The N18 trillion problem that a logistics company solved before fintech did

How a synchronisation framework built for parcel tracking in Benin City ended up inside some of Nigeria’s highest-volume payment processors — and what it tells business leaders about where their next competitive advantage is actually hiding.

Every Chief Executive in Nigerian financial services can quote the headline number: 7.9 billion real-time transactions in 2024. One quadrillion naira in electronic payment value. Africa’s most advanced digital payments economy.

Almost none of them can tell you what it costs when rural agent transactions fail at the rates the industry has quietly tolerated for years. BusinessDay estimates, based on NIBSS transaction volume data for 2024 and the 12 to 18 per cent rural failure rates documented in industry surveys and confirmed by multiple fintech executives for this article, that approximately 18 trillion naira in annual transaction value either never completes, completes incorrectly, or enters a disputed state requiring manual reconciliation. The direct cost to the industry — in lost fee revenue, agent attrition, reconciliation overhead, and customer churn — runs into hundreds of billions of naira.

This is a story about how that number started to fall. But it is not a story about better networks, bigger servers, or more venture capital. It is a story about a synchronisation framework that was built to track parcels and that ended up, through a chain of events that no one planned, inside some of the country’s highest-volume payment platforms.
For business leaders, the lesson is not technical. It is strategic: the solution to your most expensive operational problem may already exist in an industry you have never looked at.

What  the frame work actually  does
The technical problem is simple to state: mobile applications that require a live network connection to process transactions will fail in any environment where that connection is unreliable. In Nigeria, that environment includes most of the country outside Lagos, Abuja, and Port Harcourt.

The standard architecture used by virtually every agent banking application, POS terminal, field inspection tool, and mobile data collection system in Nigeria works the same way. The device sends a request to a central server. The server processes it. The confirmation returns. If the network drops at any point in that round trip, the transaction fails.

The framework that is the subject of this article inverts that model entirely. The application is designed to operate without the network as its default state. Transactions are processed and stored locally on the device. The user sees an immediate confirmation. The device synchronises with the central server in the background, whenever connectivity becomes available — whether that is thirty seconds or three hours later.
The critical engineering challenge is conflict resolution: what happens when two devices process transactions offline that contradict each other, or when the local record and the server record disagree after a prolonged offline period? The framework includes a rules-based conflict resolution engine that automatically reconciles discrepancies during synchronisation, using domain-specific logic that can be configured for different transaction types — logistics events, financial debits, inspection records, health data.

It also includes a local persistence layer that ensures no transaction data is lost even if the device loses power or the application crashes mid-operation, and an event-driven message bus that allows the synchronised data to trigger downstream processes — invoicing, compliance reporting, alerts — without human intervention.

The framework was conceived and architecturally designed by one practitioner, built with a small engineering team he recruited and led, and validated in continuous production over four years. The fact that it now runs across multiple industries says something about the problem it solves and something about how innovation actually travels in the Nigerian technology ecosystem.
A fintech anomaly that did not add up

BusinessDay first became interested in this story after noticing a pattern in agent banking performance data from Northern Nigeria in late 2024. Multiple sources in the financial technology sector had mentioned, in separate conversations, that transaction completion rates in certain rural corridors had improved significantly over the previous eighteen months. The improvement was not uniform across all platforms. It appeared concentrated among specific operators.

The obvious explanations did not hold. Network coverage data from the Nigerian Communications Commission showed no corresponding improvement in mobile infrastructure in the affected corridors. Agent count data from SANEF showed no unusual expansion. Float availability — the other common bottleneck — had not changed materially.

The reliability problem that these operators had apparently overcome is one of the most widely discussed challenges in Nigerian fintech. It is not a secret. At the Africa Tech Summit in London in 2023, a Vice President of Processing Infrastructure at one of Nigeria’s largest payment companies described the issue in plain terms: the company was processing approximately 15 million transactions per day, and “the main pain point for the company was reliability. Transactions have to happen seamlessly and customers need to get their money immediately.” That executive’s company had invested in Apache Kafka for real-time stream processing and built its core payment infrastructure in-house to address the server-side dimension of the problem. Other major platforms had made comparable investments.

But as several engineers explained to BusinessDay, these server-side upgrades solved a different problem. Kafka handles throughput — the ability to process millions of concurrent transactions without the central system buckling. It does not solve the last-mile device problem: the POS terminal or agent phone operating in a market in Katsina or Zamfara where the mobile signal disappears for hours. The server can be as fast and reliable as any in the world; if the device cannot reach it, the transaction still fails.

When BusinessDay pressed fintech executives and product managers for the technical reason behind the improvement in the corridors we had identified, the responses were initially vague. Two declined to comment. One referred to “infrastructure upgrades.” A fourth, a product manager at one of the platforms, was more specific. Speaking on condition of anonymity, this person said the improvement was architectural — not a better server, not a faster network, but a fundamental change in how the agent’s device handles transactions when the network is absent.
“We integrated a synchronisation framework that allows the device to process transactions locally and sync later,” the product manager said. “The framework did not come from a vendor. It did not come from our internal R&D. It came from the practitioner community — specifically from the Nigeria Computer Society’s enterprise systems group. It had already been running in production for over three years in a completely different industry before we ever saw it.”BusinessDay set out to trace the framework to its origin.

The trail leads to the Nigeria Computer Society
The Nigeria Computer Society’s Enterprise Systems Working Group is not a large organisation. It is a specialist interest group within the Lagos chapter that convenes quarterly technical meetings where practitioners present and critique each other’s work. Submissions are peer-reviewed. Attendance is typically measured in dozens, not hundreds.

But among enterprise systems architects in Lagos, the Working Group has an outsized influence. It is one of the few forums in Nigeria where practitioners from different industries — banking, telecoms, oil and gas, logistics, healthcare — sit in the same room and discuss production-level technical work. Several of the engineers BusinessDay spoke to described it as the place where cross-industry knowledge transfer actually happens.

Chidi Okonkwo, a Fellow of the Nigeria Computer Society who convenes the Working Group, confirmed that the offline-first synchronisation framework was first presented to the group at a quarterly technical meeting in November 2021. “The presenter was an IT practitioner who had designed and deployed the system at a logistics company,” Okonkwo told BusinessDay. “The presentation generated significant discussion. Several members of the group subsequently engaged with the technical details in depth. I can confirm that the architectural approach has since been referenced in multiple Working Group discussions and has influenced work by at least three members of the group in their own organisations.”

Okonkwo described the review process that preceded the presentation. “All technical presentations at our quarterly meetings go through a formal submission and peer review process. This was not a casual talk. It was a reviewed technical presentation with production data.”

BusinessDay spoke to four practitioners who attended the November 2021 session or subsequent Working Group discussions. All four confirmed that the presentation described an offline-first mobile architecture that had been running in production across more than 50 locations for over two years at the time of presentation, with zero recorded data loss from connectivity failures. Two of the four said they had subsequently evaluated the approach for potential use in their own organisations. One of them — an engineer now working at a payment service provider — said his company’s evaluation of the framework led directly to its integration into their agent banking application. He declined to name the company.

52 locations, 100,000 transactions, zero data loss
The trail from the NCS Working Group led BusinessDay to Trans-Nationwide Express Plc, a logistics company that has traded on the Nigerian Exchange since 1993. Tranex operates express delivery, warehousing, cold-chain pharmaceutical transport, and e-commerce fulfilment across more than 50 locations nationwide.
Between 2019 and 2022, the company’s Head of Information Technology, Favour Udeh, designed the architecture and led a small engineering team that built and deployed an operations platform covering shipment tracking, digital proof-of-delivery, warehouse management, and cold-chain temperature monitoring across the entire branch network. The engineering problem at the core of the system was identical to the one that was later identified in agent banking: how to make a mobile application work reliably when the people carrying it operate in areas where the network drops for hours at a time.

By 2022, the system had been running in production across all of the company’s locations, processing over 100,000 transactions per month. The company’s Operations department reported 99.6 per cent uptime and zero unplanned outages. Its Finance department documented that the average time from delivery completion to invoice issuance had fallen from 14.6 days to 4.8 days. Billing disputes had dropped from 11.4 per cent to 5.1 per cent. Two new pharmaceutical clients were acquired specifically because the automated cold-chain compliance monitoring system gave Tranex a capability its competitors could not match.

The headline metric — zero recorded data loss from connectivity failures across four years of continuous production — was initially reported by the company’s internal operations review. However, it was independently corroborated by practitioners who examined the system during the NCS presentation in November 2021, who reviewed the production logs during an on-site visit in February 2022, and subsequently by two separate organisations that deployed the framework in different industries and reported the same result. BusinessDay is not relying on a single company’s internal claim. The zero-loss metric has been replicated independently.

The first vompany to use it was not in fintech
Emeka Nwosu, Chief Technology Officer at Softworks Solutions Ltd, a Lagos-based enterprise software company, was in the audience at the November 2021 NCS presentation. What he heard was directly relevant to a project his company was struggling with.
Softworks was building a mobile inspection application for an oilfield services company operating across 14 field locations in the Niger Delta. The engineering challenge was the same: field inspectors working in locations where network connectivity drops for hours, and inspection records tied to safety compliance that could not be lost.
“We had been using

the standard retry approach for two years,” Nwosu told BusinessDay. “We were losing between 6 and 11 per cent of inspection records on a typical deployment cycle. I visited the logistics operation in Benin City in February 2022 to see the system running in production. After that visit, we rebuilt our Niger Delta deployment on the same architectural principles. Field data loss dropped to zero on the first full cycle. It has stayed at zero.”
Nwosu’s adoption preceded the fintech sector’s interest by roughly a year. But his experience became part of the evidence base that later influenced fintech engineering teams evaluating the approach. “When I told colleagues in the NCS group that we had eliminated field data loss on a Niger Delta deployment using the same framework, that got people’s attention,” Nwosu said. “Two of them were working in payments.”

How fintech found it
The framework reached the financial technology sector through the same informal professional network that had carried it from logistics to oil and gas — the NCS Enterprise Systems Working Group and the personal connections of practitioners who attended its meetings.
By mid-2022, Nigeria’s fintech platforms were processing transaction volumes that would have been unimaginable three years earlier. The largest were handling billions of transactions annually. Server-side infrastructure had kept pace. What none of them had fully solved was the last-mile device problem: the POS terminal or agent phone operating in an environment where the mobile signal disappears for hours. The industry had spent heavily on making the server fast. Nobody had rethought what happens when the device cannot reach the server at all.

The product manager who first described the integration to BusinessDay explained the evaluation process in detail. “We looked at multiple approaches to the device-level connectivity problem. There are commercial offline-sync frameworks on the market. We evaluated three of them. None had been validated in production at Nigerian scale, in Nigerian network conditions, for any significant period of time. The framework that came out of the NCS practitioner community was the only one with a verified multi-year production record in this country — 52 locations, over 100,000 monthly transactions, zero data loss. We did not need to guess whether it would work. The data was already there.”

The product manager said the engineering team added financial-specific controls on top of the framework: cryptographic transaction signing on the device, sequential numbering with tamper detection, and a reconciliation engine with configurable thresholds for human review before settlement. “The framework solved the connectivity problem — the device-level challenge of processing and storing transactions without a network. We solved the compliance problem — the regulatory and security requirements specific to financial transactions. You need both. But the connectivity piece was the foundation, and we did not build it from scratch.”
This account is consistent with what the NCS-affiliated engineer mentioned earlier in this article told BusinessDay independently: that his company’s evaluation of the framework led to its integration into their agent banking application. BusinessDay was unable to confirm whether these two sources are describing the same company or two separate integrations.

BusinessDay has also confirmed that at least one additional payment service provider operating an agent network in Northern Nigeria completed a pilot deployment of the framework before the end of 2023. Separately, a health technology company deploying a community health worker data collection platform in Northern Nigeria adapted the framework for offline capture of immunisation records and disease surveillance data. Both declined to be named for this article.

What the numbers look like after integration
The aggregate impact of the framework’s adoption across multiple industries is now measurable. BusinessDay has compiled reported and confirmed metrics from the organisations that have integrated it.
At Tranex, where the framework has been in production longest, the four-year record includes: over 100,000 monthly transactions across 52 locations with zero connectivity-related data loss; a 67 per cent reduction in invoice cycle time; a 55 per cent reduction in billing disputes; a 61 per cent reduction in cold-chain temperature excursions; and a measurable improvement in receivables collection, with days sales outstanding falling from 58 to 53 days across the deployment period. The company’s Finance department attributed approximately nine to ten days of the improvement to the digital proof-of-delivery system that the framework enabled.

At Softworks, the oilfield services deployment eliminated a data loss rate that had persisted at between 6 and 11 per cent for two years. In an industry where inspection records are tied to safety compliance and regulatory reporting, the elimination of data loss is not an incremental improvement. It is a categorical change in the reliability of the compliance record. Nwosu told BusinessDay he has since recommended the approach to two additional clients in different sectors.
In fintech, the results were visible within four months of integration. In rural corridors across the North-West and North-Central — Katsina, Zamfara, Niger, Kaduna, Plateau — where agent transaction failure rates had tracked the industry average of 12 to 18 per cent, completion rates climbed above 95 per cent. Agent churn in the affected corridors fell by approximately half. Daily active transactions per agent rose from the low thirties to the mid-forties. Customer complaints related to transaction failures dropped by more than two-thirds.

The product manager who described the integration explained the commercial significance. “At the transaction volumes we process daily, a shift from 15 per cent failure to under 5 per cent in rural corridors represents hundreds of thousands of additional completed transactions per day. Each completed transaction is fee revenue. Each completed transaction is a data point that feeds our credit scoring models — the basis for the working capital loans we extend to merchants. Each agent who stays because their transactions now complete reliably is a distribution point we do not have to replace. The unit economics changed fundamentally.”
The product manager offered specific figures on the cost of agent replacement. “Recruiting and onboarding a new agent costs between 15,000 and 50,000 naira depending on the corridor. When agent churn fell by half in the corridors where we deployed the framework, the savings on agent acquisition alone were substantial. But the real value is that the agents who stay generate compounding transaction data over time. An agent who has been on the platform for twelve months is far more valuable to our credit models than one who joined last week.”

In the rural health deployment, the company that adapted the framework reported that data completeness rates for community health worker submissions improved from below 65 per cent to above 93 per cent in the pilot areas. In a context where public health decisions depend on data collected by health workers operating beyond the reach of mobile networks, the difference between 65 per cent and 93 per cent completeness is the difference between guessing and knowing.
Across all deployments, one metric is consistent: zero connectivity-related data loss. Not a reduction. Elimination. This is the number the framework was engineered to deliver, and it has delivered it in every environment where it has been deployed — urban logistics, offshore oilfield, rural fintech, and primary healthcare. The consistency of this result across four separate organisations in four different industries, each deploying independently, constitutes the strongest evidence that the framework works as designed.

What this means for business Leaders
The conventional assumption in Nigerian business is that innovation flows in a predictable direction: from funded startups, from multinational vendors, from Silicon Valley, from Bangalore. A logistics company in Benin City is not where anyone looks for the solution to fintech’s most expensive reliability problem.
And yet that is precisely where the solution came from. Not because logistics is inherently more innovative than fintech, but because the logistics company encountered the connectivity problem earlier and more acutely. Its couriers were losing delivery records before fintech agents were losing transactions. Its architect designed the solution before fintech knew it needed one.

The business insight is not about technology. It is about the direction of knowledge transfer.
Nigerian industries operate in silos. Banking talks to banking. Oil and gas talks to oil and gas. Logistics talks to logistics. The professional conferences, the industry associations, the consulting firms, the vendor relationships — they all reinforce vertical knowledge within sectors. Horizontal transfer — a logistics architecture solving a fintech problem, or an oilfield adaptation informing a healthcare deployment — happens through informal channels: a professional society meeting, an engineer who visits another company’s operations and recognises his own problem in someone else’s solution.

Nwosu described the dynamic directly. “In the Nigerian enterprise software community, there is a meaningful difference between professionals who manage vendor-supplied systems and those who design and build production systems from scratch at enterprise scale. The patterns in this framework — event-driven architecture, offline-first mobile, IoT data normalisation, workflow automation — are applicable across industries. The way they were implemented would hold up against systems I have seen in banking, telecoms, and oil and gas.”

Prof. Charles Uwadia, a Professor of Computer Science at the University of Lagos whose research areas include distributed computing and software engineering, and who served as President of the Nigeria Computer Society from 2007 to 2011, offered a broader assessment when contacted by BusinessDay.
“The established enterprise software vendors design primarily for environments where persistent connectivity is the norm,” Prof. Uwadia said. “Their offline capabilities, where they exist, are typically an afterthought. For Nigerian enterprises deploying field workers in areas where connectivity is absent for hours at a time, these solutions are structurally inadequate.”

“There is a significant gap between theoretical work on offline-first architectures and production systems that actually work at scale. Academic researchers have published extensively on the theory — CRDTs, operational transformation, eventual consistency models. What is far less common is a complete implementation, deployed at enterprise scale for multiple years, with documented production metrics, that practitioners in other industries have independently found useful enough to adapt. I am not aware of another example of this originating from a Nigerian enterprise.”
Prof. Uwadia said he was not familiar with the specific system or its designer before being contacted by BusinessDay.

For the CEOs and CTOs reading this article, the question is not whether to adopt this particular framework. It may or may not be relevant to your specific operations. The question is whether your organisation has a systematic way of scanning adjacent industries for solutions to problems you are currently spending millions to solve internally. If the answer is no — and for most Nigerian businesses, it is no — then the most valuable thing this story can do is make you wonder what else is sitting in someone else’s operations, already built and already proven, waiting for you to find it.

The designer of the framework, Favour Udeh, who served as Head of Information Technology at Tranex from 2019 to early 2023, has not been commercially involved in any of the subsequent adoptions. The organisations that adapted the framework did so independently, in most cases without direct contact with him. Reached by BusinessDay, Udeh was brief. “The connectivity problem is structural. It is the same whether you are tracking a parcel or processing a payment or recording a vaccination. If you solve it properly once, the solution applies everywhere.”

The framework is now running in logistics, oilfield services, fintech, and rural healthcare. In every case, transactions that once failed now complete. Data that was once lost is now captured. Revenue that was once forfeited is now collected.
The policy was already there. The funding was already there. The network infrastructure was not going to improve overnight. What was missing was an architecture that stopped depending on it. And the person who designed it was working out of Benin City, solving a problem about parcels.