Why M-Pesa failed in South Africa

Yes, and by its owner's own test. Vodacom and Nedbank launched the M-Pesa money transfer service in South Africa on 31 August 2010, announcing a "mobile money revolution"^{[Vodacom media release, checked 2026-11-04]}. The bet was that a transfer service built for phones would find the same pull here that M-Pesa had found elsewhere. It did not. Following a thorough review, Vodacom announced on 9 May 2016 that it would discontinue M-Pesa in South Africa with effect from 30 June 2016^{[Vodacom media release, checked 2026-11-05]}.

The numbers in between explain why. Vodacom's 2015 integrated report counted one million registered M-Pesa customers in South Africa and 76,000 people actively using the service, an uptake it called "still relatively modest"^{[Vodacom, checked 2026-11-05]}. Registration was never the problem. People signed up, tried it, and kept moving money the way they already did. A payment network lives on the people who use it every week, and in South Africa that group stayed small.

This page sets out who said what went wrong, and which source supports each cause. Vodacom's own explanation comes first, then MTN's and FinMark Trust's, then what all of it means for anyone building payments in South Africa today. For the full timeline, from the first launch through the shutdowns, the relaunches and the e-money rules now in draft, read the history of mobile money in South Africa. For how the country's payment rails fit together today, start with how payments work in South Africa.

The announcement was unusually direct. Shameel Joosub, Vodacom's chief executive, said: "Vodacom's decision is based on the fact that the business sustainability of M-Pesa is predicated on achieving a critical mass of users. Based on our revised projections and high levels of financial inclusion in South Africa there is little prospect of the M-Pesa product achieving this in its current format in the mid-term."^{[Vodacom media release, checked 2026-11-05]}

Three phrases in that statement carry the argument.

"Critical mass of users." A person-to-person payment network is only useful if the people you want to pay are on it. Each new active user makes the network more useful to everyone else, and the reverse is true as well. Below a certain density, every payment needs the other person to sign up first, so most people never bother. On Vodacom's revised projections, M-Pesa in its current format had little prospect of crossing that line in the mid-term.

"High levels of financial inclusion." This is the heart of Vodacom's case, and it is about the market, not the product. Vodacom did not say the product was bad or the marketing weak. It said South Africans were already well served financially, which left M-Pesa competing with services people already had rather than filling a gap.

"In its current format." Vodacom limited its verdict to the product as it stood, and to the mid-term. It came back later with a super app rather than a transfer service, which we come to below.

Three sources speak to the failure, and they stress different things. Vodacom explained its own decision. MTN gave a short reason when it decided to close its service. And FinMark Trust published its research report on mobile money in South Africa in 2017^{[FinMark Trust, checked 2026-11-05]}, after both shutdowns. Here is each cause next to the source that names it.

Cause	Named by	What the source says
Too few users to sustain the business	Vodacom	M-Pesa's sustainability "is predicated on achieving a critical mass of users"[Vodacom media release, checked 2026-11-05]
High levels of financial inclusion	Vodacom	with revised projections and high financial inclusion, "little prospect" of critical mass "in its current format in the mid-term"[Vodacom media release, checked 2026-11-05]
Lack of commercial viability	MTN, of its own service	MTN decided in 2016 to decommission Mobile Money due to lack of commercial viability[MTN SA press release, checked 2026-11-05]
Agent network challenges	FinMark Trust	including cash float, aggregation and trust[FinMark Trust, checked 2026-11-05]
Poor technology choices	FinMark Trust	choices made for the implementation of the services[FinMark Trust, checked 2026-11-05]
A constraining institutional environment	FinMark Trust	it constrained the services that mobile money could offer[FinMark Trust, checked 2026-11-05]
Established competition in domestic remittances	FinMark Trust	it limited growth in the traditional cornerstone of mobile money[FinMark Trust, checked 2026-11-05]

FinMark then sums up where those factors left mobile money: with no significant reason for the banked, the unbanked or the underbanked to switch from competitor offerings^{[FinMark Trust, checked 2026-11-05]}. The report introduces its list as additional factors that influenced the lack of sustainability^{[FinMark Trust, checked 2026-11-05]}, so read the FinMark rows as the causes it names, not as everything it found.

All three sources describe a business that could not sustain itself. They differ on why: Vodacom points to the market, MTN to the economics, and FinMark to a mix of the market, the institutions and the operators' own choices. The sections below take the causes one at a time, starting with Vodacom's.

Because mobile money's strongest pitch is to people without a bank account, and Vodacom said high levels of financial inclusion in South Africa left M-Pesa little prospect of critical mass^{[Vodacom media release, checked 2026-11-05]}. That is the company's own reading of its market, not an outside verdict.

Think about what a mobile money wallet asks of a new user. You walk to an agent, hand over cash, and get a balance on your phone that you can send to someone else on the same service. For a person with no bank account, that is a large step up from carrying cash. For a person who already has a bank card, a banking app and a salary paid into an account, the same steps add work. Their money already sits in a bank. Moving it into a wallet means taking it out first.

That is why Vodacom's statement pairs financial inclusion with critical mass. Where most people already bank, the pool of people for whom a wallet is clearly better is small, and a small pool cannot produce a dense network. A wallet user also needs the people they pay to be on the wallet, and in a well-banked market most of those people have no reason to join.

There is a lesson here that reaches well beyond M-Pesa. A payment method wins when it is the easiest way to pay the people you need to pay. Where those people are already reachable through their banks, a non-bank wallet has to beat the banks at their own game rather than offer something they cannot.

If you are deciding which rails to build a South African product on today, that is still the question to ask, and Mctaba Labs payment integration can help you answer it for your own customers.

Established competitors, according to FinMark Trust. Its report found that established competition in domestic remittances limited the opportunity for growth in the traditional cornerstone of mobile money^{[FinMark Trust, checked 2026-11-05]}. Domestic remittances are the money people send to family and others inside the country, and they are the plain first job a mobile money network usually grows from. Paying bills, buying airtime and paying merchants come later, on top of a network that already has a reason to exist.

When that first job is already taken, the network struggles to get started. FinMark's summary, quoted above, is where its account reaches further than Vodacom's: it covers the banked, the unbanked and the underbanked alike. Financial inclusion explains why banked people had little reason to switch. FinMark found no significant reason for the unbanked or the underbanked to switch from competitor offerings either^{[FinMark Trust, checked 2026-11-05]}, so even people without bank accounts had something they preferred to keep using.

The competition for that job looks no weaker today. The Shoprite Group says customers with Money Market Accounts can make free, instant transfers to other account holders through the app, WhatsApp, the Money Market website or USSD at any time of day^{[Shoprite Holdings newsroom, checked 2026-11-05]}. That is a retailer offering, at no charge, the job a wallet would most want to own. Our page on retailer money transfers covers how the shop counter competes with wallets.

For a builder, the general rule is simple: a new payment method needs a first job that nobody already does well. Find that job before you try to build the network.

Because an agent network only pays for itself when customers keep coming back, and in South Africa too few did. FinMark names agent network challenges, including cash float, aggregation and trust^{[FinMark Trust, checked 2026-11-05]}. Each of the three is a cost that only steady traffic can cover.

Float. An agent is a shop or kiosk that takes cash in and pays cash out on behalf of the wallet. To serve customers it has to keep both cash and wallet balance on hand, known as float, and that float is money the shop cannot spend on its own stock. An agent who sees a handful of customers a day earns too little commission to justify tying it up, so it stops promoting the service or drops it.

Aggregation. Aggregation is the layer above the shops: a larger business that rebalances cash and e-money for many small agents in return for a share of the commission. When volumes are thin, there is not enough commission to share, and the layer that keeps small agents supplied is hard to sustain.

Trust. A customer hands cash to a shopkeeper in exchange for a number on a screen, and has to believe the next agent will have the cash to pay it out. Every failed cash-out teaches people not to rely on the wallet.

Now run the loop. Few active users means little agent income. Little agent income means fewer and less reliable agents. Fewer reliable agents means a customer who wants to cash out cannot count on finding one nearby, so the customer stops using the wallet. Fewer active users means even less agent income. Mobile money in markets where it thrives runs the same loop in the other direction, with each turn adding users and agents.

Banked customers also had a way to get cash that needed no agent at all: their bank. That brings the argument back to Vodacom's point about financial inclusion.

FinMark also lists poor technology choices for the implementation of the services^{[FinMark Trust, checked 2026-11-05]}. The passage we rely on does not say which choices it means, so we do not guess at what Vodacom or MTN got wrong. What we can do is show where technology decides whether a wallet gets used, because every team building a payment product makes these choices.

• The channel. USSD menus, SIM menus and smartphone apps reach different people. A channel that does not work on the phones your customers carry loses them before the first payment.
• The steps. Every extra screen, code or confirmation between "send" and "received" costs completed payments, and a slow confirmation teaches people to use something else.
• The agent tools. Cash-in and cash-out are where the wallet meets the shop. If recording them is slow or error-prone, agents serve fewer customers and trust wears thin.
• The bridge to banks. A wallet whose balance cannot move easily to and from bank accounts is an island, and an island in a market with high financial inclusion stays small.

Of all the causes on this page, this is the one most fully under a builder's control. Financial inclusion and the institutional environment are given; the channel, the flow and the reconciliation are yours to get right. Our page on mobile money, bank apps and wallets shows what South Africans actually use, which is where any channel decision should start.

FinMark gives this one as a summary: the institutional environment constrained the services that could be offered through mobile money^{[FinMark Trust, checked 2026-11-05]}. The words are broad, and the passage we rely on does not list the constraints one by one. Several documented facts show what that environment looked like from an operator's side. Read them as context for FinMark's point, not as its finding.

The first is the shape of the launch. M-Pesa did not arrive in South Africa as a Vodacom product alone, but as a joint launch by Vodacom and Nedbank^{[Vodacom media release, checked 2026-11-04]}. Vodacom's second attempt has a bank behind it too: Vodacom's app listing says the VodaPay Wallet is powered by Bidvest Bank^{[VodaPay app listing by Vodacom, checked 2026-11-05]}.

The second is how the payment system admits newcomers. In October 2026 the SARB Governor said: "At the moment, anyone who wants to access the payment system has to go through a bank."^{[BIS: Kganyago, checked 2026-11-05]} The formal change has only just begun: on 20 May 2026 the Prudential Authority published its intention to issue an exemption notice to exempt specified payment activities^{[SARB Prudential Authority, checked 2026-11-05]}.

When a bank and a mobile operator run a wallet together, the bank stands behind the customer's money and the operator brings the distribution. That split sounds sensible, but it shapes the incentives. A bank has its own accounts and cards to sell, and a wallet that succeeds too well competes with them. The operator gives up part of the economics and part of the control. Decisions one owner could take in a week take longer when two institutions with different goals have to agree. The record does not show that this is what held M-Pesa back, but it is the kind of constraint a system built around banks places on any wallet.

That matters for anyone who reads M-Pesa's exit as a simple product failure. A telco that wants to run a wallet in South Africa still does it beside a bank, in a market where banks are the incumbents it needs to beat.

Both, and the record lets you tell them apart. Start with the strongest evidence that the market mattered: a second operator gave up in the same years. MTN launched its mobile money platform in South Africa in 2012 and decided in 2016 to decommission its Mobile Money offering, due to lack of commercial viability^{[MTN SA press release, checked 2026-11-05]}.

Two operators, different launch years, different teams, and the same decision in the same year. If M-Pesa had failed alone, you could put it down to Vodacom's execution: the wrong pricing, the wrong partner, the wrong advertising. When MTN's service went the same way, the explanation both shared, the market they launched into, becomes the more convincing one.

Now sort the causes. The financial inclusion Vodacom cited and the established competition in domestic remittances FinMark found describe the market. FinMark's constraining institutional environment describes the rules and institutions around it. Agent network challenges sit in between: the market set the traffic, but building and paying an agent network is an operator's job. Poor technology choices sit with the operators.

Vodacom's own words concede the split. "In its current format" admits the product could have been built differently. But no amount of better execution would have changed the financial inclusion Vodacom described or the competition FinMark found. On the evidence, the market made mobile money hard here, and the operators' choices did not make it easier.

The usual answer is that Kenya offered the conditions South Africa lacked, and part of that comparison can be documented. Safaricom's own company history gives the launch of M-Pesa as 1 March 2007^{[Safaricom, checked 2026-11-04]}. For the banked side, FSD Kenya's analysis of the 2006 FinAccess survey found that 13.7% of Kenyans surveyed used a bank savings account and only 2.1% used bank credit^{[FSD Kenya, checked 2026-11-05]}. Set that beside the "high levels of financial inclusion" Vodacom described in South Africa^{[Vodacom media release, checked 2026-11-05]}, and the contrast in starting conditions is clear, even without a matching South African figure.

The other half of the usual comparison, operator market shares at the time, we leave out on purpose: we have not been able to confirm those figures against a primary source, and quoting them second-hand is how wrong numbers spread.

The argument does not need them. Mobile money spreads when a large share of the people you need to pay have no better way to receive money, and when enough of them are on one network that paying by phone number becomes the default. South Africa's documented conditions point the other way: Vodacom saw high levels of financial inclusion^{[Vodacom media release, checked 2026-11-05]}, and FinMark found established competition in domestic remittances, a constraining institutional environment and agent network challenges^{[FinMark Trust, checked 2026-11-05]}.

The two markets also diverged in what came next. In Kenya, developers build on Safaricom's APIs through Daraja, Safaricom's developer portal^{[Safaricom Daraja developer portal, checked 2026-11-05]}, which our M-Pesa hub covers. In South Africa, paying someone without their account details ended up inside the banking system: PayShap lets people pay to an alias instead of a bank account^{[SARB retail CBDC background note, checked 2026-11-04]}. Our comparison of PayShap and M-Pesa sets the two designs side by side.

They changed the product rather than retrying the old one. On 18 December 2019 MTN announced that its MoMo mobile money service would go live in South Africa in January 2020^{[MTN Group media release, checked 2026-11-05]}. Vodacom announced in 2021 that its VodaPay super app was live^{[Vodacom media release, checked 2026-11-05]}.

VodaPay is the clearer change of format. A super app is built to give people many reasons to open it, rather than relying on person-to-person transfers to carry the whole network. That matters in a market where, as FinMark found, domestic remittances already faced established competition^{[FinMark Trust, checked 2026-11-05]}. It is also the kind of different format Vodacom's closing statement left room for. MoMo, for its part, came back after MTN's first service had closed, which tells you MTN still saw a market worth a second attempt.

Whether either has solved the active-user problem is a question we cannot answer with verified numbers, so we do not try here. The MTN MoMo South Africa and VodaPay pages cover each product, and our page on registered versus active mobile money users explains why the headline number in any wallet announcement deserves a second look.

The institutional side is starting to move; the market side has not. The May 2026 draft Authorisation Framework makes issuing e-money and payment instruments a group of payment activities in its own right^{[SARB, checked 2026-11-05]}, the Prudential Authority has published its intention to exempt specified payment activities^{[SARB Prudential Authority, checked 2026-11-05]}, and the SARB intends to publish the Authorisation Framework by the end of November 2026^{[BIS: Kganyago, checked 2026-11-05]}. The Governor named opening up the payment system to non-banks first among the main building blocks of a better payment system^{[BIS: Kganyago, checked 2026-11-05]}.

If the final framework follows the draft, that speaks to the institutional environment FinMark described, at least for who may issue e-money. On our reading it does little about the rest. The banking system has moved on since M-Pesa closed and now has its own answer to paying someone without their account details: PayShap's initial rollout began on 13 March 2023^{[SARB media statement, checked 2026-11-05]}, and it lets people pay to an alias instead of a bank account^{[SARB retail CBDC background note, checked 2026-11-04]}. The feature that made M-Pesa feel new when it launched is now part of the banking system itself. The banks' reach is large, too: Capitec alone reports more than 26 million active clients^{[Capitec results, checked 2026-11-04]}.

So the honest verdict is that a straight copy of M-Pesa would struggle again. A new wallet would need a job the banks do not do well, such as serving people who earn and spend in cash, cross-border transfers within the region, or a merchant network the banks have not reached. It would also need agents and technology that work from the first day, the causes on FinMark's list that a company can do most about. Our page on mobile money under the e-money tier looks at what the new rules make possible.

Five lessons for anyone building payments here.

1. A rail that won elsewhere does not win by default. M-Pesa's success in other markets came from local conditions. Before you build on a payment method, check that the conditions behind its success exist where your customers are.
2. Build on what your customers already use. In South Africa that means bank rails first: cards, PayShap, Capitec Pay and bank transfers. A checkout that asks people to open a new account before they can pay you loses them at the worst moment.
3. Count active users, not registrations. One million registrations and 76,000 active users^{[Vodacom, checked 2026-11-05]} are two very different businesses. Ask any wallet or provider for active numbers before you plan around its reach.
4. Get the parts you control right. Poor technology choices are on FinMark's list^{[FinMark Trust, checked 2026-11-05]}. The channel, the number of steps and the reconciliation between a wallet and the bank are decisions, not conditions, so make them deliberately.
5. Serve each market on its own rails. If you build for Kenya and South Africa together, integrate M-Pesa for Kenyan customers and the South African bank rails for South African ones, rather than forcing one market's habits on the other.

For the wider region, where mobile money plays a very different role from country to country, see payments in Southern Africa. And if you want a South African payment stack built around what your customers actually use, Mctaba Labs payment integration does that work.