I'm not suggesting that.

You should run small experiments (<$50) per day until you figure out that you get can profitable. Only then do you scale up. The point is understanding what the microeconomics look like, so that you know how to grow the business over time profitably.

If you can go viral, that's even better obviously. I've written extensively on that also - and it's my preferred approach. But it's just true that for most subscription products, it's easier to buy the users than to acquire them virally.