By Ann Ukadike

Over the past decade, African banks have competed almost entirely on acquisition. Branches, agents, USSD codes and slick mobile apps have all served one purpose: to sign up as many customers as possible. That race has largely been won. The harder contest now beginning is not about who has the most customers, but who understands them best. On that measure, most institutions are starting from behind, and the reason is that they treat customer data as exhaust rather than as infrastructure.

Walk through a typical bank and you will find the data scattered. The app team holds one view of the customer, the cards team another, the agent network a third, and the call centre a fourth. None of these systems speak to each other in real time. The practical result is that a bank with millions of customers still talks to all of them in roughly the same way, pushing the same message about the same product to people whose needs could not be more different. Customers tune it out, engagement stays low, and the marketing spend that funded it quietly evaporates.

A customer data platform is the unglamorous fix for this. It is worth being precise about what it does, because the term is often confused with a customer relationship manager or a data warehouse. A warehouse stores data for analysts to query later. A relationship manager tracks interactions a staff member logs. A customer data platform does something different: it resolves the many fragments of a single person – the app login, the card swipe, the agent deposit, the support call, into one persistent profile, and it makes that profile available to act on the moment the customer does something. The distinction matters because action in real time is where the commercial value sits.

When that foundation is in place, the economics shift. A bank can recognise that a customer who just received an inflow is a candidate for a savings product rather than a loan, and reach them in the channel they actually use, at the moment the message is relevant. In my own work building engagement infrastructure, the pattern has been consistent: unify the data first, and previously stubborn numbers begin to move. Engagement rises because the message finally fits, and churn falls because customers stop feeling like strangers to their own bank. These are not marketing abstractions; they are line items a chief financial officer can see.

None of this is automatic, and it would be dishonest to pretend otherwise. The same machinery that lets a bank personalise can tempt it to over-message until customers disengage entirely. Data governance is not a compliance afterthought but the price of entry, and consent has to be treated as something customers grant and can withdraw, not something buried in terms nobody reads. A platform that ingests everything and protects nothing is a liability waiting to surface.

African markets add their own complications, which is why playbooks imported wholesale from New York or London tend to disappoint. Many customers are thin-file, with little formal credit history to draw on. They carry multiple SIM cards and switch devices, which makes the basic task of recognising one person across touchpoints genuinely difficult. A large share of activity still happens offline, through agents, so any honest single view of the customer has to reconcile the physical and the digital rather than assume everyone lives in an app. Solving identity resolution under these conditions is hard engineering, not a configuration setting, and the institutions that treat it lightly will build expensive systems that quietly produce nonsense.

This is also why the current rush toward artificial intelligence in banking needs a word of caution. AI is only ever as good as the data beneath it. A model trained on fragmented, duplicated, poorly governed records will not produce insight; it will produce confident noise at scale. The banks that will benefit most from AI are not the ones that buy it first, but the ones that did the unglamorous work of unifying and cleaning their data beforehand. Sequence matters more than speed.

There is a role for policy here too. Nigeria’s data protection regime and the work of the National Information Technology Development Agency have begun to set expectations around how customer information is handled, and clearer rules, far from being a brake, give institutions the confidence to invest. The other constraint is talent. We do not have enough engineers who understand both the regulatory weight of financial data and the practical craft of making it usable, and closing that gap will take deliberate investment from banks and technology firms alike, not poaching alone.

The institutions that prosper in the next decade will not necessarily be those with the largest customer numbers or the flashiest apps. They will be the ones that decided, early and seriously, to treat customer data as core infrastructure: to unify it, govern it, and act on it with restraint and precision. The asset is already on the balance sheet. The only question is whether African banks will finally choose to use it.