@OP: For starters, I agree with your overall analysis of the market penetration of SMS vs. data/internet. The developing world has roughly 22% internet penetration, but over 70% SMS penetration. Bitcoin over SMS could be absolutely huge.
No, onto your specific points below:
Their SMS service is available in the U.S. and Canada only, though they assert that they are working on offering that to a wider global footprint. Presumably, since Coinapult offers the ability to send bitcoins for free (they only charge when bitcoins are sent through their API), offering this globally would incur costs for sending the text messages.
We are struggling to find a combination of cost-effectiveness and reach. We can currently send SMS anywhere, but doing so is cost prohibitive. Even in the U.S. we average more than $0.06 per send. That is, each SMS message costs us $0.01, and a typical send requires at minimum 5 messages:
1. Sender requests coinapult send funds
2. Coinapult responds with conf code
3. Sender confirms
4. Coinapult responds with success message and new balance
5. Coinapult messages recipient
Tag in a couple of information texts (i.e. balance checks, help, transaction checks, etc) and the average cost goes up further. With international SMS fees, the per message cost is easily triple. This would bring the cost of a single send transaction to roughly $0.20. Now, there are ways to bring these costs down, but only with huge volume. For instance, a dedicated short code (intended for SMS use) costs $1k/month.
In short, the main barrier to international expansion is cost effectiveness. Even in the U.S., it is hard to justify offering such a service for free. In an international setting, the costs increase, while it is likely the average transaction size would decrease. If we can work out a viable fee structure, it can be deployed almost immediately.
Coinapult requires a text message response from the mobile number the funds were sent to so at least with that there's no bearer code being transmitted across the SMS network.
We consider this basic security, but is ultimately inadequate. At the moment, the initial 'receive' message still carries a bearer code, which is vulnerable in the same way blockchain's are. We hope to release an account system which will allow us to stop sending these by the end of July.
How many days after realizing that SMS wallets are starting to catch on before texts to and from the SMS wallet provider's number are blocked? (and thus the user's funds left stranded, at least temporarily). Unlike bitcoin's peer-to-peer architecture, telecom infrastructure operates as the givernment instructs them to.
This problem may be a little overstated. Governments don't always act so efficiently, and they'd have a tough time blocking us in such a way. Assuming we go international through a series of phone numbers in different service areas, it would be quite easy to get new numbers for any given service area. The local gov. and mobile service provider would be playing whack a mole. What's more, I'm sure the people would have something to say about it.