Getting a tx to be much smaller than the current minimum size would be no small feat although I do understand that a reduction of maybe 30% for the signature size might be possible after SegWit has been implemented.

But I think that is more likely going to be 200 bytes than 100 (and in general over time crypto stuff tends to need more bytes in order to remain secure so you can't really expect any further significant size reductions).

Personally I think get the settlement side perfected first and then work on the payment side of things (there are already quite a lot of off-chain services to do payments as it is so this isn't really so urgent IMO).

Understand that the settlement side of things includes the remittance market which is actually huge (it is the first big market that Bitcoin should actually have already disrupted yet it has failed to make any significant inroads into so far).