Unsupported-payment rate
The share of real payment volume arriving through an asset or network we cannot accept. The number the company is a bet on, and the one we will publish honestly.
About
A Lagos company, at the beginning.
Built by an engineer who gets paid in dollars from Lagos, and who noticed that the receiving business had quietly become the integration layer between its customers and Nigerian settlement infrastructure. Removing that job is the product.
Where we actually are
We are in private beta, and the early work is deliberately conversations rather than code. If you are considering trusting a company with a payment worth more than a car, you should be able to find out exactly where it stands without asking.
Now
Asking
Twenty to thirty interviews with businesses receiving international payments: how do your customers pay, on what networks, how often, at what size, and has a payment ever arrived somewhere you could not use? One more, about the other side: would your client find a proper payment record useful, and has paying you ever been the thing that slowed an invoice down? No code, just the questions the company rests on.
Now
Structure and partners
Nigerian counsel on what this activity legally is, and partners who can hold, convert and pay out under their own permissions — in writing, before production. The regulatory model is a product requirement, not an afterthought.
Planned
On the roadmap
Indexers, screening, matching, routing, ledger, dashboard, API, bank rails, paying out. All of it real, and none of it begun before twenty payments have been run by hand.
Never
Never planned
A bank, a wallet, an exchange, a proprietary bridge, lending, yield, or pooled customer balances. Holding customer money changes what this company legally is, so we minimise custody by design.
The five phases
In order, and we do not skip one because the next is more interesting. The first two need almost no engineering, and they are the ones that decide whether the rest happens.
Now
Validation
Twenty to thirty interviews, then three to five concierge customers and twenty-plus real payments. Document the network distribution, calculate the mismatch rate and the dollar volume it affects, and measure willingness to pay. Conversations, not code.
Next
Concierge product
Onboarding, a real Nomapay account, payment intents, a few payment endpoints, monitoring — then manual screening, manual routing, manual settlement and a real receipt. The customer experiences the finished product even though the backend is partly a person.
Then
Automated MVP
Automate detection, confirmations, screening, matching, provider quoting, routing, conversion, settlement and the receipt. Human operators stay for exceptions, because V1 should not attempt complete automation — the point is to discover the real operational workload before automating it.
After that
Multi-provider
A second and third settlement provider, measured on net price, reliability, settlement time and failure rate. This is the point at which routing is worth anything at all — with one provider there is nothing to route between.
Later
Widen the account
More assets and networks, added only where customer demand, settlement capability, compliance support, reliable infrastructure and positive unit economics are all present. Then bank rails and paying out — the account was designed for both directions from the start, and this is when the second one opens.
What we measure
Registered users is a vanity number. These are not.
Settled volume
International payment volume actually settled through Nomapay. The measure that matters, not registered accounts.
Contribution margin per dollar
What we retain after every underlying execution, network, screening and operational cost. Revenue alone would flatter us.
Settlement time and failure rate
Median, 95th percentile and failure rate from detection to bank credit. Measured, not advertised. Turning execution data into a product advantage only works if the data is real.
Unknown-state rate
How often an execution goes quiet, and how long a person takes to resolve it. The number that decides whether this is operable at all.
Repeat volume
Whether a business sends its next customer to the same account. The most important one, because it is behaviour rather than opinion.
Regulatory position
Stated plainly, including the parts that are still open.
Nomapay orchestrates; regulated infrastructure holds and moves
Digital asset custody sits with a licensed or authorised partner. Conversion and payout sit with a licensed partner under their own permissions. Nomapay supplies the account identity, detection, screening, matching, routing decisions and records. We avoid unnecessary balance-sheet exposure by design.
Registered, not licensed
Nomapay Global Limited is registered with the Corporate Affairs Commission. That is a company registration, not a financial licence, and we will not describe it as one. Ten questions must be answered in writing before production: whether we act as technology provider or agent, whether this constitutes VASP activity, whether receiving and transmitting instructions triggers registration, who legally holds customer funds, who converts, who custodies, and who owns the AML and transaction-monitoring obligations.
“Just software” is not an answer we get to assume
Nigeria’s SEC framework already regulates virtual-asset activities including VASPs and digital-asset custody and transfer; its August 2026 proposed rules explicitly reach digital-asset transfer and settlement, and it is admitting VASPs into the Accelerated Regulatory Incubation Programme. A July 2026 Presidential Executive Order set out to harmonise virtual-asset regulation across agencies, and the CBN continues to emphasise regulated channels for international transfers. We must not assume that being a technology provider automatically removes regulatory obligations.
Screening is not optional
Every sending wallet is screened before anything moves. An adverse result stops the payment and escalates to a person. We would rather lose the transaction than route value we cannot account for.
Data
Transaction records are sensitive, and on-chain records are permanent and public in a way bank records are not. Handling follows the Nigeria Data Protection Act. We never ask for banking credentials and do not need them.
What could kill this
Written down now, while admitting it costs nothing. In rough order of how likely we think each one is.
Most likely
The problem is not frequent enough
If most businesses simply receive USDT on Tron and their off-ramp already handles it, the routing value is low. Fewer than five of thirty interviewees with a meaningful mismatch and we do not build the routing product. This is the biggest risk here, and it is not technical.
Affected volume is immaterial
A 20% mismatch rate on $2,000 payments is irrelevant; a 5% rate on $500,000 payments is an excellent business. Frequency alone would mislead us, so we measure problematic volume as a share of total inbound volume and multiply it by what the problem actually costs.
Existing providers already solve it
If Nigerian settlement partners widen their own asset and network support faster than payer treasuries diversify, the gap closes without us and there is nothing left to sell.
Businesses will not hand over the destination
“Would you use this” is worthless. The behaviour we need is a business telling its own customer “pay my Nomapay account”. If the account abstraction is not the right experience, that never happens.
Thin margin, or commoditisation
If providers charge us close to what customers will pay, there is no business between the two. And if this collapses into pure price competition, the defence has to be payment identity, settlement data and reliability — which have to be real, not asserted.
Regulatory classification, or custody exposure
Nigerian payments and digital-asset rules are moving quickly, and a classification we did not plan for could make part of this unworkable or prohibitively expensive. The mitigation is counsel, licensed partners, and minimising custody.
Every figure on this site is illustrative unless its source is named. If we publish a number that turns out to be wrong, we correct it in public rather than quietly.