The possible potentials of telecom data in credit analysis

In regions where formal banking systems have failed to reach the most vulnerable, mobile phone usage offers a rich and largely untapped source of insight. This article explores how telecom data can help bridge the visibility gap, offering lenders a new way to assess creditworthiness and build more inclusive systems for the billions who have long been excluded from traditional finance.

I’ve been called many things in this space, from pioneer to advocate, and sometimes even crazy. And I wear all of that proudly. Because when it comes to credit, I’m not in this just to build platforms or piggyback on fancy words like “financial inclusion.” What pulled me into credit wasn’t the idea of building software or being recognised for pioneering open banking in Africa. I’ve never been interested in getting points for being early. What mattered then and still matters now is what credit has done and can keep doing for people who would otherwise never get the benefit of the doubt.

I’ve always viewed credit less as a product and more as infrastructure. Something that, when it works well, quietly supports everything else. It makes jobs feel a bit more secure, gives small businesses a chance to breathe, helps kids stay in school and stops people from having to make desperate choices just to get by. Of course, credit doesn’t fix everything, but it does give people room to move. That’s why, even though I spend a lot of time building APIs and platforms and sorting out the technical stuff, that’s never been the point. The real point is what all that effort enables people to do when they’re finally seen by the system.

It’s a bit like designing a tractor. Nobody builds a tractor just to admire its engine or marvel at its mechanics. You build it because you want to solve the problem of hunger. You want to make farming productive enough that people don’t starve. It’s first-principle thinking. Go deep enough, and what you find at the heart of Lendsqr, at the root of Open Banking Nigeria, and even the way I think about APIs, is a single question: how does someone vulnerable get access to data that can prove they deserve a shot at credit?

Because that’s the real problem, everything else is secondary. Until we can find a way to let the system see and understand people who operate on the margins, all the product features in the world won’t mean much. And I’m not just talking about people who are financially excluded by accident. These are people who have been systematically shut out. In countries like Nigeria, Niger, Bangladesh, and Congo, exclusion is structural. It is designed into the way financial systems were originally built. There are no accessible, reliable ways for the poor to prove that they are safe to lend to, even when they are.

Poverty in these places is not a function of laziness or lack of ambition. It is a consequence of being invisible to the institutions that distribute resources. If you don’t have any way to prove your reliability, you won’t get credit. And if you don’t get credit, you’re stuck. You can’t scale a business, handle emergencies, or even take the smallest risks that might improve your life.

That’s why the real work is not just building lending tools, but rebuilding the logic of how we assess people. We can’t keep applying frameworks built for structured, urban populations to people in informal economies. If we don’t expand what we count as data, we’ll keep excluding the exact people credit was meant to help.

Let’s stop pretending trust is enough

Look, I’m Nigerian, and I know firsthand what it means when people say things like, “We believe in our people.” And to be fair, I get where it comes from. It’s a hopeful sentiment. But belief on its own doesn’t protect capital. When you’re the one lending money (money that could make or break your business) you’re not making that decision based on goodwill. You’re looking for something you can point to, something that gives you a reason to believe that the person will pay back. You’re looking for signals that suggest some kind of financial discipline or consistency. And the further down the economic pyramid you go, the harder it becomes to find those signals.

That’s one of the main reasons I advocate so strongly for open banking. Because when it actually works, when all the right structures are in place, it can surface valuable data that no one had access to before. It can show spending patterns, deposits, transfers, bill payments, and all the small behaviours that help lenders build some level of confidence. You can look at that kind of data and say, maybe not with certainty, but with reasonable conviction, that this person has a stable financial routine. And sometimes, that’s all a lender needs to say yes.

Unfortunately, that entire approach depends on one critical thing: that the person you’re evaluating actually has a bank account. And in reality, that’s a big leap. Most people at the bottom of the pyramid don’t have one. They are not part of the formal banking system in any meaningful way. They might be transacting daily, hustling, and moving money around, but none of that shows up in a structured financial footprint.

And even in cases where they do have a bank account, it often doesn’t say much. Maybe they opened it to receive a one-time government transfer or a salary from a temporary job. Maybe they used it a few times and left it dormant. Either way, there’s usually not enough history to make any real judgment about their financial behaviour. The result is that the system continues to cater to people who are already visible, already banked, and already relatively stable. Everyone else is either invisible or misrepresented. So, we end up with an inclusion strategy that excludes the very people it’s supposed to help.

This forces us to confront a tough but necessary question. If banking data isn’t available or isn’t meaningful for a large part of the population, then what do we use instead? What other types of data can give us the kind of visibility we need to make fair lending decisions?

That’s where telecom data starts to look interesting. Not because it’s perfect, but because, in many cases, it’s the only data that exists.

Telecom data is not new, but we haven’t scratched the surface

The idea of using telecom data for credit analysis isn’t something I stumbled into recently. It’s been on my mind for years, not as a theory but as a response to a gap I keep running into. Back in 2021, I co-authored a paper with some good friends of mine titled Using Open APIs To Drive Financial Inclusion Via Credit Scoring Built on Telecoms Data.” That wasn’t a vanity project or an academic exercise. It came from a real need to find workable alternatives for people who don’t have bank accounts or any traditional financial history. The thinking was simple: if formal financial data doesn’t exist, but almost everyone has a mobile phone, then maybe the phone could tell us something useful.

Back then, the telecom data we had access to was basic. We were looking at call records, airtime recharge behaviour, and SIM registration data. Nothing fancy. But even those simple signals carried meaning. We started asking questions like: Does this person use the same phone number consistently, or are they changing numbers often? Are they topping up their airtime regularly, or is it sporadic? Do they abandon their SIM after defaulting on a loan? Each of these behaviours gave us a small but useful insight into patterns we couldn’t see elsewhere.

Our early work wasn’t technically complex, but that wasn’t the point. We were trying to make sense of what was available and see if we could build basic behavioural profiles out of it. In a context where most people have no credit history, even partial data is valuable. It wasn’t about creating perfect scores; it was about establishing that the data was worth paying attention to. And once you start with something that works, even if it’s simple, you can improve it. That’s how progress happens in spaces where people have been historically excluded. You start with what you have and build from there.

Telecom data today looks very different from what it used to be

Back then, what we had to work with was mostly voice-based information like call durations, contact frequency, airtime recharges, and SIM card activity. That was the standard. It reflected the reality at the time, where mobile usage in many low-income communities was still centered around voice calls and basic SMS. But that version of the world no longer exists. We are now in 2025, and the line between online and offline has blurred, even for the poorest segments of the population. Internet access is no longer a privilege reserved for the “rich”. It has become part of daily life for many people who were previously disconnected.

The people we’re trying to include might not have high-end smartphones or unlimited data plans, but they are online. They are topping up data bundles when they can. They use WhatsApp to stay in touch, they Google things when they’re curious, and they watch YouTube videos, maybe not in high res, and maybe not for long, but they’re there. It might be WhatsApp voice notes instead of phone calls, or YouTube Shorts instead of long videos, but the point is, they are present in the digital world. And that presence leaves a trace.

As a result, the telecom data available today is far more diverse than what we were working with a few years ago. It is no longer limited to who a person calls or how often they top up airtime. It now includes data patterns related to internet usage, geolocation, SIM swapping habits, and device changes. These might seem like insignificant fragments on their own, but together, they can start to paint a picture. A phone that stays in one general area week after week tells a story of stability. Regular data recharges, even in small amounts, suggest some level of income flow. Holding on to the same SIM for a long time can signal accountability, especially in regions where SIMs are easy to discard. These are not perfect indicators, but they are usable. And when traditional credit data is missing, usable data is a good place to start.

To be clear, none of this is about building systems that cross the line into surveillance. It is not about monitoring personal content or tracking browser history. That is not the intention, and it should never be the outcome. What we are talking about is responsibly working with the kinds of data people already produce as part of normal life. These patterns can help open access to credit for those who have been shut out of formal systems, not because they are untrustworthy, but because there has never been a structure in place to recognize the signals they naturally give off. If we can work with those signals thoughtfully and ethically, then we have a chance to finally design credit systems that reflect the world as it is, not just the world as it was imagined decades ago.

There’s a sentience in the data, we just haven’t listened properly

I’m not suggesting that this will be simple. Telecom data, like any data set pulled from real life, will be messy. Some of it will be incomplete. Some of it will be irrelevant. There will be inconsistencies, and there will be noise. But that does not mean it is useless. Buried in that mess is something meaningful. The way people use their phones can tell us more than we think. It’s not about seeing individuals as cold data points. It’s about recognising that these digital patterns reflect real lives. They are expressions of routine, effort, and survival, but we have to be willing to pay close attention to see them.

If we can approach this data with care, if we can build systems that are trained to observe these patterns with respect for context and nuance, then we stand a real chance of unlocking credit for people who have never had access to it before. And when I say credit, I don’t mean a one-time $10 loan through some app. I’m talking about real credit. Credit that can be used to plan, to grow something, to make decisions beyond the next few hours. The kind of credit that opens up options and gives people the freedom to stop living hand to mouth.

That’s what makes this work important. When someone at the very bottom of the ladder gets $100, not as a handout or through pity, but as a loan they are expected to repay with dignity, it can shift the course of their life. That $100 might be used to expand a thriving hustle or send a child back to school.

I’ve seen this play out more times than I can count. I’ve seen people begin with almost nothing and, through access to small or large, well-timed credit, they’ve grown into owners of multiple businesses, employers of labour, and contributors to their communities. I know people who came from poor backgrounds, no connections or safety net, and today they sit at the top of multinational organisations. They lead teams in global banks, tech companies, healthcare, and consulting firms. It didn’t happen because they were lucky. It happened because someone gave them a financial foothold, and they ran with it. That’s not charity. That’s what credit does when it is given with intention and based on a signal that someone bothered to find.

It won’t be easy, but it will be worth it

Telecom data isn’t going to fix everything, and anyone thinking otherwise hasn’t worked in credit long enough. To make it useful, we’ll have to train models on huge amounts of data, from real people, across different geographies and behaviours. That kind of work takes time. It also needs the telecom operators to participate, and that’s a whole challenge on its own. Many of them treat their data like something too valuable to share (understandably so), even when the upside could be huge for the people who rely on their networks every day.

Regulators will also have a role to play. We need rules that protect people from exploitation, but we also need room to experiment. If the laws are too rigid or unclear, nothing will move. And when nothing moves, the people who already have access stay comfortable while the rest remain locked out.

We shouldn’t also pretend it will go smoothly from day one. The first few models will suck, not because people aren’t trying hard enough, but because we’re applying new logic to groups that have never been part of the formal financial system in any real way. There will be false positives and defaults (lots and lots of defaults). 

People who should never have qualified will get loans, and people who would have paid back every rand or shilling will be turned away. But that’s part of building a new system from scratch. Risk isn’t something you can eliminate. You learn to manage it better over time. What’s happening now is that entire populations are being shut out, not because they’re too risky, but because we’ve never taken the time to figure out how to measure them properly.

Telecom data gives us a way to start doing that. It’s not flawless, but it’s real. It reflects how people live. It gives us the chance to build something that finally works for people outside the formal system. And that, at the very least, is worth the effort.

There’s real potential here if we can stop waiting for perfect

The people who need this the most are spread across Africa, Latin America, Southeast Asia, rural parts of South Asia, and large regions of the Middle East. It also includes migrant communities living in countries considered developed, where formal systems still overlook or exclude them. These are people who have never had access to a proper bank account, and who have rarely been considered eligible for anything beyond a small airtime loan, if that.

And yet, despite all of that, they’re online. They use phones, even if on prepaid plans. They find ways to stay connected. They generate digital footprints, sometimes in fragments, but still consistent enough to build from. That means there is something to work with, something we’ve ignored for far too long because it didn’t look like traditional financial data.

We don’t have to wait for the perfect infrastructure before we act. We don’t have to keep telling ourselves that once everyone is banked, then we’ll include them. That thinking has kept too many people on the outside, waiting for a system that was never really designed with them in mind. The truth is, we already have the tools. What we need now is the willingness to use them thoughtfully, carefully, but without delay. Because credit, at its core, is not just about lending money or balancing risk. It is about recognising the potential in people who have been excluded for too long, and building systems that finally take them seriously.

How I’m building a culture of technical excellence at Lendsqr

When I found out Telegram runs on 30 people and OnlyFans on 42, it broke my brain. These are billion-dollar companies shipping fast without the bloat most teams think they need. It made me wonder, why do some companies need 500 people to ship a login page? And so, Lendsqr chose the path less travelled: a lean, technical, and accountable team fully focused on execution.

I stumbled on a post that said Telegram runs with just 30 employees. Thirty. For a platform serving over a billion people. And I just sat there, stunned. I couldn’t even pretend to process it. My mind was properly mindfucked. I kept asking myself, how? How do you run a platform that huge, with that many features, with barely enough people to fill a small Zoom call?

Then came OnlyFans. Yes, that OnlyFans. Whatever people choose to use it for is irrelevant. What matters is that the company managed to generate over a billion dollars in profit within just three years, with a team of only 42 employees. No bloated org chart and endless chains of middle management designed to make lives miserable. Just a few people, some servers, and insane efficiency. 

Clearly, these companies were not playing by the same rules. They had figured out something the rest of us hadn’t.

You see, I’ve worked in places where excellence wasn’t just a fancy word on a mission statement. People like Aigboje Aig-Imoukhuede and Herbert Wigwe weren’t just exceptional because they worked long hours (they did), but because they were razor-sharp, operated on a high-performance frequency, and demanded the same from everyone around them. It was just the minimum standard and you either leveled up or you got left behind. 

That kind of energy had a way of forcing you to either grow fast or quietly excuse yourself.

Luckily, I got a close-up view of that energy early in my career at Access Bank. That place was, without exaggeration, the most valuable kind of pressure cooker. It gave no quarter to mediocrity. You might have walked in with vague ambition, but if you made it through, you came out with a very clear understanding of what elite execution looked like. It wasn’t just about working hard; it was about operating with sharpness and speed, and pressure. You were expected to make decisions quickly, own your output, and never use bureaucracy as an excuse. It could be intense and occasionally unforgiving, but it was the most formative professional experience I’ve ever had. It taught me what a real high-performance culture feels like.

So when I started Lendsqr, I made a personal decision: we were going to build that same culture of excellence,  but make it smarter, leaner, and automated to the bone. Not just hiring people who could take instructions, but people who could figure things out, solve problems without fanfare, and keep things moving without layers of hand-holding. A place where high performance is the norm, not the exception. Where being technical isn’t a job description but a shared mindset. That’s the only way we were ever going to scale without becoming bloated or mediocre.

No dead weight, no dumping ground.

I’ve always held the view that companies should not become homes for “pity departments”. Those departments that exist just so someone can hold a title and collect a paycheck. These are often the places where underperformers get quietly reassigned, tasked with managing something non-essential, while everyone pretends it’s all part of the plan. It’s the professional equivalent of sweeping dust under the rug. Everything looks neat until you bother to look closely.

At Lendsqr, we don’t make room for that kind of organizational clutter. If you’re unable to operate at the standard required for any team in the company, then you don’t make it in. We don’t insulate some teams from pressure while expecting excellence from others. Whether you’re in Marketing, Finance, HR, Engineering, or Product Operations, the expectations are the same. You must be able to think clearly, make good decisions quickly, and execute without someone constantly guiding your hand.

And yes, that meant making uncomfortable choices about how we evaluate applicants. Academic performance doesn’t tell the full story, but it does show a history of effort. 

We obsess over grades. And also don’t obsess over grades. Now, I’m confused 😵

We’ve hired smart individuals from polytechnics who often outperform those with a first-class degree. Some of our most impactful people didn’t come from flashy schools or big brands. One of our best engineers never finished school but what he did finish was a long list of complex features, urgent fixes, and systems that never go down. That’s what matters here. 

In Lendsqr, we care about whether you can think critically, learn quickly, and do difficult work without falling apart. And we’ve built our hiring process to reflect that. It includes rigorous evaluations and hands-on filters. If you’re not resourceful and ready to be useful, you’re not making it past the first few stages, and that’s intentional.

Everyone builds and automates. No exceptions.

One of the biggest problems I’ve seen in companies is how much time gets wasted on repetition and low-value tasks. You’ll find people spending hours each week copying and pasting the same reports. Others wait endlessly for the data team to pull simple numbers they could easily learn to access themselves. Some won’t touch a broken workflow until engineering steps in to fix it because the companies decreed that tech stuff is best done by tech bros, which reinforces silos and slows everyone down. So we killed that way of working at Lendsqr.

From the very beginning, we agreed that if we wanted to operate like a lean, high-performance company, everyone, regardless of their title or department, is expected to be hands-on. That means being able to access the data you need without blockers. If you need numbers, you learn how to write SQL queries. If a workflow is too manual, you build your own automation using n8n. If something breaks, you check the logs, read the error messages, and try to figure out what went wrong before escalating. You don’t sit around waiting for engineering to solve problems that are well within your ability to handle. Whether it’s drafting and testing HTML emails, playing with APIs using Postman, or simply understanding how different parts of the product talk to each other, we expect everyone to be curious and capable enough to dive in. 

And the expectations don’t stop at technical tools. Even engineers are required to understand the fundamentals of credit risk, because it’s not enough to only build features; you need to understand the domain in which they exist. Since you’re expected to build lending platform for lenders, you must know what good lending looks like. 

Just to give you a perspective, our ML models are written by guys from the Product Ops team and the best SQL writers are not even from engineering. Our finance team would coach an engineering team how to use n8n for automation.

There’s no space for helplessness, because we don’t make room for it. If you don’t know how to do something, you’re expected to learn. “I’m not technical” is not a viable excuse in our culture. Even our HR lead builds their automations to make hiring workflows more efficient. That’s how embedded this mindset is. If you work at Lendsqr, you’re expected to think critically, move independently, and make things better without needing permission. You don’t throw problems over the wall and hope someone else catches them.

Product owners have to be actual owners

At Lendsqr, product ownership means just that: ownership. It’s not a role for someone who only writes tickets or sets up meetings. Being a product owner here means taking full responsibility for how your product area performs; technically, commercially, and from the user’s point of view. They’re not simply there to relay messages between engineering and other teams. They have to think clearly about trade-offs, understand how things work under the hood, and stay close to the business and user needs, as this helps them make better decisions and move things forward without needing constant direction.

This kind of ownership requires a wide lens. Our product owners are expected to speak the language of engineers well enough to collaborate deeply and problem-solve together. At the same time, they have to understand the business strategy, how the product supports growth or efficiency, and how to prioritize based on actual outcomes, not gut feelings or noise. And then there’s the user experience; something we expect product owners to care about deeply. Whether it’s a micro-interaction or a major feature flow, they need to be able to evaluate whether it makes sense, feels intuitive, and delivers value.

To support this level of ownership, we’ve built weekly training sessions into our routine. These aren’t theoretical workshops. They’re practical training sessions based on what we encounter day-to-day. We spend time walking through how to write postmortems, not just so that issues are tracked, but so that context is preserved and others can step in quickly. We go through how to escalate issues intelligently, making sure the right people have the right information. Many of these end up with automations that owners do themselves.

We also train product owners on how to debug small issues enough to avoid bottlenecks and dependency loops when engineering is tied up. That kind of autonomy matters when things move fast. On the planning side, we show how to break down features clearly and structurally, so engineering and design teams aren’t starting from confusion or vague ideas. And we talk about usability; what makes an interface intuitive, how users think, and how to bring simplicity into even the most complex flows.

All of this isn’t because we expect perfection or want everyone to be an expert in everything. It’s because we want product owners who are genuinely helpful and capable of driving execution forward. The less they need hand-holding, the more momentum the entire team builds. And over time, that makes the difference between a product team that’s reactive and one that’s constantly pushing things ahead.

Everyone is an engineer, whether it says so on your job title or not

At Lendsqr, technical skills aren’t confined to the engineering team. While we have a team of skilled engineers who handle the deep, complex system work, almost every other team in the company engages with technology in real, hands-on ways. And that’s by design. Being “technical” here isn’t about what degree you have or whether your role has the word “engineer” in it. It’s about what the work demands and whether you’re willing to level up to meet it.

Take our Product Ops team, for instance. Their job isn’t to be passive intermediaries between customers and engineers. They’re expected to understand how our systems work well enough to build automations, write simple scripts, and investigate issues thoroughly before handing anything off. They know how to trace the source of a problem and propose concrete solutions, not just escalate tickets. The Growth team, too, doesn’t sit around waiting for data to arrive. They query APIs themselves, generate the reports they need, and even troubleshoot issues when things don’t look right. In meetings with customers, they can speak fluently about the system, sometimes so well you’d think they were on the engineering team.

We’ve made it clear that there’s no safe zone in the company where you can avoid learning some level of technical skill. Even people who would never describe themselves as “technical”, including those who started out with zero engineering background, are now comfortable reading logs, writing complex queries, or debugging issues.

We’re not trying to turn everyone into a full-stack developer. That’s not the point. What we are doing is removing the kind of fragile dependencies that slow teams down. When people know how to investigate and fix problems themselves, they move faster. When they understand how the systems behave, they make better decisions. And when they’re confident navigating technical tasks, they become much harder to block or derail.

Why do all this? Because we want to scale without bloat

Lendsqr currently serves more than 2.7 million users and supports the lending operations of over 7,000 lenders. And all of this is being run by a team of fewer than 40 people. That’s not a happy accident or a temporary situation we’re trying to fix. It’s the outcome of a deliberate and disciplined approach to building the company. From the very beginning, we made a decision to grow differently by doing more with less, but do it sustainably and without compromising quality or ambition.

Our goal isn’t to keep hiring more people as we grow. In fact, we’re designing the company so that we won’t need to significantly increase our headcount until we’re serving 10 million users or more. And the only way that’s possible is if the company is designed to be resilient and efficient from the inside out. That means building systems that don’t crack under pressure, designing workflows that don’t rely on a single person’s heroics, and creating infrastructure that can handle scale without collapsing under the weight of complexity.

Instead of throwing more people at every new problem, we focus on solving those problems in smarter ways. We build internal tools that take away busywork and allow one person to handle the kind of operational load that would normally require a team. We invest heavily in training and cross-functional learning so that people are empowered to troubleshoot, experiment, and resolve things on their own. We treat documentation as an asset, not a chore, and we create a shared understanding of how things work so that knowledge isn’t hoarded or lost when someone moves on.

That’s what we’re really building: a culture that survives change. One that doesn’t depend on any one person or team to keep the company running. 

It’s not perfect. But it’s deliberate.

We’re not pretending to have it all figured out, and we’re still 10,000 miles from Telegram-level productivity. But that’s the bar we’ve set for ourselves. That’s the north star we’re aiming at because it forces us to think differently about how we build, work, and grow.

This isn’t about chasing perfection but being intentional. I want to build a company where people feel genuinely challenged, where every role comes with stretch, discomfort, and pressure that pushes you to level up. A place where you can’t just coast. Where the expectation isn’t that you tick boxes or do your part, but that you grow sharper, faster, more useful over time.

And it’s not just about engineers chasing technical mastery. We want a culture where every single person (from Marketing to customer support to HR)  is constantly upping their game, learning something new, and improving how they solve problems. We want curiosity to be second nature and mediocrity to feel unnatural.

This kind of environment isn’t built for everyone because it doesn’t offer a quiet corner to hide in. But the people who thrive here? They leave changed. They build resilience, operate like owners, and leave knowing how to figure things out in any room, any company, and under any pressure. And honestly, that’s exactly the kind of company that’s worth building.