artice

Mas:JSC page.qxd 12/01/2011 13:42 Page 352 Journal of Payments Strategy & Systems Volume 4 Number 4 Three keys to M-PESA’s success: Branding, channel management and pricing Ignacio Mas and Amolo Ng’weno Received (in revised form): 27th July, 2010 Bill & Melinda Gates Foundation, PO Box 23350, Seattle, WA 98102, USA. Tel: +1 206 7702120; e-mail: [email protected] Ignacio Mas is Deputy Director in the Financial Services for the Poor Program at the Bill & Melinda Gates Foundation, where he leads on research and policy. Ignacio has been a Senior Adviser in the Technology Program at CGAP (a resource centre for microfinance housed at the World Bank), VP of Marketing and Account Management at interTouch (an NTT-DoCoMo Group Company which manages broadband solutions for the hospitality industry), Director of Global Business Strategy at Vodafone Group, and Senior Manager responsible for telecoms investments in Europe at Intel Capital (Intel Corp’s venture capital arm). Ignacio has been a Visiting Professor of International Business at the Graduate School of Business at the University of Chicago. He holds undergraduate degrees in mathematics and economics from MIT and a PhD in economics from Harvard University. Amolo Ng’weno is Deputy Director in the Financial Services for the Poor programme at the Bill & Melinda Gates Foundation, where she leads on product design and development. Amolo was a co-founder of Africa Online, Africa’s largest Internet service provider. She also worked as an economist at the World Bank. She holds an undergraduate degree in psychology and social relations from Harvard University and a Master’s in economics and public policy from the Woodrow Wilson School at Princeton University. ABSTRACT M-PESA, a mobile-phone-based electronic payments system, has been adopted by 8.5 million Kenyans in the relatively short span of three years. Surveys of users show it is a highly valued service, and Safaricom continues to expand the range of applications for which it can be used. This paper explores how Safaricom, the mobile operator that commercialises M-PESA, managed to create enough traction with both customers and retail stores, building trust and overcoming the adverse network effects that afflict new payments systems. The focus is on three key aspects of M-PESA’s success: (i) creating awareness and building trust through branding; (ii) creating a consistent user experience while building an extensive channel of retail agents offering cash in/cash out services; and (iii) a customer pricing and agent commission structure which focuses on key drivers of customer willingness to pay and incentivises early adoption. Keywords: M-PESA, mobile payments, mobile money transfer, remittances INTRODUCTION Who would have thought that nearly half the adults in Kenya1 would become regular users of a mobile-phone-based electronic payments system which processes more transactions domestically than Journal of Payments Strategy & Systems Vol. 4 No. 4, 2010, pp. 352–370 ᭧ Henry Stewart Publications, 1750–1806 Page 352 But first it sets the scene by reviewing the latest figures. • Channel management: Safaricom effectively leveraged its extensive network of airtime resellers to build a reliable. These problems reinforce each other in the early-stage development of a payments system. and equally hard to convince stores to sign up while there are few customers to be had. the more useful it becomes. (ii) going to non-bank retail outlets to meet their cash in/cash out needs. In order to establish itself. MPESA had to overcome three hurdles that are common to any new electronic payments system. Thus. This paper does not provide a historical account of M-PESA’s development.Mas:JSC page. It is hard to sell the proposition to customers while there are few stores to serve them.1 While network effects can help the scheme gain momentum once it reaches critical mass. nor does it provide a detailed account of the M-PESA service. M-PESA operates a system of low-value electronic accounts held by the mobile operator and accessible from their subscribers’ mobile phones through a SIM-card-resident Page 353 . • Pricing: Safaricom designed a pricing scheme for both customers and stores which provided incentives for both to join M-PESA early on. which rode on strong customer sense of affinity with and trust in the operator.qxd 12/01/2011 13:42 Page 353 Mas and Ng’weno Western Union does globally? How does one explain such an innovative service to people? How do you get them to trust such an intangible electronic service and be comfortable using the technology? How does one get the snowball going so that more and more people find value in joining the payments network? Mobile operator Safaricom has achieved all that with its M-PESA service — in only three years. In this case.2 Instead. emphasising the customerfacing efforts that continue to distinguish M-PESA from its many imitators around the world. namely: • Trust: Making users comfortable with the reliability of the new system. This paper gives an overview of the reasons for M-PESA’s astounding success in the market. • Chicken-and-egg trap: In order to grow. M-PESA overcame this hurdle through a combination of three key factors: • Brand: Safaricom built a strong service brand for M-PESA. STATUS OF AND LATEST DEVELOPMENTS IN M-PESA M-PESA was developed by Vodafone and launched commercially by its Kenyan affiliate Safaricom in March 2007. customers had to be comfortable with three elements which were new at the time in Kenya: (i) a payment system that was operated by a mobile operator. • Network effect: The value to the customer of a payment system depends on the number of people connected to and actively using it: the more people are on the network. they can make it difficult to attract early adopters in the early phase when there are few users. it focuses more narrowly on the three success factors mentioned above: branding. developments and limitations surrounding MPESA. consistent store network which served customers’ needs. creating a large hurdle to growth. M-PESA had to attract both customers and stores in tandem. and (iii) accessing their account and initiating transactions through their mobile phone. channel management and pricing. the scheme needed to drive both customer and store acquisition aggressively. KPSB. of which nearly half are located outside urban centres. US$100 million in annual revenue: this is equal to 9 per cent of overall Safaricom revenues. Page 354 .700 retail stores at which MPESA users can cash in and cash out.6 there were US$650m per month in cash deposits and withdrawal transactions at M-PESA stores. they remain quite low. These are very significant achievements.Mas:JSC page. and are capped at US$500.5 Over 17. All transactions are authorised and recorded in real time using secure SMS. the average transaction size is around US$33. The latest developments and figures reported by Safaricom after three years of operation (as of June 2010) are:4 • Over 10. this is equal to roughly 15 per cent of Kenyan gross domestic product (GDP). of which the majority are deemed to be • • • • active: this corresponds to 64 per cent of Safaricom’s customer base.qxd 12/01/2011 13:42 Page 354 Three keys to M-PESA’s success Figure 1 Offering financial services in Kenya Source: Central Bank of Kenya. Figure 1 shows the size of various retail channels in Kenya: the number of M-PESA stores is an order of magnitude higher than the total number of branches and automated teller machines (ATMs) made available by banks to their customers. probably around two transactions per month.3 million registered customers. or the number of post offices. As at March 2010. Safaricom websites application. M-PESA is going from strength to strength. US$371m per month in person-toperson (P2P) transfers: on an annualised basis. but Vodafone has stated that half the transactions are for a value of less than US$10. 21 per cent of the entire population or 46 per cent of adults.3 M-PESA is useful as a retail payment platform because of its reach into large segments of the population. Using existing retail stores as cash in/out outlets reduces deployment costs and provides greater convenience and lower cost of access to users. The conversion of cash and electronic value is performed at a network of retail stores (often referred to as agents) which are paid for by exchanging these two forms of liquidity on behalf of customers. although transactions per customer have been on a rising trend. 8 The growth of M-PESA is a testament to Safaricom’s vision and execution capacity. and reviewed the security features of the technology platform. But fundamentally. after a lobbying attack from the banking industry seeking to shut down the service. has not been catastrophic. However. the CBK allowed Safaricom to operate MPESA as a payments system. Safaricom was able to design the service as it saw fit. then regulate’ stance has been decisive. where the dominant operator has a market share of Page 355 . System downtime. the survey of M-PESA users in late 2008 revealed that less than 1 per cent of accounts had balances of over KSh1. with 84 per cent of the respondents in a recent largescale survey saying they would be worse off if M-PESA did not exist. in comparison with non-users. one-quarter of M-PESA users reported using their phones for storing money.11 So far.12 Beyond the actions of Safaricom and the CBK. The Kenyan regulator was consulted from the inception of the idea. more literate and better educated.Mas:JSC page. This has to do both with who the typical users of MPESA are and what they use it for: • A survey of 3. the public have also come to rely on it increasingly. M-PESA users had annual expenditures about 65 per cent greater than non-users and 20 per cent more assets. even in the face of pressure from banks. the Central Bank did an audit of the M-PESA service at the request of the Ministry of Finance and declared it safe and in line with the country’s objectives for financial inclusion. There are also limits on transaction sizes to address anti-money-laundering concerns. as there have been no major reports of fraud.9 outside the existing provisions of the banking law. Safaricom has had to pay a certain price for this arrangement. interest earned on deposited balances must go to a not-for-profit trust and cannot be appropriated by Safaricom or passed on to customers. The CBK and Safaricom worked out a model which provided sufficient prudential comfort to the CBK. although frequent.70). but also speaks to the very positive regulatory stance of the Central Bank of Kenya (CBK). In turn. The CBK insisted that all customer funds be deposited in a regulated financial institution. In late 2008. The average balance reported in following a government audit of M-PESA in August 2009 revealed that the average balance on MPESA accounts was only KSh203 (US$2. slightly older.qxd 12/01/2011 13:42 Page 355 Mas and Ng’weno But the overall impact on access to formal financial services at the base of the pyramid is still relatively modest. Users are also slightly more likely to be male. As it has become more ubiquitous. without having to contort its business model to fit within a prescribed regulatory model.10 The CBK has also enhanced its institutional oversight capacity. The CBK has continued to support the development of M-PESA. the Central Bank appears justified in its confidence in M-PESA. For instance. The CBK’s bold ‘experiment first.000 households in late 2008 showed that the average M-PESA user is. there are also specific country factors which no doubt made Kenya a conducive environment for a mobile money proposition:13 • Mobile market: Kenya has a concentrated mobile operator market. twice as likely to have a bank account (72 per cent vs 36 per cent).7 • In terms of services used. and has been actively involved in the development of M-PESA since its earliest pilot stages in 2004.000 (US$13). keeping abreast of innovation and technologically driven financial services. three-quarters of senders were using informal means such as the bus and friends and family. To do this. It enjoys relatively low prepay airtime commissions: Safaricom gives 6 per cent of sales to the channel.15 • A large market for domestic remittances. especially for domestic remittances. similar to richer countries such as Mauritius. quick or affordable alternatives — the Finaccess survey in 2006 revealed that 27 per cent of the population derived their main source of income from transfers from other people.qxd 12/01/2011 13:42 Page 356 Three keys to M-PESA’s success around 80 per cent. and (iii) the banks need to do the proper marketing groundwork to create relevant propositions. M-PESA will need to connect with banks which have the necessary vocation. In common with many other developing countries. called M-KESHO. Equity Bank.Mas:JSC page. Tunisia and Morocco.19 While Safaricom may not have had a Page 356 . M-PESA’s brand development At the beginning. A logical next step for M-PESA is for it to provide the payment ‘rails’ on which a broad range of financial services can be delivered cheaply and conveniently to all. Safaricom announced a banking product. of which 5 per cent typically goes to the retail outlet. inconvenient service which was not highly valued by users. The post offices had an expensive. M-KESHO is aimed at low-income customers with no minimum balance and free transfers to/from M-PESA. Kenya also has high mobile phone penetration of about 42 connections per 100 population. Safaricom was confronted with the challenge of developing trust in the new payment mechanism. customer insights and marketing (including product development) skills to serve the three-quarters of the Kenyan public who remain unbanked. Whereas initially M-PESA relied on existing Safaricom airtime outlets.18 (ii) the banks need to negotiate proper access and pricing with Safaricom. This was all the more difficult because Safaricom was introducing not only a new product. to a market that had little experience with formal financial services.17 jointly with the country’s largest savings bank. who were ready to seize new business opportunities created by the M-PESA agency model. the most common way of sending money around the country was through the bus system. but the process of connecting banks to MPESA will be gradual.16) • Poor financial alternatives. For its size and level of income (US$770 per capita14). • An entrepreneurial base of microentrepreneurs. equal to 17 per cent of Safaricom’s customer base of about six million customers at that time. (The 2009 survey revealed how the introduction of M-PESA has both increased the number of households receiving transfers (to over half) and how dramatically M-PESA has replaced informal transfer mechanisms. used by 40 per cent of senders in 2009. In addition. today it is common to find M-PESA stores which were established to do nothing else. In May 2010. given the large rural–urban migrations. It is too early to know what the uptake will be. since (i) the central bank needs first to issue its agent banking regulations. in Kenya there is cultural pressure to retain connection with one’s ancestral village. The internal launch target for M-PESA was about 1 million customers within one year. Safaricom also quickly had to achieve a critical mass of customers to defeat adverse network and chicken-and-egg effects. Prior to the introduction of M-PESA. which often results in split families. but an entire product category. there are few reliable. they simply offered a new solution for those situations — remote payments — where cash is not very effective and there are few convenient alternatives. But it achieved the purpose Page 357 . Below each of these aspects is reviewed. and had made sure to cover all of Kenya’s 69 district headquarters. management recognised the potential impact of M-PESA. Although people have proved creative in the use of M-PESA. Safaricom chose not to clutter their message with multiple claims on the attention (and anxieties) of a public not familiar with the service. The same television advertisement was used for the first two years.’ This single compelling use. At launch. which was well adapted to the common Kenyan phenomenon of split families. Safaricom had 750 stores.qxd 12/01/2011 13:42 Page 357 Mas and Ng’weno fully articulated marketing plan. sending money home continues to be one of the most important uses. Safaricom had a clear understanding of the key customer need which they were addressing. Safaricom managed to develop a strong service brand.20 M-PESA amply surpassed first-year forecasts. Safaricom was able to leverage the public goodwill which existed with the corporate brand. since everyone was new to it. A single. several days’ delays to reach customer service hotlines. they articulated the value proposition around a simple notion expressed with just three words: ‘Send money home. Possibly most important. was the unique focus of all marketing for the first year. It was also a massive logistical challenge.21 National launch at scale After small pilots involving fewer than 500 customers. Moreover. and remains the main (though no longer the only) marketing message three years later. While the M-PESA platform was able to provide a wider range of services. simple message M-PESA’s brand development and awareness-raising were driven by clear and specific messaging. The gamble paid off. As Safaricom market-tested the mobile money proposition. quickly turning the network effects in their favour as new customers begat more customers and turned M-PESA into lucrative business for more stores. User and store errors were frequent. In order to maximise the chance of acceptability in an unprepared market. in the first months after launch. making the value proposition equal for all subscribers and increasing the likelihood that the service could reach a critical mass in a short time-frame. Safaricom went for a full-blown national launch. which some market research has shown is even stronger than Safaricom’s corporate brand — itself already a powerful brand in Kenya. and committed the company to heavy investments in marketing before the proposition could be proved. in the early days. All this was supported by a service which was designed to be simple and easy-to-use. which led to a great deal of customer and store confusion and. Unlike many other mobile money deployments. they shifted the core proposition from repayment of microloans to helping people make person-to-person remittance payments to their friends and family. and treated M-PESA stores as valuable brand outposts. and was deliberate in its messaging and marketing mix. Safaricom has not set out to replace cash in day-to-day life. This was a high-risk strategy which not all operators or regulators are willing to countenance.22 M-PESA launched nationwide. since the brand of the parent company was at stake across its entire user base.Mas:JSC page. it nonetheless took a series of decisions which helped it to overcome these daunting challenges as it first introduced MPESA to the market. the M-PESA brand has become omnipresent. up-market urban dwellers with desk jobs to more ordinary Kenyans from lower-paid professions. An appropriate marketing mix Initial marketing featured and targeted the wealthier city dweller with the need to ‘send money home’ (see Figure 2).qxd 12/01/2011 13:42 Page 358 Three keys to M-PESA’s success Figure 2 Early M-PESA advert emphasising sending money from urban to rural areas. Television and radio were largely replaced by the omnipresent M-PESA branding at all outlets. explaining the product and demonstrating how to use it. While M-PESA’s launch was associated with a significant advertising campaign in traditional media such as television and radio. the marketing moved from young. The stores are immediately recognisable as they are painted Safaricom green with a prominent M-PESA logo. Stores are chosen for their Page 358 . with a wider range of services indicated. leaving strong brand recognition. it was no longer necessary to do this kind of hands-on outreach. Over time. This is a common marketing approach in Kenya for products that reach lower-end markets. because the traditional media are not viewed by the poor as relevant or trustworthy.000 outlets. Once a customer enters any store. This choice of the richer urban dweller as the initial customer created an aspirational image for M-PESA and avoided the impression that it was a low-value product aimed at the poor. Today. Over time. Safaricom achieved this by investing in store training and actively supervising the store network. linking into family and social ties of establishing national visibility and topof-mind awareness among large segments of the population. and Safaricom made best use of this channel. Newer ads (like the one in Figure 3) feature a general emotional appeal. the experience is remarkably consistent. Logistical problems subsided after a few months. Consistent store branding and customer experience M-PESA stores carried the new M-PESA brand into local communities. as people became more familiar with the product and how to use it. supported with a few large billboards. with nearly 19.Mas:JSC page.23 there was also massive outreach through road shows and tents which travelled around the country signing people up. using the same colour scheme. Maintaining the quality of the customer experience as the network has grown while at the same time containing costs has been a major challenge for Safaricom. they took pride in it. Its menu structure mimicked that of airtime transfers. which Safaricom provided free (including a SIMcard transfer service which transferred stored address books from the old SIM to the new). outside the establishment and associated tribal rivalries. In the early days. It was perceived by ordinary Kenyans as a home grown. which involved small projects in every village directly tailored to local concerns. In addition to these brand assets with the public. private sector success story. they were in the frontline of dealing with customers’ concerns and intimidation with having to deal with a new technology and a new range of financial services. Safaricom had a tried-and-tested market- Figure 3 Recent M-PESA ad with more general emotional appeal ing team which had a good understanding of its customer base. it introduced a face-to-face Page 359 . working on even the most basic phones. As the mobile operator with a dominant market share (over 80 per cent at M-PESA’s launch and scarcely less today. M-PESA customers needed to get a new SIM card. Safaricom was already a broadly respected company which had introduced products favoured by the public. including prepaid airtime accounts and per-second billing. and a relationship with a marketing company (Top Image) whom it could trust to deliver training and branding at the store level. but they thought it to be more important for the M-PESA business. An easy. As M-PESA uses an application resident on the customer’s SIM card. Safaricom’s image was bolstered by its charity work in Kenya. causing it to re-evaluate its approach. intuitive service design The customer interface was simple. Service branding building on a strong corporate image The heavy investment in branding was built on the already well-known and trusted Safaricom brand. Store clerks know how to use and explain the service to customers. and trusted it. despite the entrance of two new operators). a recently introduced product which served as a ‘stepping stone’ to introduce early adopters to the concept of transfer of electronic value.qxd 12/01/2011 13:42 Page 359 Mas and Ng’weno good knowledge of English and Kiswahili.Mas:JSC page. Safaricom had not required this of airtime resellers. While this was costly for Safaricom. The need for strong branding and consistent customer experience at the store led Safaricom to require operator exclusivity of M-PESA agents (ie they could not sell or promote the products of competing mobile operators). thereby reducing the number of direct contacts Safaricom had to deal with. Safaricom created a two-tier structure with individual stores (subagents.qxd 12/01/2011 13:42 Page 360 Three keys to M-PESA’s success opportunity for Safaricom to explain the service to new customers. Top Image acted • identification. Safaricom did not wish to manage thousands of retail stores directly. and (ii) distributing agent commissions (collecting the commission from Safaricom based on the overall performance of the stores under them and remunerating each store). This also allowed Safaricom to use informal ‘mom-and-pop’ stores as sub-agents. customers tended to blame M-PESA stores rather than Safaricom. Customers receive instant confirmation of their transaction. In the latter case. whenever there were issues with the service. the channel needed to support a range of activities. In addition. scalability (achieving rapid growth) and control (over the brand. it also wanted to maintain control over the customer experience.24 M-PESA’s store channel management Initial set-up For M-PESA to be broadly available to the bulk of the population. in Safaricom’s parlance) depending on master agents (referred to by Safaricom as agent Head Offices (HOs)). customer experience and geographic distribution of stores). among the most important of which were: • liquidity management. it sought to engage partners to help manage the individual stores. From an early stage. contracting and training of new stores • supervision of existing stores • distribution of commissions across stores Page 360 . To support the customer-facing activities of M-PESA stores. Although customer service was overwhelmed in the early months. which is important in helping customers to learn by experience to trust the system. While Safaricom wanted a scalable structure. Safaricom built a channel that was based on the key requirements of profitability (providing incentives for third-party retail players to get involved). Stores are free to switch between agent HOs.Mas:JSC page. enabling stores periodically to rebalance their holdings of cash and M-PESA balances. Safaricom had to design a channel structure which could support thousands of M-PESA stores spread across a broad geography to offer cash in/cash out services. The service sought to offer reassurance to customers in several ways. Safaricom does not prescribe the terms of agent HO–store contracts. to conduct the evaluation and training of new stores. while Safaricom’s contract was with the HO. Individual stores may be directly owned by an agent HO or may be working for one under contract. screening. Hence. so they are free to work out their own liquidity management arrangements and split of agent commissions. lost PINs and lost phones. including the things they most feared: lost SIMSs. Agent HOs maintain all contact with Safaricom. Top Image was charged with visiting stores monthly and scoring them against a range of criteria. and perform two key functions: (i) liquidity management (buying and selling M-PESA balance from Safaricom and making it available to individual stores under their responsibility). Safaricom engaged a local firm. customers quickly learned that M-PESA transfers were secure. In fact. as well as to do the periodic on-site supervision of all stores. First. Safaricom opted for two parallel mechanisms for channel management. Safaricom invested in customer service to help people deal with and resolve the problems they faced using the system. Top Image. This shift represents a natural shift in strategic priorities as the M-PESA system takes root. The difference between the customer tariffs and the total channel com- Page 361 . The first point is being addressed by creating a new class of players called ‘agent aggregators’ sitting at the top of the agent channel. centralised control (through its contract with Top Image) over the key elements of the customer experience. The minimum cash transaction size is KSh100 (US$1.Mas:JSC page. At the same time.25 Thus. The middle two boxes in Figure 5 represent the total commission that Safaricom pays to the retail channel. in Kenyan shillings (KSh). At the same time. Thus. Thus. store-level branding becomes less relevant. Sixty per cent of MPESA stores belong to agent HOs. including vetting and training new agents and ensuring that stores met guidelines and offered a reasonable service.000. Safaricom delegated the more routine. based on the number of customers they are serving and the volume of transactions they are performing per day. M-PESA customer tariff and store commission structure Figure 4 shows the M-PESA customer tariff structure as posted on the Web and at each store location. in both KSh and US cents. Safaricom retained direct. which is getting to be a challenge as the number of stores shoots past 18. • Safaricom absorbs too much cost directly. others are stepped according to the size of the transaction. Safaricom have selected the aggregators on two key criteria: (i) liquidity. in addition to paying channel commissions through agent HOs. This is to address two key concerns with the current system: • The channel is flat. as aggregators are required to have a minimum deposit in their M-PESA account with Safaricom. while the top-right box shows the equivalent values in US cents (using an exchange rate of KSh75 per US$). There will be no territoriality assigned to the aggregators: they can have sub-agents wherever they like. desk-bound. The new channel management structure represents a shift from one that emphasised control and consistency of the customer experience to one that emphasises channel scalability and efficiency. with fewer than a dozen stores. non-customer-facing store support activities to a larger pool of agent HOs. There is a costly duplication of channel players. competition between agents can be relied upon increasingly to ensure proper service at the store level. Once customers have a certain expectation of service. targeted marketing messages through broad-based marketing channels. and (ii) performance. and Safaricom will become increasingly focused on driving more nuanced. As the number of agents surges. and not through the agent HOs. they rather than Safaricom can take the lead in disciplining agents. it created some degree of competition among agent HOs. Some of the fees are a flat rate.3). Safaricom pays Top Image for on-site services directly.qxd 12/01/2011 13:42 Page 361 Mas and Ng’weno as direct sub-contractor to Safaricom. avoiding management bottlenecks and reducing unit costs become more important. Recent changes in channel management roles and structure Safaricom is presently undertaking a major overhaul of its M-PESA store channel. And as awareness of the service grows. The top-left box in Figure 5 states these fees. but not on aspects that were crucial to maintaining the quality and consistency of the customer experience. Safaricom has to deal with too many agent HOs to manage their channel. qxd 12/01/2011 13:42 Page 362 Three keys to M-PESA’s success Figure 4 Posted customer tariffs for M-PESA missions is kept by Safaricom. and the agent HO passes on the fees through their store channel. All agent HOs are free to distribute their commis- Page 362 . none are paid by customers in cash to the store. Safaricom pays agent HOs according to the commissions noted here. The bottom two boxes in Figure 5 represent the value of the commissions typically retained by retail outlets. Note that all customer commissions are charged by Safaricom against the customers’ mobile wallet account.Mas:JSC page. which is in fact equal to what the channel gets (13.qxd 12/01/2011 13:42 Page 363 Mas and Ng’weno Figure 5 Analysis of customer tariffs and retail commissions for M-PESA transactions sions down their retail channel as they see fit. the customer pricing and agent commission structure is summarised by type of transaction. for transactions under US$31. the bottom two boxes in Figure 5 represent the after-tax take by typical M-PESA outlets.Mas:JSC page.3¢ on withdrawal). the provider of the ATM network.3¢ (0¢ on deposit + 33.26 Thus. Transactions against cash A round-trip savings transaction (one deposit + one withdrawal) costs the customer 33.3¢ on the deposit + 20¢ on the withdrawal).3¢). a 21 per cent tax is levied on cash transactions at retail stores — 16 per cent as value added tax and 5 per cent withholding against income tax. The average agent channel commission per cash transactions is 16. assuming equal volumes of deposits and withdrawals. Below. But PesaPoint. but most pass on 70 per cent of the commissions they receive from Safaricom to the retail outlets. not store by store).3¢ and 20¢). This is in fact a recent price change: it used to be the same as regular withdrawals at a store.2¢ (average of 7. Withdrawals from M-PESA at ATMs are slightly more expensive than at retail outlets (40¢ vs 33. and recoups it when customers withdraw (this works on aggregate. Safaricom does not make any money on cash transactions: Safaricom merely ‘advances’ commissions to the channel when customers deposit. In addition. as it claimed to be losing money on such transactions.7¢ (average of 13. So. renegotiated its deal with Safaricom (at the same time as it raised its fees to banks) to get more commission. which represent the bulk of transactions in M-PESA. Page 363 .4¢ and 11. of which the retail outlet gets 9.1¢) after tax. This may seem odd: the customer is penalised while the noncustomer gets a free service. if the customer buys the airtime through MPESA.29 M-PESA customers can send money to non-M-PESA customers. M-PESA generally charges the same amount for transfers of money for bill payment as for P2P (ie KSh30 per transfer). There is some channel conflict here: when a retail outlet sells a prepaid card. Thus. only 8 per cent reported using M-PESA daily or weekly. who has the money and the understanding of how M-PESA works — and hence the clout to insist that the receiver register as an M-PESA user. which is not the case for a customer. except that customers send money to the M-PESA account of the recipient institution. though. For instance. but at the other end cashing out is free for a non-customer. Customers can buy airtime directly from their mobile wallet. Under this service. The logic. the cost of the transfer can be shared between the biller and the customer in the proportion defined by each biller. money is debited from the sender’s account. including any person with a GSM mobile phone in Kenya. with the receiver’s experience being similar to that with Western Union today. is to put the pain on the sender. For remote payments. Customers must enter the biller code (a number that uniquely identifies the biller) on the MPESA phone menu. This increased the speed at which technology ‘laggards’ registered with the service. customers pay for a purely electronic transfer more than double what they pay for the average cash transaction (17¢) — despite the cost to provide being lower for purely electronic transactions than those involving cash.27 But this pricing limits the suitability of using M-PESA for face-to-face transactions for which the alternative of cash is much less costly. It is analogous to a P2P transfer. and does not actively promote Page 364 .3¢). at zero cost — but also without a discount.qxd 12/01/2011 13:42 Page 364 Three keys to M-PESA’s success Electronic transfers of value Person-to-person transfers cost a flat rate of 40¢. and also have the option of entering an account number (which is necessary if the biller cannot associate the customer’s phone number with their bill account number). in the case of bill pay. which it can use to claim the monetary value at any MPESA store. the utility and the customer split the charge (ie customers pay KSh15 per bill paid). This is where Safaricom makes the bulk of its revenue. whether they are subscribers of Safaricom or of any of the other three competing networks (Zain.28 This limits the frequency of use of M-PESA: in a survey of users in late 2008. M-PESA is aware of this cannibalisation risk for the channel. The pricing on this service is interesting: it is a lot more expensive for a customer to send to a noncustomer than to a customer (US$1 vs 40¢). M-PESA is cheaper than the other available mechanisms: money transfer by the bus companies. for electricity bill payments. the channel gets 5 per cent on the value of the card sold. The ability to send money to any mobile phone subscriber was particularly important at the launch of the service when only few customers were registered. it is an account-tophone service. This reflects a notion of optimal pricing that is based less on cost and more on customer willingness-to-pay: enabling remote payments is the biggest customer pain point which M-PESA aims to address. Bill payment is a new service launched in 2009 (see Figure 6). Thus. Orange and Yu). there is no direct channel commission (though at some point there may have been a cash-in fee of 13.Mas:JSC page. and proved to be an effective customer acquisition mechanism. However. Kenya Post’s Postapay or Western Union. and the recipient gets a code by SMS. Mas:JSC page. and (iv) use of store space and staff time to handle the transactions. and many outlets no longer treat it as an add-on. (iii) enhanced security risk and ensuing insurance premiums. which is a full-time job. M-PESA can be used to send international remittances from the UK to Kenya (Figure 6 shows a Safaricom banner ad promoting this service). traffic on MPESA’s new international remittance channel with the UK is low. (ii) maintaining unremunerated working capital balances. offering only a small discount from standard Western Union pricing. Safaricom has partnered with Western Union and two Kenyan diaspora companies (one that shows Kenyan television channels on UK cable) to provide a cashin network in the UK. but Safaricom has been careful not to undermine their subsistence. A retail outlet conducting 100 transactions per day will be doing one every six minutes on average (assuming a ten-hour workday). the retail outlet incurs the following costs: (i) trips to the bank.qxd 12/01/2011 13:42 Page 365 Mas and Ng’weno Figure 6 Ads for three new M-PESA payment services: bill pay. Page 365 . As a result.30 For a store to make US$6 — roughly twice a clerk’s daily wage — it needs to handle 65 transactions per day. both in cash and in their MPESA account. due to high transaction volumes. The majority of M-PESA stores handle more than this number of transactions. Though not shown on the standard pricing sheet. Figure 7 shows the dollar value of commissions that will accrue to a retail outlet on a daily basis. The fear of cannibalisation actually drives many airtime resellers to signup as M-PESA outlets. Against these commissions. The cost is high (GB£4 to send GB£100). M-PESA has become a viable stand-alone business. retail outlets are finding M-PESA a better business than prepaid airtime. for which airtime purchases is their primary driver of transaction volume. based on the number of cash transactions they perform for MPESA customers. In fact. salary distributions and international remittances the purchase of airtime through M-PESA — unlike many other mobile money schemes. usually on a daily basis. Mas:JSC page. The cost of a round-trip savings transaction (deposit + withdrawal) is between 1 and 2 per cent of balance for withdrawals greater than US$17. it does earn some margin on larger transactions. due to the flat-rate P2P charge. it can be as high as 5 per cent for withdrawals of US$5. (It was noted above that Safaricom does not earn any fees net of agent commissions on cash transactions.70. P2P transfer and cash-out costs). MPESA is more expensive for money transfers than for savings.) Page 366 . That is the case for transactions below US$33. Money transfers of less than US$15 cost more than 5 per cent (including cash-in.qxd 12/01/2011 13:42 Page 366 Three keys to M-PESA’s success Figure 7 Retail outlet commissions as a function of number of transactions Figure 8 Customer fees as percentage of transaction size Figure 8 shows the customer commissions as a percentage of the transaction size. 31 This might help to explain why MPESA is not used as frequently as a daily deposit service.0 912. M-PESA’s success factors and future evolution The experience of M-PESA demonstrates how powerful a payment network offering convenience at an affordable cost can be.3 Table 1 shows prototypical bundles of usage of M-PESA. all values are in US$. whereas use focusing on money transfers has a high margin. it is interesting to note that M-PESA users do not tend to be aware that deposits are free. (It is assumed here that all transactions are under US$33 in size. And it built on the trust that people already had with Safaricom. To recap. Safaricom is an attractive savings option — except that it bears no interest. • Store channel development. From the beginning.) The use that is heavy on deposits is loss making for Safaricom. Customer surveys of M-PESA show that people value fast. While users are reasonably happy with M-PESA pricing. despite the substantial value that Safaricom provides for this use.qxd 12/01/2011 13:42 Page 367 Mas and Ng’weno Table 1: Total account costs for customers Net revenue to Safaricom Monthly cost to user (¢) (after retail commissions) Type of use (¢) Monthly remittance (pay-through account) Daily deposit service (‘e-susu’) Accumulation account Working capital account Transactional account Monthly transactions performed Sender: 1 deposit + 1 P2P Receiver: 1 withdrawal 1 deposit daily + 1 withdrawal at end of month 1 deposit weekly 2 deposits + 2 withdrawals + 1 check balance per week 2 deposits + 2 withdrawal + 4 payments + 1 check balance per week Sender: 40.3 645. Safaricom understood that the primary role of the mobile phone is to enable the creation of a retail outlet-based channel for cash- Page 367 . since deposits are free.3 Total: 73. branding and building up the retail channel.3 −385. once a critical mass is reached.Mas:JSC page. and the associated total account costs. while it still pays per-deposit commissions to the agents.0 Receiver: 33. It was also supported by heavy advertising focused on the key customer need that M-PESA was addressing. according to Figure 8. and Safaricom has little incentive to promote it as such. This was helped by a rapidly expanding base of retail stores which carried the MPESA brand visibly into the heart of communities where people worked and congregated.7 −53.3 33. Safaricom knew it had to develop customer trust in the new payment mechanism. easy to use service far ahead of price. the success factors from the Safaricom side have been: • Brand development.0 53.3 0 272. M-PESA is now enjoying economies of scale as much on the demand side as on the cost side. As currently priced. and understood it needed to get critical mass quickly in order for existing customers to be the prime mechanism to draw in new customers. It also shows that achieving critical mass takes significant investment in marketing.3 5. S. ACKNOWLEDGEMENTS The authors would like to thank the following for very helpful comments on earlier drafts: Mireya Almazán. Paul Makin. (2) For these. but also provided good cash flow to stores in the early days. Mark Napier. and Lonie.qxd 12/01/2011 13:42 Page 368 Three keys to M-PESA’s success to-digital value conversion. June. for financial access data derived from the FinAccess survey. Page 368 . Safaricom was in uncharted territory. a nationally representative survey of 6.19 per cent of adults have access to a formal bank account. N. David Porteous. such as government social welfare payment distributions and tax collections. • Pricing. REFERENCES AND NOTES (1) Kenya has a total population of nearly 40 million. its regulatory treatment as a payments vehicle needs to be formalised so that it can become regulated in the most appropriate way. when transaction volumes were low. It can enable a broader range of institutional payments. profitably for Safaricom) at much smaller transaction sizes. Customer pricing was designed to encourage customers to experiment with the service: free and quick registration to the service. There are some recurrent problems of agent cash liquidity and system downtime which still need to be resolved. In fact. so that customers can avail themselves of the broader product set offered by regulated financial institutions. drawing in banks to help with liquidity management. Stores acting had consistent branding. Matu Mugo. Aiaze Mitha. reflecting that customer willingness to pay is higher for remote payments where customers’ alternatives are weakest. but the model still has substantial room to develop further. pricing which tapped into customers’ willingness to pay for remote payments. Nairobi. the channel restructuring now under way is probably the main one. FSDT.600 households conducted in early 2009. Safaricom’s profit margin is loaded on P2P transfer fees rather than on the cash in/cash out fees. It needs to link to a fuller range of banks. received substantial on-the-spot training. But Safaricom did make a few early decisions which proved very valuable: the aggressive promotion of the M-PESA brand. See Financial Sector Deepening Trust (FSDT) (2009).Mas:JSC page. David Ferrand. with hindsight. Daniel Radcliffe and Graham Wright. There is the danger. whether or not they were subscribed to the service. and no doubt there were elements of serendipity. Senior managers drove the development of a fully managed retail channel strongly. of ascribing an undue level of deliberateness to Safaricom as it navigated these issues. and were frequently visited and supervised. Christoph Kneiding. see Hughes. Stores receive new customer registration bonuses which not only incentivised growth. with 78 per cent living in rural areas and a GDP per capita of US$1600 (PPPbasis). building the mechanisms to drive a consistent customer experience through the agents. Amrik Heyer. When it launched M-PESA. Olga Morawczynski. ‘FinAccess National Survey 2009: Dynamics of Kenya’s changing financial landscape’. M-PESA has been successful beyond what anyone could have imagined at its launch. Johann Bezuidenhoudt. so that customers have more incentive to leave money in their M-PESA accounts. M-PESA can introduce more finely segmented tariff and sub-agent models that work (affordably for customers. Finally. permitting sending of money to any mobile phone user on any network. Safaricom has done remarkably few course corrections. free deposits and the ability to send money to any mobile phone subscriber. (17) Kesho is Kiswahili for ‘tomorrow’. 7. 2. and Morawczynski. 1 above. 27th January. pending approval of a new National Payments System Bill. (20) A survey commissioned by Superbrands Kenya and conducted by TNS Research International in 2009 rated M-PESA number 14 in brand strength in the Kenyan market. pp. 16). B. Vol. (21) These data are from the FinAccess survey (FSDT. 8 above. p. and Mas. December. 1 above. I.Mas:JSC page. No. Nairobi. October.safaricom. (19) Safaricom company results for the year ending March 2007. Brookings Institution. and FSDT. 6. The results of the survey are explained in Okoth. pp. (15) Comparisons are based on GSM Association (2009) ‘Wireless Intelligence Database’. I. Nairobi. The SIM card contains encryption keys. respectively. 1–2.com (16) Finaccess surveys: FSDT (2007) ‘Financial Access in Kenya: Results of the 2006 National Survey’. (2008) ‘The Performance and Impact of M-PESA: Preliminary Evidence from a Household Survey’.co. 8 and 39) and Okoth. T. Innovations. M. October. FSDT. Washington. A. (18) The Finance Act 2009 contained an amendment to the Banking Act which permits banks to use non-bank retail agents. the CBK was reported to be working on the draft implementing regulations. ref.wirelessintelligence. ref. FSDT. The Central Bank of Kenya Act was amended in 2003 to give CBK broad oversight mandate over payment systems. (13) For a fuller discussion of environmental or country factors that are likely to influence the success of mobile money propositions. 63–81. and involved the Commercial Bank of Africa. above. FSDT (2009) ‘Research on Mobile Payments Experience: M-PESA in Kenya’. Additional figures are taken from Safaricom’s published year-end results for the period ending March 2010 and Central Bank of Kenya reports. and Mas. 4.qxd 12/01/2011 13:42 Page 369 Mas and Ng’weno (3) (4) (5) (6) (7) (8) (9) (10) (2009) ‘M-PESA: Mobile Money for the Unbanked’. (14) World Bank (2008) ‘GNI per capita’. year ending March 2010. unpublished draft. Page 369 . (12) See Suri. Nos. in addition to Safaricom. p. and drives the phone’s menu. 2. For key monthly statistics for M-PESA. p. (2009) ‘Expanding the Financial Services Frontier: Lessons from Mobile Phone Banking in Kenya’. Faulu Kenya and MicroSave. ref. O. Innovations. N. whereas Safaricom fell below 17 (the lowest reported in the survey). The Standard. J. ref. World Bank. and Jack. 1 above. available at www. pp. see http://www. 1 above. unpublished paper. (22) The earliest pilot project conducted in 2004/05 revolved around microloan repayments. June. (2009) ‘Seeking Fertile Grounds for Mobile Money’. ref. see Heyer. and Ndung’u. 16) confirmed that 40 per cent of adults had used M-PESA. (11) See Kimenyi. The subscriber identification module (SIM) card is a smart card found inside mobile phones which are based on the GSM family of protocols. The data in this paragraph are from the FinAccess survey (FSDT. secures the user’s PIN on entry. (2009) ‘Designing Mobile Money Services: Lessons from M-PESA’. The 2009 FinAccess survey (FSDT. but the operational modalities for its regulatory powers over payments systems have not been implemented. The short messaging service (SMS) is a data messaging channel available on GSM phones.pdf. which has languished in Parliament. ref.ke/ fileadmin/template/main/downloads/ M-PESA_Statistics. FSDT. unpublished paper. Vol. Safaricom Annual Report. (2009) ‘Regulator Gives M-Pesa A Clean Bill of Health’. At the time of writing this paper in December 2009. 77–92. Vodafone. Special Edition for the Mobile World Congress 2009. DC. and Pickens. August. p. 282. (27) In her field research. (2009) ‘Poor People Using Mobile (28) (29) (30) (31) Financial Services: Observations on Customer Usage and Impact from M-PESA’.qxd 12/01/2011 13:42 Page 370 Three keys to M-PESA’s success (23) A survey of 1. And yet. 28. that all customer transactions are for a value of less than US$33. (26) M-PESA services became VAT-exempt in the new national budget law which came into effect July 2009. (25) FSDT. p. television or radio (FSDT.Mas:JSC page. 7 above. Washington DC. C. this exemption was for a year. further report. M. FSDT. October. ref. (2008) ‘Examining Trust In Mobile Banking Transactions: The Case of M-PESA in Kenya’. Smith. Page 370 . 7 above. and Jack. This assumes an equal number of deposits and withdrawals. on the basis of the same survey.210 users in late 2008 revealed that 70 per cent of survey respondents claimed they had first heard about M-PESA from advertisements. and they tend to be smaller than the average remittance. (2009) ‘Mobile Money: The Economics of M-Pesa’. which can be substantial. while those received are about 5. However. Vol. p. MA. FSDT. Boston. it is reported that M-PESA is quite often used to pay for taxi rides and late-night beers. M. (eds) ‘Social Dimensions of Information and Communication Technology Policy: Proceedings of the 8th International Conference on Human Choice and Computers HCC8’. ref. Thus. G. p. Olga Morawczynski finds that sending KSh1000 through M-PESA was 27 per cent cheaper than the post office’s PostaPay. seen a 16 per cent increase in their revenues. 25th–26th September. South Africa. O. Indeed. Pretoria. the channel has. P. and Suri. 15.6 per cent. (24) Morawczynski. CGAP Brief. International Federation for Information Processing TC 9. 21. so this analysis assumes that the VAT is still applicable. despite this. 287–298. the higher transaction costs of M-PESA in face-to-face transactions are outweighed by the greater convenience and security which it offers.5 per cent of monthly household expenditure on average. 7 above. Monthly remittances sent are smaller. O. amounting to approximately 4. that households send and receive remittances on average once every three to four months. in effect.. About a third of remittances are sent and received via M-PESA. ref. pp.5 per cent of monthly expenditure. and van den Besselaar. and 68 per cent cheaper than sending it via a bus company. B. amounting to about 1. and excludes registration commissions. 6). CGAP. and Miscione. in Avgerou. See Morawczynski. in certain circumstances. Springer. T. 7 above. unpublished paper. ref.