A $1.5 Trillion Tip on Cracking Merchant Payments

Mobile network operators (MNOs) have been trying to crack merchant payments in Africa for over five years. It is easy to understand why: Although person-to-person (P2P) payments have proven to be a good initial use case in many markets, merchant payments are a much bigger prize. The global card industry, which is the closest analogue, has grown since 1958 to become a $26 trillion business, with over 11 billion cards in circulation and growth rates of 6 to 16 percent a year in the last two years, including in emerging markets (China Union Pay grew at an astonishing 30 percent from 2014 to 2015). And the IFC estimates that retail payments in Africa are on the order of $1.5 trillion today, a number that is likely to grow with Africa’s rapid population growth and rising prosperity. But MNOs have, for the most part, failed at building a successful merchant payment use case in Africa. Why?

There are many potential culprits, but the most obvious is the ease of using cash. A consumer is highly unlikely to cash in to his mobile money account so he can pay for a small item digitally; he will just hand over the cash. And the merchant probably prefers cash anyway. P2P payments solved the problem of distance. But paying for a good or service from a merchant doesn’t involve distance, so people revert naturally to cash. Until people hold balances on wallets or in bank accounts and have incentives to pay digitally, cash will always win. It is well understood, it’s low cost, and it works perfectly well, especially for low-value transactions. That will always be a pretty tough nut to crack. So paying digitally must be better than paying with cash, especially for consumers.

Have early experiments in merchant payments met that standard? The honest answer is no. Not for merchants and not for consumers. For merchants who are also agents, there is a conflict, as they receive a fee if the customer cashes out, but often pay a fee if the customer pays digitally. Beyond that, digital payments leave the merchant in a position of managing two sets of accounts — cash accounts and e-money accounts. For the merchant who is also an agent, the best choice is to do a cash out and take payment in cash. From a consumer perspective, the process is also not very attractive: It involves the usual long USSD strings, with multiple screens and confirmations. Definitely not better than cash. So the customer interface will also have to get a lot better to entice consumers into wanting to transact digitally. For more on this, see CGAP’s new slide deck, “Digitizing Merchant Payments: What Will It Take?

But there is another challenge at play here: Cash is fully interoperable. Anyone can accept it and use it in the same way with any merchant. In contrast, holders of different mobile wallets all use different processes to pay merchants digitally. One company might use a P2P payment, another NFC, and another a QR code. And the pricing model for merchants varies across different providers. The customer experience and acceptance architecture is different in every single case. That is a recipe for confusion for both merchants and consumers. This is where the mobile industry has important lessons to learn from the card industry.

Mobile money is in many ways in a similar position as the card industry in the 1950s and 60s. BankAmericard was created in 1958 by Bank of America, which was originally limited to banks in California but started licensing banks outside of California in 1965. Not coincidentally, in 1966 Master Charge was introduced by a different set of banks as competition to Bank of America and its partners. During this period, banks basically issued their own cards and acquired their own acceptance networks, much like MNOs do today. As consumer demand to use cards grew, merchants had to juggle multiple acceptance processes, and consumers could not use their cards in the same way everywhere. In the late 1960s, Dee Hock proposed the idea of an association that would bring the banks in the system together into a more coherent scheme. This was done in 1970 with the formation of National BankAmericard Inc (or NBI), which brought the BankAmericard issuing banks in the United States into one system, and in 1974 with a company called IBANCO that managed the international BankAmericard program. Master Charge evolved along similar lines.

These separate entities were consolidated into fully global schemes with the formation of Visa in 1976 and MasterCard in 1979. Both companies were established as associations to build out a common platform for managing payments across different member banks’ systems. Eventually, the need for a common user experience, brand and rules became evident, and these organizations evolved to provide the consistent and reliable experience for both merchants and consumers that we see today. The result is that I can take my Visa or MasterCard card anywhere in the world, and it works the same way. I know roughly what a transaction will cost me and I know that fraud will probably be flagged and brought to my attention. I also know that my bank will cover the cost of fraudulent transactions under certain agreed conditions and that there are agreed dispute processes in place to resolve problems arising from any transactions I perform.

MNOs have experimented with interoperability in many markets, Tanzania being a well-known example. But that early experiment was limited to P2P transfers. Operators are aware that they need interoperability to make merchant transactions work — because they need that same seamless experience that exists in the card space, both for merchants and for consumers. And that is not simply a matter of interconnecting so that transactions flow across different networks. In the same way as the card companies have done over the last 60 years, they will need to build interconnection, common processes and a common acceptance brand if they aspire to have merchant payments work in Africa in the same way as they do in the card industry. Merchants and consumers need to feel that there is a benefit for them to paying with mobile money, and that is hard to do if it requires more energy than simply using cash. The relatively low value of transaction sizes in Africa is a challenge, but this is a long game, and MNOs are perfectly positioned to build merchant payments networks if they are patient and understand the need to work together.

The interesting question is whether these fiercely competitive companies will come together to build the collaborative infrastructure required to make this work. Multiple closed loop systems will always result in sub-optimal outcomes, so if MNOs want to build the kind of market we have seen emerge in the card space over the last 60 years, they will need to move in this direction. It will take time for this to evolve, but there is much to learn from the card industry, and even a fraction of a $1.5 trillion market is probably ultimately worth the effort.


Reading Deck

Digital Financial Services providers increasingly sense a key opportunity in digital retail transactions, or merchant payments. This deck gives a high-level view of what CGAP sees as some of the key challenges in merchant payments and outlines idea to overcome them.
Sub-topics: Payments


11 September 2017 Submitted by Christopher Williams (not verified)

The report is very well done and identifies the change in the market .. the blockchain structure is moving to free processing for merchants where digital cash and/ or cryptocurrencies are the flows. I agree the card companies might still get involved, but the margins are lower and more suited to data management entities.
Overall, expect remittances to come down in price to the SDG target of 3% be early 2018 - and merchants to be charged zero while basing their structure on the ad savings.- that is in the developing world .. which may be the new leader in transaction structuring for the future.

25 September 2017 Submitted by Greta (not verified)

Remittances are already coming down in many markets, as MNOs move into the cross-border space. Technology is providing many interesting new solutions. But I agree with the previous comment that the merchant and the customer are at the heart of this, and they will have to make the decisions to move towards cashless. It will be about getting the incentives right for both. That is what this series of research pieces is about. This is already happening fast in Asia. What will it take to move towards low cost merchant payments in Africa???

11 September 2017 Submitted by Talitha Bulaya (not verified)

Sitting at the core is the consumer and the merchant. It has to be convenient for both for cash to be eradicated and mobile money to reign. Much as Visa and MasterCard are successful, in Africa cash is still king. Therefore the card industry and the MNO's have have a common problem, cash.
It all boils down to education and making the consumer and the merchant aware of the benefits. A change of mindset is what will make cashless society in Africa possible.

12 September 2017 Submitted by Christopher Williams (not verified)

This is a very good piece and asks a lot of the correct questions. I also found the PPT to be very helpful.

My comment would be that while card companies could be partners in these endeavors, not unnaturally, they do tend to want to protect their business model with its high fees and costly infrastructure. They are very good at covering every possible type of transaction (for a price) but is it necessary to make it likely for someone from Latvia to go to a merchant in Zambia or vice versa?
Blockchain technology can operate at much lower costs and at far lower risks and delays for MNOs and cryptocurrency exchanges, which opens up the opportunity of lowering costs to the extent where merchants pay no interchange at all. This changes the conversation considerably; now it becomes a question of "why not?" accept these payments.
The other pricing impact is on remittances; the current level of around 8% is still woefully far away from the SDG target of 3% - but that can change. We expect to see trades going through at the target level within the next six months.
So, the future for payments is bright, particularly for those who have paid too much for too long and have the greatest need.

19 September 2017 Submitted by Kamaljeet Rastogi (not verified)

Fully agree that Merchant Payments are a a huge opportunity for Mobile money/Wallet operators. NFC based solution for mobile money operators is solving this requirement in a very good way. Small card sized POS device is given to the merchants and a NFC card to the consumer - merchant payments are just a quick Tap - faster and better than Cash - Refer Airtel Tap Tap service in Tanzania

25 September 2017 Submitted by Greta (not verified)

Yes, Kamaljeet. Ease of use is going to be very important. And will hopefully get even better as smart phones come on line. The QR code seems to be the method of choice in much of Asia. It will be interesting to see how this evolves in Africa.

02 November 2017 Submitted by Kevin Brown (not verified)

Hence the importance of someone running commercially with the concept of Mojaloop ( Gates Foundation proof of concept). As it stands however the code released is no where near production ready and also, too many MFS platforms are not very open with their API to allow integration to Visa type switch situations. QR codes are great but Africa needs more smartphone penetration. Countries such as Myanmar with a huge base of smartphones needs to be examined as a model to try to push African countries into more smart phones.

Also, the MFS platform offer very little back office support for small merchants. This needs to change.

All the open API standards forget the out flows/push of required merchant data to allow the merchant to manage MFS payment systems.

02 November 2017 Submitted by Kevin Brown (not verified)

" if they are patient and understand the need to work together"

In the MFS space, the issue has always been Telco reluctance to invest in service level API's. The simple phrase of a rising tide floats all the boats is always lost in telco space ( outside of network and phone GSMA type standards).

It seems an obvious step for Telcos involved in MFS to adopt something like the GSMA MM API, but it is painfully slow. Fear.

Add new comment