Naturally, cost-benefit ratio does depend heavily upon the requirements and usage scenario. Obviously, it's also very different if you're talking about powering just a few critical loads versus supplying/augmenting power to the entire property.
My scenario is in a rather cold climate, and also for a building a good ways from my house, where the grid connects, that does not have adequate power for my needs even when the grid is up, so I'm looking at a hybrid solar solution, capable of off-grid operation as well as grid assist (but not net-metering).
The one thing I would add to your observation is that it can very easily be worthwhile to use LiFePo4 cells over a standard car or deep cycle battery, for any type of system, for several reasons. They're safer, last longer, have greater discharge capacity, and lend themselves well to incremental replacement, since the cells don't tend to all fail at once, so rebuilding them can be far cheaper than replacing a single deep cycle battery of equivalent storage capacity.
The version in my rendering uses a total of 120 cells, so assuming several years into their use, I needed to replace say 10% of the cells to restore the capacity, I'd be looking at a cost of around $40, since the individual cells are only about $3 each. Compare that to $400 for a comparable deep cycle that's either good or bad as an unserviceable unit, and that's where they really start to look appealing.
You could of course achieve the same capacity with only 4 3.2v 200AH cells at around $110 each with a good deal less complexity, but I designed this model to be highly scalable, for systems ranging from 12.8v (4 cells per string) all the way up to 512v (160 cells per string), so it makes a lot more sense to use the smaller 3.2v 7AH cells in this design, since expanding capacity $480 at a time is a lot more manageable than doing it $17.6k at a time!
= new reply since forum marked as read