An Honest Look at What we have learned running Referrals.
In 2022, a customer slid into our support inbox with a spreadsheet. He had referred 110 people and done the maths that he was owed roughly ₦899,000 in referral rewards. He wanted answers.
When our customer success team looked into his account, the story got more interesting. Of his 110 referrals, 38 had been flagged. They came in through disposable email addresses. Some shared phone patterns. A few had never traded past the sign-up. The system had done its job by not paying out on suspicious activity. But the customer was still frustrated, and honestly, so was the team looking at the case.
That single support ticket sits at the intersection of everything that makes referral marketing genuinely interesting, and difficult. It is also a snapshot of a wider pattern happening across African tech, one that most companies would rather not talk about publicly.
This article is aimed at helping other fintechs, founders and startups to learn from our experience and draws insights from Feranmi Ajetomobi, a marketing leader who has run referral programmes at multiple African fintechs including Cowrywise, Cenoa and Timon.
Why Quidax and other Fintechs Run Referral Programmes
The first referral programme we did at Quidax was built out in 2020 and expanded into a more robust version in 2021. We wanted a growth loop where our retail customers and influencers could invite new customers to the platform using unique codes, and get rewarded when those referrals signup and start trading.
At the time, this was the standard playbook and it still is. Every serious African fintech ran a referral programme. The logic is compelling because your existing happy customers become your best acquisition channel, and you only pay when they deliver.
Feranmi remembers his first referral campaign. “The first one I launched was for Cowrywise,” he says. “We were doing something where somebody makes a deposit, you get ₦250, and they also get ₦250.” Over time, the Cowrywise bonus scaled up. “At some point when I left, it was even increased to ₦1,000.”
Quidax’s version was designed to be flexible from the start. Our Product team designed this to run parallel in two structures: a fee-based system where referrers earned a percentage of the trading fees generated by the people they invited, and a milestone-based system where rewards triggered when a new customer completed specific actions like email verification, Know Your Customer (KYC), or a first transaction.
These two structures were set in such a way that the admin dashboard was highly configurable. The team could set target currencies, specific regions, transaction thresholds, and the number of transactions required before a reward triggered. On paper, this was thoughtful engineering but in practice, it turned out to be the first mistake. We will tell you what didn’t work as we proceed.
What Actually Worked
Before getting into what did not work, it is worth naming what did.
The Quidax referral programme genuinely drove customer acquisition. The financial incentive worked as designed for the subset of customers who used it in good faith. In fact, it worked so well that to this day, years after the programme was discontinued, our customer success team still receives questions from customers asking when we will come back with another referral programme.
Feranmi’s experience matches ours too. “What worked well on the referral side was it helped us with the ambassador programme at Cowrywise. While at Timon, it has helped us with a lot of organic growth. A good part of our growth is coming from referrals.”
There is also something else that does not always show up in growth metrics but matters just as much. Customers who referred people and earned rewards tended to stay closer to the product. They gave better feedback. They spoke about Quidax more publicly. They became ambassadors in ways that were harder to quantify than just sign-up numbers. Quidax’s Head of Product, Ebube Ojimadu, swears by this particular reason.
The First Crack: Discoverability
The interesting thing about referral programmes is that the first battle is often not fraud or gaming. It is whether customers can even find them.
Feranmi remembers one of the referral programmes he ran initially being buried. “The referral programme was hidden under profile or settings. It was not out there,” he says. “The moment we brought it out into the homepage, it skyrocketed literally within 24 hours.”
That single design decision — where the referral programme lives inside the product — turned out to be one of the biggest levers for whether it worked at all. And most companies get it wrong on the first attempt, tucking the programme somewhere inside the product and not realising nobody sees it.
The Second Crack: Customers Misunderstood the Reward Mechanics
At some point our referral programme paid referrers a 10% discount on trading fees generated by the people they referred. But many customers thought they were earning 10% of the total transaction value. When rewards landed in their wallets, some felt short-changed and following that, support tickets started stacking up. The mismatch between what customers expected and what the programme actually paid became a constant source of friction.
This is a version of a broader pattern Feranmi has seen across companies. Referral mechanics are almost always more complicated than customers assume, and clear communication is sometimes harder than teams expect.
The Third Crack: Over-Engineering
Looking back, Ebube acknowledges that the initial version of the Quidax referral programme was over-engineered. Every possible feature was built in from day one. Every granular setting was configurable. The team wanted a system that could handle any future iteration without requiring redevelopment.
The result was a longer time-to-market than needed, and a system with more moving parts than the launch actually required. Some of the features never got used. Some of the flexibility became a liability when things broke because there were more surfaces where things could break.
The Fourth Crack: People were trying to be Smart with the system
Every referral programme attracts gaming. Our programme attracted a lot of it.
The milestone-based structure, which was the first iteration, particularly had a fault. Customers figured out that they could complete the required milestone actions on new accounts without ever intending to trade meaningfully. What started as isolated attempts by opportunistic individuals eventually became patterns we could categorise, name, and track.
Here are five of the most common ways we have experienced firsthand.
Disposable Email Farming
A customer would generate dozens of disposable email addresses, each email would sign up as a new customer using the same referral code. Some cleared the sign-up milestone, others did not. The customer at the centre would rack up referral counts on their dashboard.
The 110-referrals customer from our opening scene fell partly into this category. Of his 110 referrals, 38 were flagged specifically because they came in through disposable email addresses. The pattern was mechanical, easy to detect at scale, and immediately obvious to anyone reviewing the data.
Multiple Device, Multiple Account Cycle
This is a step up from Disposable Email Farming. Instead of using temporary email addresses, we experienced a couple cases of customers using several devices and different phone numbers to create distinct-looking accounts. Each account would sign up and complete Know Your Customer (KYC). With the signup reward based system, they would already be accumulating rewards just by signing up with different accounts and would cycle money between accounts.
The Influencer Model
Late one evening we got a brand mentions alert about a signup campaign that we had not initiated. When we looked at the thread we noticed that a customer had devised a means to game the system by running their own personal campaign. They would run a promotion asking customers to sign up, complete KYC and execute a trade for a chance to win a reward. The customers who signed up through them simply had to share a screenshot that they had completed the steps and a select few would receive ₦5,000.
For them the maths was simple;
- Get 200 people to sign up with their referral code and complete a trade. At the time customers could deposit and trade with ₦200.
Target: 200 signups - Receive ₦1,000 for each customer that completes the steps.
Potential Total reward – ₦200,000 - Select 5 customers and reward them with ₦5,000 each
Cost: ₦25,000 - Pocket the difference of ₦175,000.
Feranmi mentioned experiencing something similar but this time people went as far as creating YouTube videos of how to make money from referrals, using that avenue to drive traffic and earn rewards for themselves. The referral programme became the product, rather than a vehicle for growing the actual product. Not going to lie, the math made sense.
The Face Identity Trades
This is perhaps the most audacious out of all the attempts and honestly the most uncomfortable to write about.
We experienced a curious case of passing the baton.
There were signups with real identification documents, real phone numbers and approved Face IDs but with one referral code. They would sign up, complete KYC, trade the required amount needed to unlock the reward. Then like a baton, the money would be transferred to another new customer that signed with the same referral code and who would repeat the same steps.
To understand what was happening, we placed a few calls to the phone numbers used on sign up and for the most part they were genuinely unaware of what Quidax was or ever having signed up. We made more calls and discovered that a number of them had been given a small token in exchange for their phone number, NIN and to complete Face ID verification using the referrer’s mobile phone. This revealed to us that some people were willing to trade their private information for small amounts of money.
The Co-Ordinated Signup Ring
Within a 48 hours window we noticed a trend of sign-ups from a specific region in Nigeria which we had previously not seen any significant crypto trades from. These were signups in multiples and raised alarm bells. It felt like a distributed denial of service (DDOS) attack but for referrals.
Later down the line after speaking to several companies that noticed this same trend it appeared there’s a ring that’s on the prowl for new companies running referral programmes. Once they spot and identify an opportunity. They would share it on Whatsapp groups with likeminded people like themselves.
This led us to pivoting toward the fee-based model to ensure the programme generated trading and revenue activity. Here, whoever has been referred must perform a revenue based action to trigger rewards for the referrer. And we also started tracking devices and flagging email patterns.
All these challenges became the reasons why we decided to make some technical changes.
Pivoting from Cash to Product Utility
The pattern of what companies do after they hit the gaming problem is remarkably consistent across the industry.
We had to disable parts of the programme temporarily and move away from the milestone-based structure toward a fee-based model. The fee-based model was harder to game because rewards only came from real trading activity, not from completing sign-up steps.
Feranmi describes exactly the same evolution. His team had to embed the referral in such a way that it was not tied to some actions that can be altered by people. At Timon, where he currently leads growth, the model has evolved further. “On Timon, you have to deposit $20, and then you get a card. Once the person transacts up to a certain amount, the people that referred them gets $10.”
The principle underneath all of this is the same, you have to tie rewards to real product usage, not to sign-up steps. Sign-up steps can be gamed. Real product usage is much harder to fake because it usually involves money that has to genuinely move.
The Instant Gratification Problem
One of the most interesting insights from Feranmi shared was how customers respond to different reward timelines.
Timon initially offered a very generous long-term reward structure. “What we had before was that you earn on what they transact for the next two years, which was good in the real sense,” he says. On paper, this was a smart design. It aligned the referrer’s incentive with the referee’s actual value to the business.
But it did not work because people like instant notifications. Two years is a long time.
This is a lesson that shows up across the industry. Referrers do not want to hear about future income. They want a notification today. The programme design that works well on a spreadsheet often fails on human psychology.
The Stakeholder Reality
One thing that is not often discussed in public conversations about referral programmes is how difficult they are to run internally, even before customers get involved.
Feranmi describes the reality bluntly, “Sometimes people are like, oh this reward makes us look cheap. Oh this reward requires a couple of actions. There are a couple of people that would always want to be involved in this. There’s revenue operations, which look at it from a financial perspective. There’s the branding guy who is thinking “does this thing actually match our energy, our positioning as a brand. What are we supposed to do? How are we supposed to stand out?””
Then there is the alignment with company strategy. “Sometimes it might be what the goal of the company is at that point in time. Are we trying to drive adoption for certain products? ” he said.
Every referral programme sits at the intersection of finance, branding, growth, and product. Getting any one of those groups to disagree can slow or kill the programme. Getting all of them aligned is genuinely hard.
What Is The True Cost Of A Referral Program?
For startups when they first start their referral programme, they consider the reward they need to pay out to the customer but that’s only a part of it. The invisible cost that’s often missed is the cost on your tech stack.
There’s the cloud storage that needs to be scaled up to prevent the app from clashing or to maintain fast load time. The identification service who you need to pay to as the number of verifications increase. Analytics tools and the email marketing or CRM tool that charges as you grow past the tiered price you originally agreed to.
And lest we forget, the time and attention your customer success and other related teams have to put into ensuring that enquiries from both legit customers and bad actors about the referral programme get responded to. Because you don’t want your Time to Resolution (TTR), First Response Time (FRT), First Contact Resolution (FCR) and obviously your Customer Satisfaction score (CSAT) to end up in the mud.
For Startups running referral programmes, what can you learn from our experiences?
Between the Quidax experience and the patterns Feranmi has seen across different companies, some clear lessons emerge for any African startup thinking about launching a referral programme.
Start simple: The most repeated lesson is that the first version of any referral programme should be dramatically simpler than what feels right to build. A minimum viable version that can be tested, measured, and iterated on beats a fully-featured system that ships slowly and creates surfaces for bugs and gaming.
Get discoverability right from day one: A referral programme buried in settings is a referral programme most customers will never use. Put it where they can find it, ideally on the homepage or in a prominent flow.
Bind accounts to devices: One of the most concrete lessons from the Quidax customer success side was that any future referral programme should implement device or phone binding to prevent one person from creating multiple accounts to game the system. People have limited phones. Removing multi-account gaming at the infrastructure level would solve a large share of the fraud.
Communicate reward mechanics obsessively: A significant share of disputes came from customers misunderstanding what they were actually earning. Clearer explanations, built into the product flow rather than the terms and conditions, would have prevented much of the friction.
Design for instant gratification: Referrers do not care about future income. They care about seeing something happen now. Any long-tail reward structure needs to be paired with an immediate visible reward that gives the referrer a reason to keep engaging.
Understand your local market’s fraud patterns: Fraud in African tech does not look the same as fraud in Silicon Valley case studies. Designing for local realities from day one matters more than borrowing playbooks from markets with different behaviours.
Work with legal, support, product, and finance every step of the way: Referral programmes touch too many parts of the business to be run by marketing alone. Legal needs to review the terms. Support needs to know how to handle the disputes. Product needs to build the actual system. Finance needs to figure out the payout system and compare economics. Pulling everyone in early is slower at the start but dramatically faster when problems appear.
Set clear limits on referrals per person and put a budget cap on the programme itself: No single customer should be able to generate an unlimited number of referral rewards. No programme should be able to spend past a defined ceiling without triggering a review. Limits give you the discipline to iterate on what is working without exposing the business to uncontrolled costs when something breaks.
Track the dashboard obsessively: This goes without saying, you must actively track and flag anything that looks out of the ordinary. This increases your chances of observing and recognising unusual patterns.
Moving forward
Referral programmes are one of the most powerful growth tools in an early-stage startup’s kit. They can also be one of the most damaging when they attract the wrong customers, get gamed at scale, or lock the company into paying out on activity that has no real business value.
The businesses that get this right in African markets over the next few years will be the ones that start small, tie rewards to product utility, communicate mechanics clearly, and are honest about when to pause or restructure.
We are sharing this as a part of our goal to equip founders and startups with valuable resources and knowledge to acquire customers and grow their businesses. If you know any founders, marketing, growth or product leader who needs this, share it with them.
