1 Is there Really any Money in Mobile Money? Author: Paul Leishman

2 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? Contents 2 Is There Really any Money in Mobile Money? How Much Must an MNO Invest in Mobile Money Before Turning a Profit? 4 How Significant are Airtime Distribution Savings to Profitability? 5 6 How Significant are Churn Reduction Benefits to Profitability? How Significant is ARPU Uplift to Profitability? 7 How Significant are Direct Revenues to Profitability? 8 9 How can an MNO Manage Costs to Achieve Profitability? How can MNOs Ensure Their Tariff and Commission Models are Well Designed? 11 Appendix A: Resources 13 1

3 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? Is There Really any Money in Mobile Money? Exhibit 1: Growth of active customers and transactions From Afghanistan to Zambia, mobile network for MTN Uganda’s MobileMoney operators (MNOs) in developing countries are launching mobile money services at a rapid pace. Yet while their enthusiasm to enter this business is Active customers 400000 clear – to date 78 deployments have been launched and another 83 are being planned – their rationale 350000 for doing so is not. There’s no doubt that Safaricom’s 300000 runaway hit, M-PESA, is profitable. But Kenya 250000 represents somewhat of an anomaly – the perfect 200000 coalescence of latent demand, a dominant MNO and a progressive regulator. So the question remains for 150000 just about every MNO outside of Kenya: is there 100000 really any money in mobile money? 50000 To answer this question, GSMA has studied our 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 portfolio of MMU fund grantees, which includes months rapidly scaling deployments like easypaisa in defined as any customer that has performed Active Customers, Pakistan, M-PESA in Tanzania and Kenya, and True a cash-in, P2P transfer, cash-out, or airtime top-up within 90 Money in Thailand; interviewed mobile money days. practitioners; and conducted a deep-dive into the operational and financial results of MTN Uganda’s P2P transfers MobileMoney, a promising deployment from the 400000 East African country of 32 million where 80% of the population lacks access to financial services. 350000 300000 In an effort to provide a level of depth that’s useful to 250000 mobile money practitioners, we’ll focus primarily on 200000 MTN Uganda’s MobileMoney, but will be sure to put their experience in a global context where relevant. 150000 So before we answer the provocative question posed 100000 in the title of this chapter, first a bit of background on 50000 MTN Uganda’s MobileMoney. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Launched in partnership with Stanbic Bank in March months 2009, the service enables customers to send and P2P transfers, defined as a transfer of money to a registered receive money domestically and buy airtime using or unregistered customer. their mobile phone; it’s delivered via a network of it’s growing 1,400 agents; and, most importantly, rapidly, now counting 400,000 active customers, While MTN does have a full roadmap of features processing as many as 385,000 P2P transfers per planned, we’ve not made any projections in our month, and serving as the channel through which study: every insight presented is based on actual data 1 3% of total airtime is sold per month. and has been analysed using our GSMA financial model. MTN Uganda’s MobileMoney is now cash- flow positive on a month-to-month basis – and they crossed this critical threshold just 14 months after launch. 1 2 For more information, refer to Exhibit 1

4 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? So, is there really any money in mobile money? In account for 48% of MobileMoney’s gross profit the case of MTN Uganda’s MobileMoney, the answer to date. We also found that 55% of the costs in the The service is now cash-flow positive on a is yes. business to date are variable and step rather than month-to-month basis – and they crossed this MTN’s financing requirement fixed; in other words, critical threshold just 14 months after launch. has been (and increasingly will be over time) driven MTN’s peak financing requirement, or the amount . by their own customer growth that they had to finance before MobileMoney became Exhibit 3: Gross Profit contribution to date (MTN cash-flow positive, was less than US$4 million. MobileMoney Uganda) Exhibit 2: Financing requirement for MTN Uganda’s MobileMoney 3% 3000000 2000000 33% 52% 1000000 US$ 0 12% -1000000 -2000000 Money Transfer Contribution Airtime Distribution Savings ARPU from Churn Reduction Retained -3000000 Uplift in Voice/Data Consumption -4000000 Month 6 Month 12 Month 0 Month 18 Exhibit 4: Breakdown of Total, Year-1, and Year-2 costs (MTN MobileMoney Uganda) including customer fees for money transfer Monthly Revenue, Year 1 Total Year 2 service, savings from airtime distribution, income from reduced churn, Year 1 Total Year 2 and increased share of wallet from voice and SMS. 14% 22% which includes agent commissions, Monthly Cash Expenses, 13% 43% 14% 45% customer acquisition costs, agent costs, technology costs, SG&A 22% 12% costs, and up-front investment in technology (excludes non-cash 66% items including depreciation of capital assets). 73% 13% 43% 12% 45% 12% revenue equal to total monthly Monthly Net Cash Flow, Year 1 Total Year 2 less monthly cash expenses 66% 73% Fixed costs Step costs Variable costs 14% 22% the rolling sum Cumulative financing requirement , 12% of monthly net cash-flow; the lowest point represents the peak 13% 43% 45% financing requirement 12% 66% 73% Note: the timing of some costs incurred may have been altered 12% slightly in a way that protects supplier confidentiality but does Fixed costs Variable costs Step costs not affect the underlying story. including marketing, field agency costs, SIM upgrade Fixed Costs, fees for non-mobile money customers (assumption for amount Fixed costs Variable costs Step costs attributable to MM), agent handset subsidies, fixed m-wallet provider fees (assumption for up-front investment), agent POS merchandising. For MTN Uganda, these numbers are exciting. But what’s interesting for mobile money practitioners Step Costs, including management staff and back- office staff everywhere is exactly how this service became We found that indirect benefits cash-flow positive. Variable Costs, agent commissions, SMS fees, SIM unique to MNOs – including savings from airtime replacement, registration commissions, variable m-wallet distribution, reduction in churn, and increased provider fees, ARPU loss from discounting. share of wallet for voice and SMS – combined to 3

5 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? How Much Must an MNO Invest in Mobile Money Before Turning a Profit? Finally, and perhaps most important, The first chapter of this article introduced MTN a successful Uganda’s MobileMoney, a service that has turned mobile money service’s financing requirement will ultimately be driven by variable and step rather an exciting corner into cash-flow positive territory. But the CFO of any mobile network operator (MNO) than fixed costs; in other words, it’s difficult to ‘spend like Safaricom’ unless customers are adopting knows that simply getting out of the red on a month-to-month basis is not enough; his alternative and using the service. investment options are usually very attractive, so he So instead of asking “how much must I invest?”, needs to know just how much is required to scale a mobile money service – and whether future income the more relevant question practitioners have begun asking is “what costs will drive my financing will justify the spend. requirement?”. To answer this question, let’s again examine the case of MTN Uganda’s MobileMoney. Unfortunately, there’s no generic amount that an MNO – in any market, operating with any business In Exhibit 4, we see that so far, 55% of MobileMoney’s model – can assume they need to invest before turning a profit. Until now, Safaricom’s M-PESA has financing requirement stems from variable and step provided the industry’s only reference point; and the thus, more than half costs, and 45% from fixed costs – best estimates reckon that Safaricom and Vodafone of their financing requirement has come, in part, 3 from customer adoption and use. We also see that in have spent to the tune of US$30 million scaling the 2 service so far. their first year of operation, they incur an initial flurry Our team’s recent analysis of MTN Uganda’s MobileMoney indicates that they’ve of fixed costs, including investment in the m-Wallet platform, upgrades to their SIM access gateway, spent somewhat less, roughly $10.5 million in total 4 spending on above-the-line marketing , and opting to costs and investments to date, driving the service into cash-flow positive territory on a month-to- embed their application on all new SIM cards. These fixed costs were not insignificant – yet as the service month basis. It does merit note, however, that in grew, they were quickly overtaken by variable costs, its first 16 months M-PESA grew twice as fast as MobileMoney in terms of customer registration as a including customer registration commissions, agent commissions, and per-customer technology licensing percent of mobile subscribers (roughly 31% vs. 17% In the second year of operations, variable and by month 16). fees. step costs like these account for fully 66% of the Alas, in the absence of context, top-line investment total costs in the business. figures like these are of limited applicability. For It’s clear, then, that the financing requirement for a starters, Kenya’s population is 36 million – so successful mobile money service is driven largely the country is a bad comparable for practitioners in Fiji (population 844,000), India (population by variable and step costs – but is all the spending even worthwhile? That is, can mobile money services 1,100,000,000), and most countries in between. Moreover, for better or worse, MNOs in other generate a sufficient net present value (NPV)? countries have not replicated the M-PESA model: in For MTN Uganda’s MobileMoney, the signs are some cases they’ve promoted different services, and promising: if we assume that the service continues to grow roughly at Uganda’s rate of inflation and in others struck different bank partnerships – and then include the terminal value in our calculation, each of these factors impacts profitability. the NPV for MobileMoney is positive. It’s difficult to say exactly when the cumulative net cash-flow curve in Exhibit 2 will become positive, particularly It’s difficult to ‘spend like Safaricom’ unless since MTN is planning to launch additional services customers are adopting and using the service that will surely generate incremental revenue; still, simply based on the foundation they’ve laid with their domestic money transfer and mobile top-up offerings, it’s only a matter of time before MTN recoups its investment. 2 Mas, Ignacio and Radcliffe, Daniel, Scaling Mobile Money (September 22, 2010) 4 3 And as the service grows, the model will be predicated even more on variable and step costs 4 MTN has spent a total of US$850,000 on above-the-line marketing; this amount is assumed to be slightly skewed to up-front spending.

6 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? How Significant are Airtime Distribution Savings to Profitability? Outside of Africa, mobile top-up has been an equally One of the most important sources of value for mobile important value driver for mobile money services – network operators (MNOs) who offer mobile money and often by strategic necessity. In the Philippines, services is the ability to sell airtime using the platform. When a customer buys airtime using mobile money where existing domestic money transfer alternatives rather than with scratch cards, operators unlock value are better than Kenya, both SMART Money and the First, they pay lower commissions: in two ways. G-Cash have aggressively promoted their mobile commissions paid to agents for performing cash-in top-up services; in Indonesia and Thailand where (a necessary step before buying airtime) are typically regulatory guidelines currently don’t allow customers lower than the discounts at which MNOs sell airtime to withdraw money from an e-wallet, Telkomsel and True Move have both promoted mobile top-ups for to the channel—although the degree of difference will vary by market. T-Cash and True Money respectively as an important Second, MNOs save on the manufacturing and storage of scratch cards. Any ‘use of electronic funds’; and in Fiji, where physical savings realized in these ways flow straight to their distribution of scratch cards to remote areas can be a pre-tax bottom line. challenge, Digicel and Vodafone have both launched with mobile top-up as a core feature. So how big a deal is this? We’ve found that for successful services, savings from airtime distribution So how can MNOs evaluate the importance of For MTN Uganda’s mobile top-ups to their profitability? The first step is can be a big deal indeed. to identify the size of the discount at which airtime MobileMoney, this value source has contributed a total of 12% of their gross profit to date. Even is sold to the channel: the higher the discount, the greater the opportunity for mobile money to deliver though the service is less than a year and a half old, MTN has still managed to derive significant value value. Second, an MNO must estimate the percent from their mobile top-up feature: in their best month of total airtime sales they can reasonably convert from scratch-cards to mobile money. And third, an so far, roughly 3% of total airtime was sold through MNO must consider the myriad costs involved in MobileMoney – at more than a 9% savings compared to airtime that would have otherwise been purchased facilitating mobile top-ups. These can include but are via scratch cards. not limited to: perpetuities paid to top-tier agents on airtime sales for customers they register for mobile money; incentives paid directly to frontline agents or customers themselves to stimulate adoption; and For MTN Uganda’s MobileMoney, savings from commissions paid to agents for facilitating cash-in airtime distribution has contributed a total of (because customers can’t buy airtime from an empty 12% of their gross profit to date e-wallet). Beyond Uganda, MNOs are collectively eyeing – or already capitalizing on – mobile top-up as a means of reducing their cost of distributing airtime. Safaricom has led the way, apparently selling 19% of its airtime on M-PESA. And in the context of total profitability for their service, this feat has been important: if we assume that Safaricom saves 8% in costs on airtime sold through M-PESA, and assume that in their last fiscal year, they sold about $800 million in prepaid airtime in total, this suggests that they’d have generated savings of $12.8 million (note that these figures are illustrative). By some estimates, that’s more than a quarter of what M-PESA generated in profits on a standalone basis. 5

7 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? How Significant are Churn Reduction Benefits to Profitability? If you’ve ever attended a mobile money conference, Some services report a less ones described above. dramatic reduction in churn; some report no change you’ve likely heard a speaker tout the potential in churn; and some even report a slight temporary benefit of ‘reduced churn’ that mobile money can unlock for an MNO. But what you probably haven’t increase in churn. This variance underscores an important message for MNOs that launch mobile heard is whether any service has actually delivered money services on the basis of potential for churn on this promise – and if so, whether the subsequent reduction: the benefits are real and attainable, but benefits amount to a big or small deal in the overall financial model. only for those who execute effectively. That is, the services that have not realised any churn reduction benefits are those that have registered customers In our analysis of MTN Uganda’s MobileMoney, with no real interest in the service, or been plagued a service that has turned the corner into cash-flow by bad customer experiences, poorly planned agent positive territory on a month-to-month basis, we in any given month, uncovered a startling finding: networks, and half-hearted attempts at creating a the churn rate for active mobile money customers strong brand and relevant service offering. It’s easy to see, then, why executives in some countries have gone That is, while the churn rate for regular is negligible. as far as charging internal transfer pricing premiums mobile customers was roughly 4.5% per month, the churn rate for an active mobile money customer to their mobile money business units, reasoning that a poorly executed foray into financial services will do was no more than 0.2% over the course of the three nothing more than jeopardize existing relationships months for which we analysed data. with valuable mobile customers. Exhibit 5: Churn Comparison (MTN MobileMoney Uganda) So what does this mean for a mobile money Active MobileMoney Customers practitioner? First, it means that execution is Registered MobileMoney Customers everything. The promise of ‘reduced churn’ has MTN Customers (excluding active and registered MobileMoney customers) 5% been realised – but only by deployments that are 4.5% well funded and have executed effectively. 4% 3.5% Second, without considering the benefits of reduced 3% churn, the profitability picture is incomplete. Today, Monthly Churn 2.5% many MNOs choose to exclude churn benefits from 2% their P&L or business plan: some do so because 1.5% executives are sceptical about whether variances 1% stem from ‘causation’ or ‘correlation’; others reason 0.5% that if this service is to be sustainable, it must be on 0% the basis of direct benefits alone. The latter rationale Month-1 Month-3 Month-2 is prudent, but when capital budgeting season arrives and executives start to ask for IRR figures, it behoves practitioners to have these figures at hand. This is a dramatic reduction, but the question remains: does it make much of a difference to the the significance of churn reduction And finally, overall profitability of the service? In the case of benefits underscores the importance of tracking the MobileMoney, the answer is a resounding yes. Of For practitioners to gauge whether the right metrics. the total revenue generated to date, churn reduction service is moving the needle on churn, they must first benefits account for 33% – and if the service wasn’t have a process established, usually one in which an delivering this benefit, MobileMoney would have external data warehousing team is engaged, to track barely been out of the red by now. In other words, the the metric. This can be time consuming, but given benefit of reduced churn matters – a lot. the potential importance of this metric, it’s clearly worthwhile. not every service we’ve Alas, there is one catch: studied has generated results as impressive as the 6

8 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? How Significant is ARPU Uplift to Profitability? this benefit is real. In the case of MTN Uganda’s The previous chapter described how significant MobileMoney, active customers do consume slightly churn reduction benefits can be to the profitability more voice and SMS than non-mobile money case for mobile money; and when people talk about customers, but drawing a solid conclusion here indirect benefits, ‘reduction in churn’ is usually would be incredibly challenging from a data-mining closely followed in the same sentence by ‘uplift in perspective. ARPU’ (Average Revenue Per User). Having shed light on the important role churn benefits can play in the context of profitability, in this chapter we’ll focus While we haven’t conclusively pinpointed the impact of “increased share of wallet for voice and SMS” in a on the role of uplift in ARPU. But before we answer the question posed in the title of this chapter, let’s financial model, it’s plain to see that the potential to first determine whether ‘uplift in ARPU’ is even the reap benefits is massive – and there are some steps MNOs can take to position themselves to do so. right metric for practitioners to measure. To gauge whether mobile money actually causes Beyond executing well to ensure customers do indeed have an incentive to keep their mobile money SIM in customers to spend more, ‘uplift in ARPU’ would the phone more often than not (a subject I discussed need to be measured over time. But this particular in the previous chapter), promoting mobile money type of analysis is tricky. First, the average selling as a method of topping up is also important. In price for airtime and SMS, and therefore ARPU, in a country varies for any number of reasons on a particular, MNOs have found success by promoting month-to-month basis, so it’s impossible to simply mobile money as an option for topping up in small attribute any change solely to mobile money. Second, increments, and topping up after hours when For instance, scratch cards may be unavailable. in many cases ARPU figures will already include WING, a Cambodian mobile money service, has revenue generated from mobile money – so taking credit again would be inaccurate. enjoyed success with their mobile top-up feature, and found that 33% of top-ups on their system occur outside typical store hours, and 70% occur at the It’s clear, then, that ‘uplift in ARPU’ isn’t a perfect US$1 price point, a level at which scratch-cards are a metric. But what, if anything, is? We propose that particularly expensive as a distribution option. the less catchy, but somewhat more accurate, phrase of “increased share of wallet for voice and SMS” is the more relevant metric. By measuring ‘minutes of use’ and ‘billable SMS events’, an MNO can isolate changes in customer behaviour, something that’s not possible with an ‘uplift in ARPU’ calculation. Additionally, “increased share of wallet” accurately describes just why a mobile customer might consume more mobile services on their mobile money SIM; that is, it’s easier to imagine a customer who carries two SIM cards, each month spending $3 on one, and $2 on the other, shifting some of her spending to the stickier of her two SIMs. So if we accept “increased share of wallet for voice and SMS” as a good metric, the question still remains: is it a significant driver of profitability? Unfortunately, our findings in this department are inconclusive. From a survey conducted in 2009 by McKinsey & Co., CGAP and GSMA, we know that in the Philippines 44% of mobile money users carry more than one SIM, and 68% report using their mobile money SIM as their ‘primary SIM’; this is encouraging, but not conclusive evidence that 7

9 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? How Significant are Direct Revenues to Profitability? Exhibit 6: % of transactions So far in this article, we’ve written about the role that Exhibit 7: to unregistered customers Money Transfer Revenue indirect benefits play in enabling a mobile network operator (MNO) to turn a profit from a mobile money age volume of tr ansfer to registered customers Aver ansfers to unregistered recipients Revenue from tr Aver age volume of tr ansfer to registered customers ansfers to unregistered recipients Revenue from tr Revenue from tr ansfers to registered recipients Aver ansfers to unregistered recipients age volume of tr Revenue from tr ansfers to registered recipients Aver age volume of tr ansfers to unregistered recipients service – but what about the most obvious source of value, direct revenue from customer fees? After all, this is often the single source of value upon which MNOs evaluate the business case for mobile money. 38% 38% 44% 44% 56% 56% For MTN Uganda, who currently offer domestic 62% 62% money transfer and mobile top-up services, direct revenues include fees to send money, and fees to To date, these withdraw money from an e-wallet. direct revenues, less commissions paid to agents, contribute 52% of total gross profit for the service. But not every MNO allows P2P transfers to It’s clear then, that this is an area of the business case unregistered recipients: some reason that by doing so, not to be neglected. So how can MNOs ensure they’re they are forfeiting potential net new mobile revenue well positioned to fully capture this source of value? from recipients who, if they want to receive money, Well in the case of MTN Uganda’s MobileMoney, one have no choice but to activate a SIM from the MNO decision has had more of an impact than any other: in question (and then, as the theory goes, start to use enabling P2P transfers to unregistered recipients. this new SIM for mobile services, too). But this walled garden logic is risky: mobile money is a service that Uganda is a fragmented mobile market: according to is predicated on network effects, and particularly in Wireless Intelligence at time of writing, MTN holds countries with fragmented mobile market share, the 44%, Zain, Warid and Uganda Telecom each hold ‘closed model’ presents an insurmountable customer roughly 18%, and Orange holds 3% market share. So experience barrier to adoption, ultimately making it it’s not surprising, then, that when MTN launched difficult to scale the mobile money service. And if a the service, they made sure customers could send mobile money service cannot scale, its sustainability funds to recipients on any network. To date, 38% becomes questionable – so in the end, any benefits of of P2P transfers made using MobileMoney have net new revenue will be short lived. been from a registered customer to an unregistered recipient; and this use case has generated 45% of It’s clear, then, that direct revenues are a significant total revenue (and even more in gross profit). Two value source, and mobile network operators have the overall things are striking about this data: first, an opportunity to maximize them by enabling P2P number of P2P transfers to unregistered users is transfers to unregistered recipients – a feature that, quite high, which suggests that had MTN not offered coincidentally, is just what customers in Kenya, this option, they likely would have left some revenue Uganda and other successful mobile money countries on the table. Second, P2P transfers to unregistered have demonstrated that they want. users are more lucrative for MTN than P2P transfers to registered users (i.e. 38% of transactions are generating 45% of revenue). This occurs because MTN charges customers a premium – 7% for low and 94% for highest value transfers – to make a transfer to an unregistered recipient, and the commission paid to agents remains the same. Thus, by enabling P2P transfers to unregistered recipients, MTN not only expands the base of potential users for their service, they also generate a significant amount of revenue. 8

10 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? How can an MNO Manage Costs to Achieve Profitability? When most people hear the phrase “to turn a profit, still applicable: these activities are routinely touted we need to manage our costs”, they usually take it to as strategic imperatives for any successful mobile mean “to turn a profit, we need to reduce our costs”. money service – but for MTN, they’ve cost a pittance 7 But when it comes to mobile money, practitioners compared to the amount spent on technology (30%) 8 have found that some costs can be done away (12%) to date. or customer registration commissions with more easily than others. So the trick, then, is to understand which are strategic (and must be Exhibit 8: Detailed breakdown of costs protected), and which are discretionary (and can be curtailed). Agent network costs Handset subsidies POS merchandising field mark eting agency contr act 12% 12% Agent network costs ay upgr allet solution SIM access gatew Up-front investment in m-w ade recurring fees But before we begin our evaluation process, let’s first 8% for m-w ade maintenance allet solution recurring fees for SIM access gatew ay upgr ansfer price) SMS communication fees (internal tr briefly take stock of the costs (and there are many) that 3% om discounting ARPU loss fr 28% are involved in launching a mobile money service. ate Total airtime bought through mobile money at a discount multiplied by discount r Before launch, MNOs incur a series of technology Selling expense Mark eting and advertising 30% costs, including investing in an m-wallet platform, 7% G&A upgrading their SIM or USSD access gateway (in Management staff back office staff most cases), and deciding whether to embed their Cash-in/cash-out commissions ansfer commissions agent airtime commissions Money tr application on all new SIMs – and in most cases Agent network costs act Handset subsidies POS merchandising field mark eting agency contr ation costs Customer acquisition & r egistr 12% consequently upgrade to a larger card (while this isn’t 12% ation collater al registr ation commissions SIM sw Customer registr ap cost incremental ading to larger card SIM cost for upgr Agent network costs a cash outlay at first, it’s a decision of major financial ay upgr allet solution SIM access gatew Up-front investment in m-w ade recurring fees 8% allet solution recurring fees for SIM access gatew for m-w ay upgr ade maintenance significance). The next tranche of costs are go-to- ansfer price) SMS communication fees (internal tr 3% market related, and include recruiting and paying ARPU loss fr om discounting 28% for management and back-office staff, training and ate Total airtime bought through mobile money at a discount multiplied by discount r merchandising a network of agents, and designing Selling expense and launching above and below-the-line marketing Mark eting and advertising 30% campaigns. Most of the costs identified thus far carry 7% G&A on after the service has been launched, but the day a Management staff back office staff service goes live, a third set of costs come into play: Cash-in/cash-out commissions ongoing costs. These typically include cash-in/cash- Money tr ansfer commissions agent airtime commissions out commissions for agents, SIM cards, starter packs egistr Customer acquisition & r ation costs ation collater al registr Customer registr ap cost incremental ation commissions SIM sw and agent registration commissions for customer SIM cost for upgr ading to larger card acquisition, and internal transfer fees for using SMS services or selling airtime at a discount. For a full breakdown, refer to Exhibit 8. So if these activities deliver such good value for money, why do some practitioners have a difficult So which of these are strategic and which, if any, are time getting budget to do them properly? In many discretionary? Unfortunately, answering this question highly strategic, cases, this stems from the fact that is not as simple as sorting costs according to size. If we financially insignificant costs often require a look at the drivers for MTN Uganda’s MobileMoney, commitment to spend in advance of having any highly strategic operational activities we find that indication of whether the mobile money service – things like building and managing an agent will be a success. For instance, MTN had to commit network, or providing great customer care – are to a fixed monthly contract with a field marketing Since launch, 7% of comparatively inexpensive. agency ($623,000); pay for and train their dedicated MobileMoney’s total costs have been on building call centre representatives ($440,000); and design 5 and managing their agent network , and 4% has and fund an above-the-line marketing campaign 6 And while it’s been on back-office customer care. ($850,000) all prior to launching their service. Each true that Safaricom spends somewhat more on these of these activities has been instrumental in MTN particular activities, and has benefited from an agent Uganda’s success, and their decision to invest network of industry leading quality, the insight is aggressively in them ultimately stemmed from their confidence that the service would become a hit. 5 Includes handset subsidies, agent POS merchandising, and field marketing agency costs 9 6 Includes total cost of back-office staff 7 Includes cost of m-wallet platform and monthly charges, SIM access gateway upgrade and monthly maintenance charge, and SMS communication fees 8 Includes commissions paid from MTN to agents ($1.33 per registration).

11 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? But that’s not to say all of their spending has been strategic; some costs were discretionary, and potentially could have been substituted for less expensive, equally effective alternatives. For instance, MTN recently introduced an airtime bonus for customers who top-up using mobile money, an incentive many MNOs have used in an effort to encourage customers to top-up using their e-wallet. But this tactic was particularly costly since it negates MTN’s decision to invest aggressively in a big portion of the savings realised from eliminating marketing, agent monitoring, and call centre discounts paid to dealers. staff ultimately stemmed from their confidence that the service would become a hit Moreover, like they have in other markets, MTN has pursued a strategy of aggressively registering new customers in Uganda. In practice, this has meant registering more inactive customers (552,213) than active ones (421,254). And this strategy has been expensive: MTN has spent a total of $1.3 million on registration commissions and new SIM cards for customers that have not performed a single revenue- generating transaction.. 10

12 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? How can MNOs Ensure Their Tariff and Commission Models are Well Designed? For a mobile money service to scale and achieve It also merits note that MTN Uganda’s customer profitability, it’s critical to have well designed tariff model grants customers minimal leeway to customer tariff and agent commission models. So defraud the operator of prospective direct revenues. how can MNOs ensure their tariff and commission That is, given that the P2P transfer fee typically models are well designed? Here again, MTN accounts for less than half of the total end-to-end customers cost of sending money using the service, Uganda’s MobileMoney exemplifies some key insights. have little incentive to perform a direct deposit. Moreover, MTN has structured its tariff tiers in such If the MobileMoney customer tariff model looks a way that there is no opportunity for a customer to familiar to you, that’s probably because you’ve reduce their fees by splitting a cash-in or cash-out into multiple smaller tranches. seen it in action before: in structure, it’s a replica of Safaricom’s M-PESA. And as Ignacio Mas noted in the 2009 Mobile Money for the Unbanked Annual But it’s not just MTN’s customer tariff model that report, this tariff structure (and the way it’s taken merits attention. Their agent commission model has been thoughtfully designed, too. The article Neil to market) works for a few reasons: it’s simple and transparent, customers are not bound by minimum Davidson and I wrote for the 2010 Mobile Money for the Unbanked Annual Report details most of our balance requirements or prohibitive deposit fees, thinking on agent incentives, but it’s worth briefly and it offers customers an ability to send money 9 It’s inevitable that MNOs will to non-customers. noting here how MTN espouses some key principles. innovate and trial new models, but the design MTN pays MobileMoney agents a commission First, features listed above can be considered prerequisites for every activity that they perform, even though for an effective tariff model in any environment. MTN may not charge customers a fee directly for MTN MobileMoney Rates For instance, even though MTN doesn’t each one. charge customers a fee to cash-in, they do provide Charge Activity Transaction Tiers agents a commission for providing this service in recognition of the time and cost involved. Of course, Max Min while MTN take a temporarily hit by subsidising Loading Money 0 1,000,000 5,000 cash-in, the fees collected from an end-to-end money Sending Money transfer (which includes a cash-in, a transfer, and a To registered user 800 1,000,000 5,000 cash-out) do exceed the corresponding commissions 1,600 30,000 5,000 All told, the margin MTN earns for a typical paid. 30,0001 60,000 2,000 end-to-end P2P transfer (excluding variable 125,000 60,0001 3,700 To non-registered user technology fees) to a registered customer is just MTN or Local network north of 50%. 250,000 125,001 7,200 10,000 500,000 250,001 Second, while MTN may pay agents for both cash- 1,000,000 500,001 19,000 they deliberately pay a higher in and cash-out, Withdrawing commission to agents for facilitating cash-out than 700 5,000 30,000 they do for cash-in. This stems from the simple fact 1,000 60,000 30,001 that ‘cash-out’ agents have a higher cost of restocking 60,001 125,000 1,600 their inventory of physical cash than cash-in’ agents By registered user 125,001 250,000 3,000 do for restocking their inventory of e-money. As such, 500,000 5,000 250,001 ‘cash-out’ agents must be compensated accordingly. 500,001 1,000,000 9,000 By non-registered user 0 5,000 1,000,000 MTN or Local network 0 Buying airtime 5,000 1,000,000 Daily transaction limit UGX 1,000,000 9 Ignacio Mas: Good Service Design Features of M-PESA’s Money Transfer Service. 2009 MMU Annual Report. 11

13 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? Third, MTN recognized that to keep agents engaged in the period following launch when transaction volumes are typically low, it would be important to provide them with a different source of revenue. To this end, they have provided agents with a commission for every customer that they 10 Thus, in the early days following launch, register. MobileMoney agents earned money by registering customers; as the service scaled they increasingly earned their money from facilitating cash-in and cash-out transactions for customers. 10 12 This decision also helped drive customer growth.

14 Mobile Money for the Unbanked Is There Really any Money in Mobile Money? Appendix A: Resources GSMA Financial Model The GSMA Financial Model is an excel tool that practitioners can use to develop a comprehensive view of the profitability of their mobile money service. The model generates a P&L statement that is based on a series of user inputs, including investment, direct benefits, indirect benefits and costs. GSMA Metrics Dashboard The GSMA Metrics Dashboard is an excel tool that presents practitioners with an easily digestible summary of their operational metrics that matter most. Existing and future customers of Comviva, Fundamo, Sybase 365 and Utiba can integrate the Dashboard as a reporting feature free of charge. To receive a copy of the GSMA Financial Model or the GSMA Metrics Dashboard, send an email to [email protected] 13

15 For further information please contact [email protected] GSMA London Office T +44 (0) 20 7356 0600 This article was originally published in October 2010 as a series of blog posts on the GSMA Mobile Money for the Unbanked website.

Related documents