Fair point, but how about this: they've committed to 3 months anyway, so why not try charging for 3 months? Just don't touch the money coming in during that period. If they reach the end of that time, and it's not working, issue refunds and shut down. There is some cost to this (transaction fees etc), but seemingly minimal compared to other running costs they've already committed to (because if it fails, the number of transactions will necessarily be low).

They might struggle to get people to sign up when there is doubt about the service surviving. But that cat is already out of the bag, and as I alluded to earlier they could also play this to their advantage (charging now is not a money-grab, it's simply a matter of keeping a valuable service viable).