Why guess when you could know?
I see a lot of people asking questions here in a fashion that seems to be an attempt to bypass the conversion optimization process, which involves testing. This is something that varies a bit from one website offer to another. Someone else’s experience or opinion isn’t always helpful as so many of these situations are specific to the particular use case.
The best thing to do is a test. I suggest you set up an A/B split test and measure the results. And do not simply compare the conversion rate, but the retained earnings after the refund period have expired. Using real data is always better than guessing, or asking opinions.