this test was conducted only in the US and on a single title.