Thinking about in terms of what basically becomes a $30 user acquisition cost is grim enough, but it’s even worse when the fact is the system will run out of Steem!
If it truly takes 15-30 SP to have a minimum viable user experience, then the Steem blockchain currently caps out at 10-20 million users. And that’s only if everyone dumps their stake and we secure a totally equitable distribution of the ~300 million Steem in existence! In truth, under these SP needs, we’re probably capped at 2-4 million active users.
Whatever needs to happen on a technological level, the goal should be clear. If Steem ever wants to touch the user base numbers of Reddit, Instagram, FB, etc.; accounts need to have a passable user experience when holding 1 SP or less.
To do that, we'll need for the major stakeholders to decide, one way or another, if they want to be participating in a fintech crypto-commodity market or a functioning social network which uses a lightly distributed transaction ledger to keep score.
A crypto-commodity market can survive with the well under a million active users, even flourish. A social media system trying to compete with the network effects of Reddit, Instagram, Facebook, and Twitter simply cannot.
The usual counterargument to this is that as the perceptual value of the social media network increases, the value of STEEM likewise increases and thus the user acquisition cost decreases proportionately. Which would be true – if growth was guaranteed. Unfortunately, the costs for developing digital applications on the steem blockchain are frontloaded and you need users now, not in five years. In five years if accounts were magically useful at 30 SP values, that might be a real cost of under a couple of cents per user. Perfectly bearable. But this is in five years of magical unicorn growth. It's now.
I'm not sure this is a problem that can ever be solved given this technology with a competitive edge. The underlying technology has costs and inefficiencies which are not actually being addressed by the RC system, only vaguely gestured at.
A more useful system would introduce a dynamic market which witnesses could set their own prices for transactions in based on their own network, storage, and CPU costs, with clients choosing who to do RC transactions with based on who is offering the best deal for a given commit time, but we'll never see that.