A conversation I recently had with a taxi driver in Kenya—a country where more than half the adult population uses Safaricom’s M-PESA—illustrates why talk of the death of banks is unfounded. He explained that M-PESA is one of a portfolio of financial tools that he uses to manage his money, and that his bank is an indispensible part of that portfolio. In fact, I’ve come to believe that mobile money services can increase, rather than dampen, demand for traditional banking services.

John explains one reason why he still needs a bank: he prefers not to leave large balances in his M-PESA account because a criminal could coerce him into transferring money into his account. It may be that this risk is higher for taxi drivers, who come into contact with all kinds of characters, than for ordinary users. But he makes a more prosaic point, relevant to just about any saver: you can’t earn interest with M-PESA. John takes advantage of the linkages that Safaricom has created with a number of banks in Kenya, allowing him to sweep money from his M-PESA wallet into his bank account when his balance becomes large enough.

For John, M-PESA is a complement, not a replacement, for the bank. In fact, you can plausibly argue that M-PESA is increasing, not decreasing, the demand for bank accounts in Kenya. That’s how I interpret the extraordinary popularity of M-KESHO, an interest-bearing savings account linked to M-PESA which 176,000 people have started using in the last 3 months. By tearing down some of the barriers to bank usage (long queues, inconveniently located branches, minimum-balance requirements, etc.), M-PESA makes banking attractive to a whole new segment of consumers. And that’s something for customers, operators, and banks alike to celebrate.

Let’s imagine that it costs US$2 to register a new customer. That cost is typically made up of a commission paid to the registering agent, the kit (sometimes including a new SIM) that is provided to the user, and any costs in the back office associated with creating the account and fulfilling KYC requirements. If an operator has 500,000 users on the books that aren’t using the service, that equals US$1 million in costs with no associated revenues. And if the operator pays a per-user fee to its technology provider, either on a one-off or ongoing basis, the situation becomes even more dire.

You may be asking yourself why a customer would register for a service he has no intention of using. Well, if it’s free to register, and when commissions to the registering agent are significant, the more relevant question may be, “Why not?” When agents can earn a quick buck (or some fraction thereof) signing up new users, they may decide that that’s more lucrative than facilitating transactions, and focus on the former rather than the latter. (Agents whose only responsibility is signing up new customers don’t even have this trade-off to make; so long as their compensation is tied to registration itself, they’ll register anyone for the service, regardless of whether than person has any demonstrated need for mobile money.)

There are two bits of good news. The first is that it’s relatively simple for operators to start filling in this hole, and it doesn’t need to be expensive. The first step is to put the brakes on acquisition by fixing registration commissions: agents need to be paid for registering customers, but they should see more value in signing up a user from whom they will earn a stream of cash in and cash out commissions than in the single act of registration. The second step is to take the money that’s being saved on customer registration and pour it into marketing activities that are designed to drive usage, particularly among users who’ve already registered. As long as unregistered customers can still go to an agent and register themselves if they are inspired by these marketing efforts to try the service, operators should see no reduction in transaction volumes when scaling back their customer acquisition efforts.

The second bit of good news is that we see a number of operators moving in this direction. For example, a number of operators have moved to registration commissions that are contingent on users making one or more transactions. This might mean that the numbers of registered users—i.e., the ones that capture the headlines—start to grow more slowly. But it will also mean bringing the industry closer to long-term financial sustainability, which, although it might not make headlines, will be something to celebrate.

For more about agent commissions, check out our recent report on incentivising mobile money agents. And stay tuned to this blog for our forthcoming look at the profitability of mobile money.

In drafting these definitions, we have sought to reflect the most common usage of these terms, rather than advocating for meanings that we prefer. Therefore, in a few cases, we have listed multiple definitions for terms that are commonly used to mean more than one thing.  We will be embedding these definitions into the fabric of this blog: you will be able to hover over any term that is underlined twice to see our definition. There is also a printable PDF version of the definitions below.

Finally, you may find as you peruse these definitions that you think you can improve upon our efforts: you may do so at the Mobile Money Information Exchange, where a wiki has been set up to allow users to make modifications to the definitions and to add new ones.

We look forward to hearing your feedback, either on the wiki or in the comments section of this post.

Mobile Money Definitions_no marks_

In this article, we argue that serving as a mobile money agent makes a number of unusual demands on retailers: rather than selling a product, like fast-moving consumer goods or even airtime, mobile money is a service, offered in accordance with a set of processes that agents have to learn and adhere to. To ensure that agents are fulfilling their obligations so that customers, in turn, enjoy a positive experience with mobile money, it’s essential that operators have a way to support their agent network on an ongoing basis. In this article, we discuss what support agents need and the various mechanisms that operators can employ to provide this support.

This article joins “Building a Network of Mobile Money Agents” and “Incentivising a Network of Mobile Money Agents” and completes our series on mobile money distribution. All three sections, plus an introduction, are now available online, and they will be published in hard copy as part of the Mobile Money for the unbanked 2010 Annual Report and distributed at Mobile Money Summit in Rio de Janeiro.

Managing a MMANetwork NEW

Please click below for the full report on Agent Networks

Agent Networks Manual

In every country, mechanisms for bill payment, money transfer, and storing money have existed long before the advent of mobile money. When mobile money services are introduced, customers compare the new option with existing alternatives. There are five key attributes that customers consider when evaluating a new money transfer service: its convenience, speed, safety/reliability, ease of use, and cost. Importantly, customers rank these attributes in different ways: some care a lot about one feature, others about another.

Operators need to be sure that their money transfer service will be superior to existing options in the marketplace along at least one of these dimensions: it should be more convenient, or cheaper, or faster, or more reliable, or easier to use. A new offering doesn’t need to beat the competition along all of these dimensions in order to be successful: given that not all customers value the same thing, simply offering a service that is superior along one dimension will inspire customers for whom that feature is most important to switch to mobile money. And indeed, since the design process for any service usually involves making tradeoffs between these attributes, companies usually have to pick and choose which attributes they will optimize and which they won’t.

Nevertheless, the amazing thing about M-PESA in Kenya is that Safaricom managed to design a money transfer service that outshone the competition along all five of the relevant dimensions. Before M-PESA was launched, Kenyans employed a variety of mechanisms for sending money: bus companies and matatu drivers, the post office, banks, and traditional money transfer houses all offer money transfer services, and many people would simply ask a friend or family member to carry cash by hand. But in a survey of M-PESA users in September 2008,

* 98% said M-PESA was faster than other money transfer services

* 96% said M-PESA was more convenient than other money transfer services

* 98% said M-PESA was safer than other money transfer services

* 96% said M-PESA was cheaper than other money transfer services

* 99% said M-PESA was easier to use than other money transfer services

I would argue this helps to explain M-PESA’s appeal to such a large swath of Kenyan society (40% of adults, at last count). Safaricom was able to win customers regardless of which attributes they prioritized, because M-PESA outshone the competition along every dimension – in part because the service was so good, but equally because the competitive alternatives were so weak.

What does this mean for operators in other markets? Three things, I think:

1. Operators must analyse existing alternatives to any potential service (whether it is p2p payments or something else) they might offer by mobile money. They need to know exactly how reliable they are, what they cost, how convenient they are, how long they take, and how easy they are to use.

2. Operators need to realistically assess whether they can offer a service that will beat the competition in all of the key attribute areas. Our experience suggests that it will not always be possible to beat all competitors along every dimension as Safaricom did in Kenya: in some markets, pre-existing money transfer options are better than they were in Kenya. But this need not deter operators. What’s important is to design a service that does beat the competition on the attributes that will resonate with target customers.

If, say, the national postal service offers a popular money transfer service in its branches that is cheap—i.e., so cheap that a mobile operator would be unable to match the price while still turning a profit—but slow and inconvenient, operators should try to find a service design that optimizes for speed (which is one of the natural features of mobile money transfer) and convenience (by focusing on a large agent network).

3. Operators need to understand the preferences of their target customers. In the scenario I describe above, it would be important for an operator to make sure that there is a critical mass of consumers who would value speed and convenience enough to pay a bit more before going to market—lest they be unhappily surprised to find no such willingness to pay.

The last thing to note is that this methodology applies as well to bill payments and other mobile money services as it does to money transfer. For any service that can be offered on a mobile money platform, there will be competition—and understanding that competition is the first step to beating it.

Over the last several years, distribution has proven one of the most difficult pieces of the mobile money business to get right. Even with the accumulated experience of distributing airtime in their markets-often through thousands of retailers-mobile network operators find the process of building a network of mobile money agents, incentivising them appropriately, and managing them to be fraught with difficulties. Indeed, from keeping agents liquid to deterring fraud, operators consistently rank issues related to agent networks as among their most vexing.

In response, the Mobile Money for the unbanked Team at the GSMA is today pleased to release the first instalment of a handbook for mobile network operators on building, incentivising, and managing networks of mobile money agents. This report is the product of six months of research into and analysis of the ways that mobile network operators and other providers of mobile money services approach distribution. It is designed to serve as a reference not only for operators that are just beginning to consider their distribution strategy, but also to those that are interested in benchmarking their approach against other alternatives.

Today, we are releasing an introduction to this handbook and the first section, entitled “Building a Network of Mobile Money Agents”. Between now and Mobile Money Summit, which is being held May 24­-27 in Rio de Janeiro, the other two sections of the report will appear on this blog: “Incentivising Mobile Money Agents” and “Managing a Mobile Money Agent Network.” The complete handbook will de distributed in hard copy in Rio de Janeiro in May, accompanied by an article on agent network regulation.

Each section of this report is designed as a series of questions and answers, to make it as easy as possible to navigate. In the first section, we tackle basic questions like, “What do agents do?” and “How are agents recruited?” as well as more difficult ones, like “How big should an agent network be?” Among other topics, we discuss the pros and cons of using field registration agents; the optimum rate of agent network growth; and the benefits of an aggregator/masteragent tier in mobile money distribution.

This report is the product of conversations with dozens of industry participants, and we are grateful to everyone who has shared their experiences with us. At the same time, we recognize that the mobile money industry is still young, and that all of the conclusions we draw in this report are at best provisional. So we look forward to hearing your feedback about this and future instalments in the comments section of this blog.

For the next chapter on how to incentivise a network of Mobile Money Agents,  please follow the following link:

Yet a number of mobile money platforms that mainly offer P2P transfers have struggled to attract users over the course of the last year. Of course, it’s difficult to isolate why this has happened-mobile money is a complex business, and lots of things can go wrong. But one possible explanation is that Kenya, with its tightly knit families and its high degree of labour mobility, is atypical in the volume of remittances that flow within its borders.

We are starting to see some evidence for this. As part of a Bill and Melinda Gates Foundation-funded project called AudienceScapes, Intermedia has recently undertaken nationally representative surveys to measure media use, ICT use and personal communication habits in Kenya and Ghana. They found that 42% of respondents in Kenya had used a money transfer service in the last year, versus just 14% in Ghana. Some of this discrepancy might be explained by the fact that M-PESA has allowed people who previously didn’t use what they would consider a formal money transfer service do so in the past year, but given the rate of penetration of M-PESA in Kenya, this effect cannot explain the entire difference. It is likely, rather, than there is simply more demand for money transfer in Kenya than Ghana.

So what does this mean? First, that operators who offer mobile money services need to be rigorous when assessing the demand for mobile money services in their market. There is no substitute for qualitative and quantitative market research-relying on anecdotal evidence, or drawing inferences from other markets, is fraught with risk. The GSMA has developed a toolkit to help operators assess the demand for mobile money services in their markets, and they can contact mmu@gsm.org to access it.

Second-and this is good news-operators have begun to explore more seriously other services that users might value on the mobile money platform. One example is bill payment. Telenor Pakistan enjoyed tremendous success bringing to market bill-payment services before offering P2P transfers. Nadeem Hussain, CEO and President of Tameer Bank (which is 51% owned by Telenor Pakistan), told a group at Consult Hyperion’s Digital Money Forum in London last week that easypaisa is on track to be processing bills by the millions by the end of the year. This is in part because for consumers in Pakistan, paying bills is a big pain point, just as sending money home was for consumers in Kenya. The option to pay a bill at a local agent is often preferred, by a wide margin, to doing so at the bank or other existing bill payment outlets: for example, Hussain reported that 40% of bills processed by easypaisa are paid outside normal business hours, which is impossible at traditional bill-payment outlets like banks.

The lesson of easypaisa (and Grameenphone’s BillPay and True Money) is not that mobile bill payment is the ‘next big thing’; rather, it’s that successful mobile money platforms are the ones that tailor their service offerings to their local market. This requires an extensive effort to understand customer needs; it also requires assessing which of the host of financial services that could potentially be offered on a mobile money platform (in addition to bill payment, operators are experimenting with international remittances, disbursements and repayments of microloans, microsavings, microcredit, insurance, and more) can be offered as part of a sustainable business model-which in turn requires a thorough understanding of substitutes that already exist in a given market.

If 2009 was the year of the proliferation of mobile money services beyond P2P transfers, we hope that 2010 will be the year that many more operators are able to align their services offerings with the needs of the low-income consumers they serve-setting the stage for more M-PESA-scale success stories in the months ahead.

1. Distributing cash and electronic value using MFI infrastructure: In many markets, mobile money service providers are appointing the branches of microfinance institutions as cash-in/cash-out agents. There are real synergies here: MFI branches are typically more secure that other retail outlets, are staffed with people who are trained in cash handling, and maintain float in the course of doing business-plus, it’s faster for operators to sign up one MFI with many branches than many independent shops. In addition to a new revenue stream, the MFI gains foot traffic and the opportunity to cross-sell mobile money customers on microfinance.

2. Disbursing microloans and accepting repayments on mobile payment systems: In other cases, microfinance institutions are discovering that they can lower their costs by disbursing loans and/or receiving repayments via mobile money, eliminating some of the administrative costs that they incur in the field and the back office. Some of these cost savings can be passed on to the customer; in Afghanistan, borrowers from First MicroFinanceBank who agree to receive and repay their loans using M-Paisa qualify for a reduced interest rate. But customers enjoy other benefits, too, which may actually be more important to them: it is safer to take out and repay microloans in mobile money rather than cash, and it saves time. For the provider of mobile money service, partnering with an MFI with a large number of borrowers in this way can be a way to rapidly sign up new customers and drive usage of mobile money. However, it is a significant operational challenge to integrate the systems

3. Meeting mobile money agents’ demand for capital with microfinance: Often, new mobile money agents have to raise capital to invest in cash and/or electronic value float. One way to get it is to take out a microloan. From the perspective of the MFI, this kind of loan is relatively low risk, since the principal is preserved in highly liquid form and repayments are made from the commission stream earned by the agent. I suspect that many mobile money agents around the world have financed their cash and electronic value float with microcredit. But more formal loan products could be tailored to exploit this opportunity. In one country, a mobile operator and MFI have teamed up to help microentrepreneurs open small stores that sell airtime: The MFI provides upfront capital as well as a revolving line of credit for times when the entrepreneur runs short of cash or airtime. It’s interesting to imagine a loan facility that addresses the particular liquidity requirements and cash flows of mobile money agents; if mobile money turns out to be as successful in other parts of the world as it is in Kenya, there could be significant demand for such an offering.

Thanks to Cameron Goldie Goldie-Scot at Musoni for discussing the content of this post.

But what if you did have a suspicion that fax machines were going to become hugely popular, and that they would become indispensible tools for doing business through the early 2000s (when they started to be displaced by the wide availability of scanners, printers, and internet access)? Quite naturally, you would then be more interested in owning one.

This introduces an important wrinkle to the definition of network effects that I gave earlier. In networks with positive network effects, potential users’ interest in joining the network increases not with the size of the network, but rather with their perception of the size of the network and their expectations about its size in the future.

How can operators influence the perception of the size of their network? Well for one thing, they can share their customer adoption numbers, assuming they are good. Safaricom made regular announcements in the two years after M-PESA’s launch, and these updates were widely reported in the Kenyan press. Another tactic is to encourage, or even require, agents to advertise for the network they represent. Mobile network operators do an amazing job in many emerging markets achieving near-ubiquity for their brands-doing so for their mobile money offerings suggests to potential users that the service is widely used.

How can operators influence expectations about the future size of their network? (This will be particularly relevant when the actual, to-date uptake figures are not as large as hoped.) One technique is to make large investments to signal commitment. In the United States, each year advertisers spend millions of dollars to buy 30-second spots on the broadcast of the Superbowl, the most-watched sporting event in the country. What’s interesting is that in addition to the Coca-Colas of the world, which you would expect to have the massive marketing budgets to support such advertising, you’ll often see ads placed by small internet start-ups. Why? Well, many of those start-ups are platform-mediated networks, the value of which is a function of the size of their user bases-or, more specifically, expectations about the future size of their user bases. By spending big for a high-profile ad, those start-ups signal their commitment to do whatever it takes to develop the market.

The takeaways for operators are simple: to exploit network effects to your advantage, use your marketing and communication messages to suggest that the number of users in your network is large and signal that you are committed to developing the market for mobile money.

The good news is that by carefully crafting a value proposition and a marketing and communications plan, mobile money service providers can dramatically increase the odds that they will build a successful mobile money platform. They can do this with a nuanced understanding of the way that network effects in mobile money work.

Let’s go back to our original definition of network effects. A network that exhibits positive network effects is one in which potential users become more and more interested in participating in a network as the number of users in that network grows. But as anyone who has ever used mobile money knows, not all users are the same. If you are a widow living in rural Kenya who relies on your son to send you home money each month, millions of people can sign up for M-PESA without your interest in the service increasing one bit. Only when your son signs up does your interest perk up dramatically.

The market for fax machines is in this way very similar to that of mobile money. Each time a new fax machine was installed, the value of buying a new fax machine for any given user went up almost infinitesimally (since odds were that the new user and the potential user didn’t know each other and would never need to interact with each other). But when someone with whom you already communicated with bought a fax machine, your interest in owning one did indeed increase.

The successful fax machine manufactures grasped this point and shaped their go-to-market strategies accordingly. They focused their initial sales efforts on diplomatic corps, multinational corporations, and major newspapers. Why? Because those organizations had consulates, offices, and bureaus (respectively) scattered all around the world with whom they needed to communicate. By selling the head office of a multinational firm a fax machine, the value of a fax machine to their subsidiaries’ offices increased dramatically.

What does this mean for mobile money? When going to market, it’s important that operators identify the participants that, once they’ve signed up to the service, will attract others to join, too. For example, Telenor, which has recently launched easypaisa in Pakistan with an emphasis on bill payment, lined up some of the biggest billers in the country to participate in the platform. In the model of a platform-mediated network, those big utility companies are “users” that many potential users are going to want to transact with.