When we did voluntary subscriptions, getting revenue was only half of the goal. They also let us do a real world pricing test with the best data possible. People’s actual willingness to pay. So when we set the prices for our plans, we set them at $5/month (the single most popular amount people were willing to pay) for our personal plan and $15/month (the most expensive amount that a significant portion of voluntary subscribers chose to pay) for our everything-included pro plan.
We had a huge initial user base and so part of our thinking with our free plan was to make sure that there wasn’t a lot of room for copycat services to come in and undercut us. And that worked out pretty well - we’re at least 10x bigger than all of the clones to this day. If we’d had a viral component to the service (we didn’t, nobody wanted the people they were emailing to know they were scheduling emails to send later!) then we would probably have skewed more toward the free plan.