I see this happen so frequently, and my co-founder and I are grappling with exactly this issue. Our model will rely on subscription rev for a good portion of total rev once we are at full capacity, but we won't be there at launch. The approach we are taking now is deliberately holding back from V1 and V2 the features that we think will be the real value-ads that will entice paid memberships. That way, instead of asking people to pay for a think they previously got for free, we will be adding new features and asking people to pay if they want more - or stay on the free membership and continue doing what they are already doing.