My thought on this is that offering to pay it back if the initiative fails speaks highly of your ethics.

But marketing is always a case of experimenting until you get the mix just right.

So although I fully support this proposal (especially the simplified onboarding and follow-up user education aspects) and think it'll succeed, in the event that it does fail I would rather not see the cost repaid, I'd rather see that amount cycled back into the next experiment.