No Float Expected

In Parcelles Assainies, a Dakar neighborhood, a Yango driver parks beside a blue kiosk with a yellow Wave logo. He hands over his phone number and confirms the withdrawal on his phone. Ninety seconds later he walks out with 35,000 CFA in cash — two days of earnings. He fuels up, buys dinner, drives home. The transaction that just closed the loop from platform to driver to gas station to grocery never touched a bank.

This is how money moves in emerging markets. Cash, agents, and a tight compression of the time between earning and spending. Miss that compression and the cost lands somewhere specific: driver retention falls by at least thirty percent when payouts move from daily to weekly in the first ninety days of onboarding.

The unit of purchasing power

Credit systems as Silicon Valley understands them do not exist across most of Sub-Saharan Africa. No bureau score. No revolving card. No overdraft waiting to absorb a bad week. More than nine in ten mobile subscribers in the region run prepaid lines.¹ Value here is underwritten by what a household holds right now: cash, land, livestock, inventory, standing in a rotating-savings group. Future income does not enter the calculation.

The purchasing power of a Dakar driver lives in the notes in his pocket and the assets behind him. A Visa limit does not enter the math.

The time-value of a day

A Lyft driver in Denver earns on Tuesday and sees the money Friday. The delay is invisible because the driver has a card, an overdraft, and a checking account that can wear the gap.

A driver in Nairobi cannot wear the gap. By Friday the money is already worth less. Not because of inflation, but because his fuel tank ran dry on Wednesday and he had to borrow from a neighbor. Settlement delay is not a bookkeeping detail. It is a liquidity cost, and the driver prices it into whether he takes the next shift.

I have watched the pattern from Dakar to Karachi to Lima. That thirty-percent delta is the time value of a dollar in a cash economy, compounded across a week.

So the platform that delays settlement pays for it somewhere. Usually in subsidy. The incentive a growth team has to stack on top of Friday's lump sum to match the behavior daily cash would have produced for free. I have watched platforms burn eight-figure budgets on exactly this, year after year.

Prepaid telco plans beat postpaid in every major African market for the same reason. When no one extends credit, no one expects float. Airtime is bought now and used now. The system does not hold balances against a future bill because the billing concept assumes an underwriter no one trusts.

Marketplaces sit one layer up on the same foundation. A driver accepting a trip on a promise of future settlement is accepting float. In a market built on prepaid expectations, float is friction, and friction is priced in whether the platform team models it or not.

Rails arrived sideways

When modern payment rails reached emerging markets, they came through the phone. M-Pesa launched in Kenya in 2007. By the year ending March 2023, Safaricom reported annual M-Pesa transaction value of roughly Ksh 35.9 trillion, more than twice Kenya's nominal GDP.² Its agent network reached around 250,000 outlets over the same period, against fewer than 1,500 commercial bank branches in the entire country.³

In Senegal, Wave launched in 2018 with a flat one percent fee, pricing an order of magnitude below Orange Money's tiered rates. By 2021 Wave had raised a $200M Series A at a $1.7B valuation, the largest Series A in African history at the time, and had overtaken the incumbent without owning a bank.⁴ Its infrastructure is a human one: a mesh of kiosks stitched across Dakar, Thiès, Saint-Louis, and up into the river towns.

India reached the same outcome through UPI. Brazil through Pix. China went from cash to Alipay and WeChat Pay without passing through credit cards at all. The shape repeats wherever formal banking never reached the last mile. The rails that did reach it were shaped by the behavior they inherited.

For a gig marketplace, the agent network is the settlement layer that makes daily payouts economically possible. A platform can credit a driver's digital wallet in real time. If the driver cannot convert that balance into fuel and dinner before the next shift, the credit has not cleared. The kiosk at the corner is what clears it.

The density of the mesh determines how cheaply a platform can operate. In Nairobi, where M-Pesa agents line every block, a driver's time-to-cash is two minutes. In a city with thin agent coverage it is thirty minutes and a motorbike ride. The difference shows up in retention, in subsidy burn, and eventually in the price a rider pays.

The cadence question

Daily settlement is the right default at launch. It builds trust faster than any marketing spend. A driver who sees money land every night believes the platform. A driver waiting until Friday is one neighbor's folded-platform story away from logging off.

Daily payouts also attract fraud. Collusion between drivers and riders. Fake trips. Incentive stacking across accounts. Weekly payouts create a fraud window. Time to detect patterns before the money walks.

Payout cadence has to track where the platform sits in its own arc. Early, pay daily. Trust has to be manufactured. As the risk stack matures, the cadence can stretch. KYC that holds up. Trip-integrity scoring. Behavioral flags. Anti-collusion models. By the time a platform can pay weekly without bleeding retention, the guardrails have been built. Not before.

This is not a universal rule. Across Latin America, Asia, and Africa you find both ends: drivers with credit buffers or side-gig income who can absorb weekly payouts, and drivers who live paycheck to fuel tank and cannot. The right framing is to read the driver base's liquidity constraint and price settlement to match, not to pay daily everywhere by default.

Design for the pattern

The transaction pattern in emerging markets already works. It is tuned to a world where credit is expensive, data is thin, and the cost of waiting is paid in immediate liquidity. A marketplace that fights the pattern subsidizes its own friction. A marketplace that builds with it inherits the resilience of a system already tuned to the ground. Daily cash-out through dense agent networks. Compensation that clears before the next shift. Fraud systems that earn the right to stretch settlement later. The financing loop and the informal stack tell the same story as this one, one layer up and one layer down from settlement.

Cash is the rail. Agents are the mesh. Delay has a price, and in emerging markets it is steep.