[From Financial Access Initiative blog, 6 September 2011]
A lot of today’s research is focusing on tweaks to financial contracts and marketing with the aim to improve take-up and impact. The grail is big gains generated by small changes.
But big impacts often require much more than tweaks. That’s especially true for mobile money, in which scale and interconnectedness really matter.
Mobile money systems seek to create an ecosystem within which money is passed around and stored in electronic form. It’s hard to get such an ecosystem going, but M-PESA in Kenya shows us how once it gets big enough it can become a powerful snowball. Critical mass thresholds are associated with two types of transactions that are particularly problematic.
The last mile
The first one is the conversion of cash into electronic value, and vice versa. Users would like to resolve this inside their pocket since that’s where they keep both their cash and their mobile phone, but unfortunately these two things don’t talk to each other. (See my hopelessly unrealistic vision of smart banknotes which can get turned on and off from a mobile phone.) This is really a last inch problem, but we turn it into a last mile problem by engaging shops to act as cash merchants.
This last mile problem is tractable, but solving it satisfactorily requires assembling a very large transaction base. First you need a sufficient number of stores so that the majority of users have a cash merchant reasonably nearby (we are not calling this the last mile for nothing). Second, each of these stores needs to see enough volume of commissionable transactions each day (I use 50 as a rule of thumb) to justify the cost of holding sufficient liquidity to meet customer needs, the prime signage positioning, the enhanced risk of robbery, and the ongoing staff training incurred.
Cash merchants offer backward compatibility of mobile money with the legacy payment system: cash. Having a well-functioning cash merchant network is essential, especially in the early phase of development of a mobile money system, because it offers a path for the majority of the population which is unbanked and lives entirely on a cash basis to get into the ecosystem. It also confers trust in the system in so far as people feel like they can convert electronic value back into good old cash at any point in time.
As mobile money systems address the cash transaction problem, they
start growing mainly on the basis of remote transactions (remittances, bill payments and the like) and larger
transactions (for sporadic rather than daily purchase and sale of goods and services), for which cash is
less well suited. Mobile money and cash both compete and complement each other, and each quickly finds
its niche. But as long as people use mobile money primarily for larger, monthly payments, usage
will remain limited to a couple transactions per month per customer, on average. The growth of the
more successful mobile money systems has indeed been driven by growth in customers, not so much by
growth in usage per customer. And then, at some point, saturation kicks in.
To get to the next level, mobile money systems will need to address the most problematic transaction of all: payment for everyday goods at the corner shop. I call this the last yard problem: that’s the width of the counter between buyer and seller at the duka, the sari sari store or the bodega. This transaction is problematic because cash performs well in this environment: it is fast, convenient and free (for small enough transactions, as long as you are long cash at least). Here you are taking on cash in its home court.
Yet capturing in-store payments would help entrench mobile systems in several ways. First, it opens up a much larger transactional pool – from 1-2 per user per month to 1-2 per day. Second, by making mobile money directly useful in everyday life, it incentivizes people to keep their money in electronic format rather than cashing out the moment they receive a transfer. This reduces the end-to-end cost of transactions and makes it more likely for people to reach for their phone when they need to pay for something. Third, users’ higher frequency transactional data may one day become useful to conduct credit scoring and grant credit on demand, such that the phone becomes more like a credit card.
Let’s just accept that in-store payments–unlike remote payments—will need to be very cheap. But that won’t be enough: you can only solve the last yard problem by creating incentives for the buyer to want to pay electronically, and for the seller to want to receive money electronically. The allies on both sides will be the business users behind them: employers (whether formal or informal) paying wages electronically, and manufacturers and distributors rolling out electronic payments across their supply chains. Capturing the business payments space will be the key battleground, and rolling out a suite of business APIs (application programming interfaces) and engaging communities of solution developers will be key instruments.
Cracking the last yard problem requires taking a full ecosystem view, and that requires immense breadth and scale. Customers need to want to pay at any store with their electronic money, and stores need to want to receive electronic money from any electronic source. Fragmented payment solutions will not fly.Which brings me back to my theme for 2011: mobile money will not work in most countries unless the various players agree to interconnect. (I laid out my mantras in an article in the 2011 MMU Annual Report.) Few players have the market reach and capacity to solve the last mile and last yard problems on their own. Yes, interconnecting is messier and a lot less fun than if you could do it on your own, but the point is: most mobile operators cannot do it on their own. Do just like you do in your core business: interconnect.