In Lagos, saying you “do POS” means something specific: you sit behind a small terminal and hand out cash. In Nairobi, “POS” more often means the thing that rings up a sale. Two different businesses, one borrowed acronym, and a great deal of confusion for anyone trying to decide what to put on their counter.
The distinction matters more in 2026 than it did two years ago, because the rules and the numbers have both moved. Nigeria’s PoS transaction value passed 88 trillion naira in the first eight months of 2025 according to NIBSS data, on a network of more than 2.1 million active terminals. Meanwhile Kenya’s registered agent network is shrinking while its merchant network grows. Those two facts point in opposite directions, and they should change how a shop owner allocates space and capital.
The two businesses, side by side
Agent banking is a cash-handling service. You act for a bank or fintech, customers withdraw and deposit at your counter, and you earn a commission per transaction. Merchant payments are the opposite flow: money comes in because you sold something.
| Agent banking | Merchant payments | |
|---|---|---|
| Revenue model | Commission per transaction | Margin on goods sold |
| Main constraint | Float: cash and e-value held simultaneously | Inventory and supplier terms |
| Regulator’s view | Financial service, licensed principal required | Ordinary trade |
| Nigeria scale | 2.1m active terminals, over 88 trillion naira in 8 months of 2025 (NIBSS) | Not separately reported |
| Kenya scale | 333,011 M-Pesa agents, FY to March 2026 (Safaricom) | ~2.4m M-Pesa merchants, of which ~0.9m Lipa Na M-Pesa tills |
| Direction of travel | Agent count falling 5.6% in one quarter (Q2 2026) | Lipa Na M-Pesa handled KSh 1 trillion in six months to Sept 2025 |
The row that deserves attention is the last one. Agency is consolidating. Merchant acceptance is still expanding.
What the CBN changed, and when it bites
Nigeria’s Central Bank issued new Guidelines on the Operations of Agent Banking on 6 October 2025, replacing every previous agent banking framework with a single document. Most provisions were binding immediately. One was not.
The single-principal rule takes effect on 1 April 2026. From that date an agent may work with only one principal, and the same exclusivity applies to super agent networks: an agent belongs to one super agent at a time, though super agents themselves may still hold multiple bank relationships (Techpoint Africa).
This dismantles a model that a large share of agents built their income on. Running terminals from three or four providers at once let an agent route around a provider outage, chase the better commission of the day, and keep serving customers when one platform hit its limits. After April, that flexibility is gone. An agent whose principal has a bad week has a bad week.
The guidelines also cap customer withdrawals at 1.2 million naira per day and require principals to publish verified agent lists. The cap is a fraud control, not a revenue control, but it changes who an agent can serve: the high-value cash customer now has to go elsewhere, or split across days.
For a shop owner weighing the two businesses, the practical reading is this. Agency banking in Nigeria is becoming a more regulated, lower-optionality business with a single counterparty. That is not a reason to leave it, but it is a reason to stop treating it as the easy second income it was in 2023.
Kenya: agents are consolidating, tills are multiplying
Kenya reached the same crossroads by a different route. Registered mobile money agents fell from 602,470 to 568,463 in the quarter to June 2026, a 5.6% drop in three months (TechTrends), while mobile money accounts climbed past 54 million with M-Pesa holding roughly 89% of subscriptions.
Safaricom’s own count went the other way, rising to 333,011 M-Pesa agents in the financial year ended March 2026 from 298,890 a year before. The two figures are not contradictory: the market is concentrating on the dominant network and thinning out elsewhere.
Merchant acceptance shows no such thinning. M-Pesa reports around 2.4 million merchants, made up of roughly 900,000 Lipa Na M-Pesa business tills and 1.5 million micro-merchants on Pochi La Biashara. In the six months to September 2025, KSh 1 trillion moved through Lipa Na M-Pesa. A till is cheaper to run than an agency: no float, no rebalancing, no cash risk at the counter.
We covered the mechanics of till numbers, paybills and reconciliation in the M-Pesa merchant guide.
The float problem nobody mentions upfront
Terminals are easy to get. Principals hand them out, sometimes free, because the terminal is a distribution cost and the agent absorbs everything else.
What the agent absorbs is float. To serve withdrawals you need physical cash. To serve deposits you need e-value. You need both at the same time, all day, and the ratio between them shifts with your street’s traffic: a market location runs out of cash by noon, a residential one runs out of e-value. Rebalancing means a trip to a bank or a super agent, during trading hours, with cash on you.
That is the real barrier to entry, and it is why agency income should be measured as a return on working capital rather than as commission per transaction. An agent turning 300,000 naira of float three times a day earns on volume, not on margin, and every hour the float sits idle is a cost.
Merchant payments have no equivalent. The capital sits in your inventory, which you were going to buy anyway.
Running both without blending them
Plenty of shops do both, and that is a reasonable position. The failure mode is not doing both, it is accounting for both as one.
Three rules keep them separable.
Keep the tills physically apart. Agency float in one drawer, shop takings in another. The moment they mix, an agent covers a shortfall with shop cash and the shop’s margin quietly funds the agency.
Reconcile against different sources. Agency reconciles against the principal’s statement. Retail reconciles against your own sales record, item by item. If your only record of retail is the same phone that shows agency transactions, you cannot tell which business paid for this month’s restock.
Measure them on different metrics. Agency: commission per day divided by float deployed. Retail: gross margin per product and stock turn. A busy agency counter can hide a retail business that has been losing money for a year.
What the retail side needs from software
If you sell goods, the sales record is the part that has to be automated, because it is the part you cannot reconstruct from memory or from a bank statement.
Four requirements, in order of how often they get skipped.
Inventory decrements per sale. Every sale should remove the item from stock without a second entry. Anything else drifts within weeks.
It works with no network. Fibre cuts and grid outages take terminals down regularly, and a shop that stops selling when the connection drops loses revenue the agency business never recovers. We set out how true offline mode is built, and how to test a vendor’s claim, in our piece on offline POS software.
Mobile money is a first-class payment method, not a note typed into a “cash” sale. Otherwise your till totals and your mobile money statement will never agree.
The price is the price. Inventory, multi-user access and reporting are commonly sold as upgrades on software advertised as free. The POS comparison puts the published prices and features of the main options side by side, each with its source.
digabloPos is our answer to that list: free, inventory included, full offline operation on a phone you already own, with M-Pesa, Orange Money and Airtel Money built in. It does not make you an agent, and it is not meant to. It records the other business, the one that pays your rent.
The short version
If you have float, a location with foot traffic and the discipline to rebalance daily, agency banking still works, with the caveat that Nigerian agents need to pick their principal carefully before April 2026. If you have stock on shelves, merchant payments are where the growth is, and they cost you nothing to hold.
Most shops should stop treating the first as the serious business and the second as an afterthought. The numbers stopped supporting that a while ago.