Most operations conversations about iGaming payments in Africa land on the same conclusion: mobile money is the winning deposit rail. It’s instant, it scales, it’s the future.
After running cross-border payment operations for iGaming operators across West Africa — through cash agent networks, mobile money rails, and card flows — I think that conclusion is wrong. Or at least more contested than most ops teams realize.
Below are three operational realities I keep running into. Each cuts against the conventional narrative in some way. Each one shapes how I’d design a deposit channel mix today.
Reality 1: Agent network leakage exists, but it’s not the metric that matters
Conventional wisdom on cash-agent deposits for iGaming: leakage is the killer. Float skimming, ghost transactions, agent margin manipulation, partial deposits where the agent pockets the rest. These all happen. Anyone who tells you otherwise hasn’t actually run a cash agent network.
But the metric that gets framed as “agent leakage rate” is usually less than 1% of deposit volume in a well-disciplined network. It’s bounded. It’s predictable. It’s manageable with daily reconciliation, exception thresholds, and clear settlement-day discipline.
What actually costs operators money on cash deposits isn’t the leakage. It’s the settlement timing asymmetry between the moment the customer hands over cash and the moment the deposit shows up in the player wallet. In well-run operations, this is sub-15 minutes. In poorly run operations, it stretches to hours — and during that gap, the customer abandons, plays elsewhere, or files a complaint that costs more in support time than the deposit was worth.
In one month I looked at across a sample of agent-funded deposits, the operational cost split was roughly 70% settlement-gap support load, 20% recon exceptions and chargebacks, 10% actual agent leakage. The headline metric was the smallest slice.
The operational fix is structural:
- Daily agent settlement reconciliation, not weekly.
- Exception thresholds set tight enough to catch leakage early but loose enough not to drown the recon team in noise.
- Settlement-day discipline — the agent doesn’t receive the next day’s float until yesterday’s reconciliation closes.
Do that, and agent leakage becomes a rounding error rather than a strategic problem.
Reality 2: Mobile money’s hidden cost is conversion drop in the deposit funnel
This is the section that gets the most pushback from people who haven’t operated in this market: mobile money is not the obviously-better deposit rail in West Africa. Not on conversion. Not on customer experience. Not on retention.
The reason is the deposit funnel itself.
A mobile money deposit from inside an iGaming app typically requires the customer to: enter the deposit amount → confirm operator details → trigger a USSD prompt → switch out of the app to the mobile money interface → enter their PIN → confirm the transaction → wait for confirmation → return to the iGaming app.
That’s six to eight interaction steps, with two app switches and a PIN entry, before the deposit lands in the player wallet. Each step has drop-off.
In our experience across multiple operator integrations, mobile money deposit funnels see 15–30% drop-off between “deposit intent” and “completed deposit.” That’s the silent attrition that doesn’t show up on the operator dashboard, because the operator only sees the deposits that succeeded.
Cash agent deposits, by contrast, are typically a single-touch interaction. The customer walks up to the agent, hands over cash, the agent enters the player ID and amount, the deposit lands. No multi-step app interaction, no PIN forgetting, no USSD timeout. Drop-off on the cash channel is usually under 5%.
This isn’t an argument that cash is universally better — agents have their own costs, geographic constraints, and operational overhead. The argument is that the deposit channel mix should be designed around conversion economics, not cost-per-transaction economics. If a mobile money deposit costs 1.5% of transaction value but loses 20% of deposit intent, the effective cost-per-funded-account is wildly different from the cost-per-completed-deposit.
Most operators I see don’t measure this. They optimize the cost of completed deposits without ever quantifying the cost of the funnel they’re losing.
Reality 3: Fraud risk is higher on mobile money than on cash, not lower
The standard pitch for mobile money over cash is safety. Less physical handling, traceable transactions, lower fraud exposure. In high-digital-fluency markets, that argument holds.
In West African markets where many players are still building digital fluency, the operational data tells a different story.
The fraud surface for mobile money in iGaming deposits looks like this:
- USSD social engineering during deposit: “Wrong number” callers reach out mid-deposit pretending to be operator support, walk the customer through a fake confirmation flow, and divert the deposit to a different recipient.
- Phishing during the deposit moment: Fake operator websites or WhatsApp accounts intercept deposit attempts and reroute the customer to a controlled mobile money number.
- Reversal scams: Customer sends a deposit, then claims it was sent by mistake and requests reversal — sometimes with social engineering of the operator’s support team.
- PIN-sharing fraud: Customer’s PIN gets harvested through a related scam, the attacker drains the wallet, and the customer files a deposit dispute with the iGaming operator.
Cash agent deposits don’t have this surface. There’s no remote attack moment during the deposit itself. The agent verifies the customer in person, takes the cash, and triggers the deposit. The customer leaves with a printed receipt or a confirmation SMS. Fraud on the cash channel exists — agent fraud, fake receipts, the leakage discussed in Reality 1 — but the per-transaction fraud rate from external attackers is materially lower than on mobile money in low-digital-fluency segments.
The operational cost of mobile money fraud isn’t just the transaction chargeback. It’s the support load (every disputed deposit takes 30–60 minutes of investigation), the account closures that kill lifetime value, and the reputation damage that travels through tight customer communities.
For ops teams running iGaming payments in West Africa, the implication is direct: a well-run agent network is not just a deposit channel — it’s a fraud control. The customer who deposits via agent is meaningfully less likely to file a fraud-related dispute than the customer who deposits via mobile money, at least in customer segments where digital fluency is still developing.
What this means for ops design
None of this is an argument that cash beats mobile money universally. Mobile money’s geographic reach is unmatched in markets where agent networks are thin. Withdrawals work better on mobile money than on cash for obvious reasons. Operationally lean operators with no agent infrastructure have no choice but to lean on mobile rails.
The argument is that the deposit channel mix is an operational design decision, not a defaulted assumption — and most operators I see treat it as the latter.
Three principles I’d build any iGaming payments operation around in this market:
- Measure deposit funnel drop-off by channel, not just deposit volume. The deposits you don’t see are more strategically important than the deposits you do.
- Build agent reconciliation discipline so leakage is bounded, not hidden. A 0.5% leakage rate that’s measured, exception-tracked, and closed daily is cheaper than a 0% leakage assumption that papers over the real cost of fraud and abandonment elsewhere.
- Track fraud cost end-to-end — including customer-support load and lifetime-value impact. Per-transaction fraud rates lie. Per-customer-relationship fraud cost tells the real story.
If you’re running payment ops at an iGaming operator in Africa and your channel mix is built on assumption rather than measurement, that’s where I’d start.
If you’re working on this problem — channel mix, reconciliation design, agent network discipline — I write about cross-border payment operations across 15+ African corridors. The reconciliation breaks at scale piece and the LemFi teardown get into more of the operational mechanics.
