Decentralised Mobile Payments for the poor
This post is an archive from May 4, 2016.
This post just details my thoughts and plans for creating a mobile payment system for the poor in India. The photo above is from a hackathon which Barclays conducted in November 2015, which was the first time I got introduced to the world of mobile banking.
The Situation
Mobile payments are something that is very hot in India right now. In a country where most of the population falls in the unbanked and underbanked sector, it’s not a puzzle why banks would want to figure out ways to tap into this audience.
India has already seen the use of systems like m-Pesa (owned by Vodafone) and PayTm, which try their best to bring mobile money to everyone.
While PayTm is the biggest mobile wallet in India today, it is used by people who have access to a smartphone. On the other hand, while mPesa is the biggest USSD mobile wallet in countries like Kenya, it is not that widely used in India because of the inability to send money to numbers on other networks other than Vodafone (cross-network payment).
Why don’t mobile apps work?
Majority of Indians are still under the poverty line and hence it’s not possible for them to spend as much 1.5K Rs on a smartphone when their monthly household income comes to just 10k Rs. There is, however a wide use of brick phones, which are far more cheaper than smartphones. However, brick phones don’t have apps hence mobile solutions like PayTm cannot be used on these devices.
But for many households, getting a smartphone is a luxury. And while the prices for a smartphone are becoming cheaper and cheaper (See Freedom 251), you do have people who are happy with spending a little bit more because they treat a smartphone as a long-term investment.
Still, as India is slowly improving technologically; It will be a long while before we see smartphones everywhere.
Why current USSD services won’t work
If you’ve ever been to Kenya then you’ve definitely heard of mPesa. mPesa is payment system where you type in a USSD code (*400#) and press call and a simple menu comes about. The menu has a few options for you to check your m-Pesa balance and send money from one number to another.
It’s a simple and beautiful system which allows people to use mobile money with a brick phone. It’s fast, simple and widely used in Kenya.
What about India? Will USSD money work in India?
m-Pesa unfortunately, is not that widely used in India. It works in countries like Kenya because the company that runs m-Pesa, Safaricom is a monopoly in the Kenyan telecom space. m-Pesa had the first-mover advantage in Kenya.
In India (and in other countries), mobile providers soon became aware of this concept and have started replicating the system on their own.
Which is why now you have along with Vodafone’s “mPesa”, Airtel’s “Airtel Money”, Reliance’s “Reliance money” etc. all being offered to the consumer with the same features. mPesa (in fact, none of the mobile wallets) have been successful in going mainstream in India.
What’s wrong with USSD?
USSD is primarily restricted to the network carrier it is offered on. Which means that if you can type in *400# on your Vodafone SIM to access mPesa, you won’t be able to do the same if you had an Airtel SIM.
USSD is great since it is fast, simple and nearly all phones (including bricks) can run USSD codes. But the competition and the challenges that mobile networks providers face make it nearly impossible for USSD money to become mainstream in India.
Some ideas to tackle this
So I’ll note down some good ideas that have come about over, that basically attempt to solve the problem.
1. Using SMS for mobile money
There is however a third option (besides apps and USSD), which also has it’s pros and cons. A common service that is used by everyone, SMS (Short messaging service). If you haven’t realized, SMS and phone calls are some of the few services that can talk across different networks. That is, a person with an Airtel sim can send an SMS to a person with a Vodafone sim.
Which means that there might be some way for us to create a mobile system using SMS as a medium, right? It’s pretty easy to do so in fact. During Barclay’s hackathon (photo above), My team and I made a prototype that could transfer money from one number to another using one simple SMS command.
But SMS is really slow. SMS takes on average 5–10 seconds to send and receive. Which means a typical transaction could take well about 10–20 seconds (A SMS going to us and a confirmation back to the user). Also SMS in the end is tied down to the carrier.
2. Having a joint collaboration with the banks and telecom providers
You could have a collaboration to create an entity wherein every bank and telecom provider has a certain fair share and they all could work collaboratively. If that is even possible, then the telecom providers would all agree to integrate a common system and that then everyone could send/receive mobile money via USSD to anyone in India.
This won’t really happen, because no-one would be interested unless it is forced as a government initiative. Which is why NCPI was able to create a unified interface for anyone to send money from their bank accounts through a simple USSD interface. They call it the NCPI Unified USSD platform.
NCPI lets you transfer money from one bank account to another, by having you punch in your account number & IFSC code through the USSD menu (good luck remembering that).
The problem here is that NCPI forces the user to have a bank account. And while there are schemes like the Jan Dhan Yojana which has opened 200m+ bank accounts for the poor (a large portion of it, still being dormant); many people associate that a bank account is too expensive.
And typing in the bank account number makes it absolutely irritating when it comes to simple things like paying your milkman.
3. Use IVR and make it accessible via a toll-free number
With the help of IVR, we could still create an automated system and the fact that we could have a toll-free number would imply that anyone in the region could access it for free.
The only issue here is that having toll-free numbers are usually very expensive and again is maintained (hence restricted) by a network provider. However, if such a system is legally approved (by the RBI) then the network provider would have no rights to take down the system (because you could legally sue them and win).
4. Start your own telecom service
Being your own telecom provider would allow you to send SMSs at lightning fast speed (Because you are the SMS server). A typical transaction can go down to anywhere between 5–10 seconds (instead of 20–30). You could even run your own toll-free number.
But in India, to legally start your own telecom service, you would need to reserve a frequency band which unfortunately would cost you in the billions.
5. Have a decentralised system using SMS
Having a decentralised system allows it be completely independent of any outside influence and it will be very hard to be taken down. While SMS might be slow, it is probably the only thing that works.
Such a system would raise a lot of regulatory concerns and it also might end up being rejected by many people (especially merchants) if it turns out to be illegal.
I have yet to think of how such a system would work.
What I’m trying to do
I believe that creating a decentralised system is the best way to go for something like this. The main issue I haven’t solved (which is a very core one) is to figure out how the accounts for every number & transaction would be distributed. How can we create a system where there has a good ledgering mechanism (much like the blockchain), but at the same time being very simple for anyone to use?
There are a lot more challenges to this, but I’ll leave it for the reader to ponder over.
Creating a decentralised system would make the most sense as it cannot be controlled or taken down by any network provider. If thought about correctly, it would at least be something that would survive the obvious pressure from the banks, who would not like anything that dominates the market except for themselves.
