Africa’s largest mobile-money franchise didn’t out-app the market—it out-operated it. Any phone, patchy data, cash-heavy retail: USSD plus agents turns constraints into scale. For product teams, that’s the real moat.
The product model: built for constraints, not theory
Facts as of 26 Sep 2025.
M-Pesa is a SIM-tied stored-value wallet that runs on USSD menus with SMS receipts and, where it makes sense, a smartphone app. Core rails cover P2P, cash-in/out, bill pay (PayBill), and merchant payments (Lipa na M-Pesa), with add-ons like savings, micro-credit, and remittances depending on the market. According to company statements and regulator datasets, usage skews toward short, frequent transactions where speed, reliability, and zero-data access matter most.
Why USSD still wins
Any GSM handset works. No data bundle required. Sessions are fast and resilient even when coverage is spotty. In markets where many people still carry feature phones and treat data as scarce, USSD is the inclusive default—not a legacy afterthought. The “works-everywhere” factor turns first-time trials into habits.
The agent network: the offline distribution moat
A dense cash-in/cash-out grid—kiosks, corner shops, petrol stations—keeps liquidity close to where people live and trade. That shrinks travel time, shortens queues, and makes cash conversion mundane rather than exceptional. Industry reporting and company materials point to hundreds of thousands of agents across the footprint; in Kenya alone, (all providers combined) agent counts are in the hundreds of thousands. When data drops or a phone is basic, a nearby agent plus a USSD code still completes the job. Reliability in the “last 500 meters” is what compounds trust.
Commissions and trust: economics users accept
The fee design is simple: cash-in is free; cash-out and certain transfers are tiered. For merchants, Lipa na M-Pesa shifts more cost to the acceptance side, nudging in-store digital payments without taxing every consumer flow. According to company statements, M-Pesa remains a major revenue engine; that revenue funds agent commissions, training, and float management—so agents stay motivated and liquid. Predictable tariffs → successful first experiences (USSD works; agent nearby) → word-of-mouth → repeat use. Not flashy UI—operational reliability—drives daily active use.
Constraints and the road ahead: USSD fades slowly, not suddenly
- Friction. USSD sessions can time out and menus can be deep. Apps and QR improve UX when data is stable, but they don’t erase the basic-phone reality overnight.
- Interoperability & pricing. Regulators push fairer cross-network flows; providers juggle margin and inclusion. Tariffs evolve, yet the core “free in, paid out” pattern persists.
- Scale & resilience. Back-end upgrades lift throughput and fraud controls, but day-to-day NPS still hinges on agent liquidity, dispute resolution, and transparent receipts.
What this means for you (builder or merchant)
- Design USSD-first, then layer app features where they deliver clear wins.
- Invest in agents: incentives, float, training, queue management.
- Price against the real cash cost: compare fees to bus fares, queue time, and theft risk—not just to “zero.”
- Plan for offline: SMS receipts, robust reversal flows, and clear fallbacks when data dies.
By the numbers (Kenya-centric)
- Value: tens of trillions of KES annually; volume: tens of billions of transactions per year (company/regulatory data).
- Distribution: hundreds of thousands of agents in Kenya (all providers); ~600k+ across the wider footprint (company claims).
- Devices: smartphone adoption rising, but a large base remains on feature phones—sustaining USSD usage.
Facts are based on company statements, regulator datasets, and industry reporting.