How African fintech can avoid repeating a measurement mistake

By Abisola Areola

African fintech has time European fintech no longer has: the time to choose how it measures its customers before the dashboards harden.

Most established fintech companies in the United Kingdom and Europe spent the last decade building elaborate customer measurement systems. Product teams track transfer speed, exchange rates, and conversion funnels. Research teams run NPS surveys and concept tests. Customer support teams log ticket volume. The dashboards are sophisticated. The investment has been significant.

The customers, meanwhile, were experiencing something different. And almost none of the systems were built to capture what they were experiencing.

I analysed approximately 100,000 customer reviews and social listening posts about leading cross-border payment platforms, including Revolut, PayPal, Wise, Remitly, and WorldRemit, over a two-year window. What emerged was not three separate findings about customer reviews. It was three faces of one structural problem.

The companies were measuring the product. The customers were experiencing the relationship.

These are not the same thing, and the gap between them is where customer trust quietly erodes.

The product versus the experience

When fintech platforms design market research, they typically ask customers to rate the product. Transfer speed. Exchange rate accuracy. Interface clarity. These are the measurable, designed features.

What customers actually write about is something else.

In the review text I analysed, the themes that dominated were not about product features. They were about experience friction. Words like easy, experience, customer service, and fast appeared at far higher frequencies than exchange rate or transfer limit. Customers were not primarily evaluating whether the product did what it was designed to do. They were evaluating how it felt to use the product when something went wrong: when a transfer was delayed, when a customer service interaction did not resolve the way it should have, when an unexpected fee appeared without explanation.

This matters because companies build what they measure. If your surveys only ask about the features you designed, you will optimise those features while the actual drivers of satisfaction and churn sit in the unstructured text, undetected.

The reviews know what the surveys did not ask about.

When the numbers and the words disagree

The most striking pattern in the data was the mismatch between numerical ratings and the actual content of those reviews. A customer could leave a four-star rating while writing something that read as predominantly negative in tone. The reverse happened too. The numbers and the language were telling different stories about the same customer experience.

This is not a minor data quality problem. It is a structural one.

Most fintech product teams make decisions on aggregate ratings dashboards. They see a clean positive trend while the underlying language is carrying warning signals that nobody is reading. The numbers look good until they do not, and by the time the trend in the numbers shifts, the customer has already left.

Treating review text as a primary data source, not as a qualitative footnote to the ratings dashboard, is among the cheapest customer intelligence investments a fintech can make. Many product teams have never made it.

When the platform shapes the problem

The same fintech company, examined across different review and social channels, surfaced different problem sets entirely.

On Trustpilot, customers tended to write under conditions of high emotional intensity, usually after either an unusually good or an unusually bad experience. The themes concentrated on customer service and resolution. On social media, customers surfaced different concerns, often around trust, brand perception, and recoverability after something went wrong. In-app NPS responses surfaced operational issues that the other channels barely registered.

Businesses relying on a single feedback channel are not getting a representative picture of their customer base. They are getting a picture of the customers who engage with that particular channel, in that particular emotional state, at that particular touchpoint. The rest of the signal is elsewhere, and the rest of the signal is often where the real story is.

Why this matters more for African fintech

The risk of repeating the European measurement mistake is sharper in African fintech, for a reason European fintech never had to manage.

African fintech customers sit across more touchpoints than almost any other market in the world. USSD. Mobile app. Agent network. Web interface. WhatsApp. In some markets, all of them at once. The European companies that built single-channel measurement systems were operating with customer bases that lived mostly inside one or two channels. The cost of measuring incompletely was lower because the customer reality was simpler.

For African fintech, the customer is rarely in one place. A Kenyan M-Pesa user, a Nigerian OPay user, a South African Capitec customer engages across multiple channels, often within the same week, often for different purposes. The single-channel feedback strategy that has cost European fintech years of innovation does not just cost African fintech innovation. It systematically over-represents one customer segment while making the others invisible.

The segment that complains loudest on the channel you watch most is the segment that already has options. The customers you most need to understand are the ones who do not complain on the channels you measure. In Africa, that group is structurally larger than it is anywhere else.

What to do about it

Three changes a fintech product team can make now.

First, audit the gap between what your surveys ask and what your customers write about. If your top ten review themes are not represented in your top ten survey questions, that gap is your innovation backlog. Most teams have never run this audit. The ones that have were surprised by the results.

Second, stop reporting aggregate star ratings to senior leadership without sentiment polarity scores from the text alongside them. If the two diverge, treat that divergence as a leading indicator, not a data quality issue. The numbers will not tell you the customer is about to leave. The language will.

Third, map your feedback channels to your customer segments before you decide which channel matters most. The voice you have built your dashboard around may be representing the smallest share of your customers. The segments that matter most to your growth may be entirely absent from the feedback you currently read.

The choice between measuring the product and understanding the customer is the most consequential one African fintech will make in the next five years. The data already shows what happens to companies that get it wrong. African fintech does not have to find out for itself.

Abisola Areola is a data and digital transformation strategist working across the United Kingdom and Sub-Saharan Africa. She is the Founder of Aethryna Foundation CIC.