Some products show their strengths or weaknesses most clearly when something goes wrong. As a Senior Product Manager working in Fintech in Africa, I have experienced this reality. This article shares my lessons navigating the challenges of driving repayment from customers via a Pay-AS-You-Go (PAYG) model
In this model, customers make small, daily payments to keep using their mobile device, and the system can temporarily lock the device if payments fall behind. For many readers, this may be a new concept, since the mechanism is common in regions where access to traditional credit is limited.
We discovered that our enforcement system, although working as intended, was creating worry and confusion for customers who relied on their phones for daily communication and income. A model designed to manage repayment was instead affecting trust and predictability. My goal was to figure out how we can drive the best user experience that encourages them to pay off their loan consistently.
This article explains how we reworked enforcement into a clearer and more dependable experience. By rethinking timing, communication, and how payments were confirmed, the team was able to reduce friction while preserving the financial structure of the lending model. The process revealed important lessons about how products behave during moments of stress and how customers interpret fairness.
1. Timing is a UX variable
Our first enforcement model focused on protecting the loan portfolio rather than the customer experience. Devices locked automatically once a user’s credit reached zero, which seemed reasonable because payments were expected daily. The impact proved far harsher than intended. Support tickets spiked, customers who simply forgot a payment felt unfairly punished, and small business owners suddenly found their work disrupted.
A feature meant to manage credit risk was damaging trust instead. Through this failure, we identified three deceptively simple questions that formed the backbone of a redesigned experience: timing, communication, and reconciliation.
How long should the system wait after a missed payment before triggering a lock? A strict lock once a payment is missed is technically easy to justify, but it ignores the volatility of the informal economy. Many of our customers earn in small, irregular bursts. A missed day does not necessarily signal an unwillingness to pay; it often reflects the natural variance of gig work or trade.
We experimented with different delays and watched the data. Shorter delays increased repayment rates but spiked complaints. Longer delays reduced friction but weakened repayment discipline.
What emerged was the realization that the timing of enforcement wasn’t purely a financial parameter; it was a UX variable. It shaped how customers interpreted fairness. The "right" delay had to be discovered through testing, balancing portfolio health with the emotional impact of losing access.
2. Communication as a design surface
Initially, our system sent efficient, automated messages like, "You have missed a payment. Your device is locked." From an engineering perspective, this was accurate. From a user perspective, it was chaos. Customers didn’t know their device was about to lock, whether they could prevent the lock, or exactly how much they owed. Some assumed it was spam; others panicked.
We learned that enforcement is only perceived as fair when it is predictable. We redesigned the messages to specify the exact outstanding amount, the precise time the device would lock, and the specific steps needed to avoid interruption.
We stopped treating communication as an operational necessity and started treating it as a design surface, something to be written, tested, and validated just like an onboarding flow. Once communication became specific and time-bound, customers were more at ease. The process began to feel less like a punishment and more like a transparent system.
3. Payment processing and trust
The third pillar revealed our biggest blind spot: payment processing. Many customers made payments through mobile money agents or third-party channels where transactions took minutes or hours to reflect. If the device locked during that delay, customers interpreted it as dishonesty. The emotional damage of a "false lock" was far greater than the financial value of enforcing quickly.
In cases where a customer makes payment before their phone locks, that’s straightforward, as the device unlock period is extended after payment is received. However, when customer makes payment after device is locked, the wait time between when customer makes payment and when device unlocks becomes very critical to the customer experience
To fix this, we redesigned the backend flow and optimized the various services that processes payment and credit and communication to the locking providers. This reduced payment processing time down to a few seconds and ensures that the device unlocks almost immediately after the customer makes payment.
Speed became our proxy for integrity. When we automated the unlock process to happen within moments of payment confirmation, customers began to relax. They learned that the system was predictable: if they paid, the device worked. We started measuring enforcement reliability with the same seriousness software teams measure uptime. We tracked time-to-unlock and false-lock rates as core product metrics.
Reframing enforcement as a journey
After stabilizing the basics, we stepped back to rethink the conceptual model of enforcement. Our instinctive view was that enforcement was a binary rule: pay or lock. But users experienced it as a narrative. When a payment is made but the phone stays locked, they feel betrayed. When the phone locked earlier than expected, they felt trapped.
We shifted our internal language from "Locking", a term heavy with punishment and exclusion, to "Lock Journey." This wasn't just semantics; it forced us to recognize that a lock event has a beginning, a middle, and an end. Because it is a journey, it can be designed.
This approach highlighted the critical "Pre-lock" phase. Previously, the gap between a customer falling behind and the device locking was a void. Customers sensed they were late but didn't know the urgency. We filled that gap with education and transparency, ensuring customers knew exactly what would happen and when.
We found that the shock of losing access was often worse than the inconvenience itself. Prevention became a tool for maintaining trust.
The "Lock" phase also needed a redesign. The original lock screen was a dead end, a blank slate with no information. It felt like a door slamming shut. We replaced it with a structured interface displaying the reason for the lock, the amount owed, and instructions for payment. We added a real-time indicator showing what would happen once payment was received. The lock screen transformed from a wall into a bridge back to access.
Interestingly, this journey perspective also taught us what not to do. In early pilots, we experimented with random acts of leniency, such as grace periods. While well-intentioned, this confused users. Some learned to wait for the lock before paying; others felt the rules were shifting arbitrarily. We learned that customers do not need a system to be endlessly forgiving; they need it to be clear. Predictability creates more psychological safety than inconsistent flexibility.
Finally, we addressed the end of the journey. Originally, when a customer finished paying off their loan, nothing happened. The device just kept working. Many assumed the loan was still active or were unsure if they owned the device. We added a celebratory completion moment, a digital receipt, a badge of ownership, and a clear message of closure.
This simple UX change increased repeat purchase rates and deepened the relationship.
Moving to adaptive, user-aware systems
Once we established a fair, static baseline, we questioned whether enforcement had to be uniform. A single global rule treated every customer identically, regardless of their history. A user who had paid consistently for eight months but missed one day was treated with the same severity as someone who had never paid on time. This blunt uniformity created friction and ignored context.
We moved toward "Adaptive Enforcement", a system that responds to customers based on their patterns, reliability, and context, much like a thoughtful human agent would.
We began by identifying behavioral indicators that signaled intent. We analyzed repayment histories, support interactions, partial payment patterns, and device usage. We found that some indicators were highly predictive of reliability. For example, a customer who always paid within hours of a reminder was demonstrating discipline, even if they were consistently a day late. Conversely, random payment patterns often signaled income instability.
We rewrote our logic to reflect these nuances. Customers with stable histories received softer communication. Those with irregular histories received firmer reminders and earlier prompts. This wasn't about being overly personalized; it was about being responsive.
This shift allowed us to redefine risk. We previously focused on strict thresholds to protect revenue, but began to see that the riskiest customers weren't always the ones who were late, they were the ones who were disengaged. We shifted our strategy from enforcement to re-engagement, using early signals to flag drifting customers and nudging them long before a lock became necessary.
What enforcement taught us about product culture
Designing enforcement exposed the hidden truths about our product and organization. It acted as a stress test for our entire operation. Weak links in payment reconciliation, API reliability, or support bandwidth that might be tolerable in other features became catastrophic in a credit product. A half-hour delay in unlocking a device wasn't just a bug, but a broken promise.
This reality forced us to break down silos. Payments, fraud, engineering, and customer support had to design, test, and debug together. Enforcement is unforgiving; it demands organizational alignment. It also taught us that customer trust is shaped less by features and more by how a product behaves under stress.
Customers rarely mentioned onboarding when discussing trust. They talked about moments of friction: "They fixed it when I paid," or "They didn't respond when I was locked." These enforcement moments became the emotional reference points for our entire brand.
We learned that fairness is a design outcome, not a policy statement. A rule can be mathematically fair but emotionally unjust. Designing fairness meant optimizing for comprehension, timing, and tone. It also meant using enforcement as a mirror for customer circumstances.
Usage patterns revealed market cycles, harvest periods, and religious holidays. We began seeing missed payments not just as risk events, but as context events, allowing us to design culturally appropriate nudges.
Ultimately, we discovered that discipline and empathy are mutually reinforcing. A strict but predictable system communicates respect for the customer's agency. Clarity allows users to plan; predictability gives them dignity. These attributes mattered far more to our users than generosity.
In conclusion
The most transformative shift in our journey was not technological, but philosophical. We stopped viewing enforcement as a mechanical necessity or a collections function and started viewing it as our most honest interface with the customer.
Trust is not built by avoiding difficult moments, rather by handling them well. Customers judge financial products by their worst days, the moment cash runs out, the moment the device locks. In those moments, the product either behaves like a fair partner or an unpredictable threat. That distinction determines repayment, retention, and reputation.
By reframing enforcement as a user experience, we gave customers a sense of control. When people understand the rules, they behave more responsibly. When they know the system unlocks promptly, they feel respected. When communication is clear, they feel safe.
In the world of PAYG lending, the real product isn't the device or the credit, it is the relationship. Every lock and unlock writes a line in that relationship. When those lines are clear and humane, they accumulate into trust. And in markets where trust is scarce, that belief is the ultimate competitive advantage.
