Over the past decade, few ideas have carried as much hope for Africa’s future as the power of data. Politicians, international organisations, and technology companies alike have repeated the same promise: once the continent is digitised, progress will follow. If farmers can record their harvests through apps, if governments can track citizens’ health records, and if businesses can follow customer purchases, then suddenly everything will become visible. And when things are visible, the reasoning goes, they can be measured, predicted, and improved.
Digitisation has already transformed how people communicate, move money, and access services. However, the next challenge is even more complex: ensuring that the data we collect actually helps us build stronger systems.
This promise is not abstract. In many parts of the world, it has already proven true. Think of a supermarket in Europe or the United States. Managers there can tell you exactly how much bread to bake or milk to stock on a Monday morning. They rely on years of sales records, carefully tracked by scanners at checkout counters. They know that customers tend to buy more soup in the winter, that sales of barbecue equipment spike in the summer, and that demand for certain goods rises before the holidays. The data is stable and predictable, so the supermarket rarely finds itself with empty shelves or wasted stock.
The same is true in healthcare. In countries with reliable health systems, patient data helps doctors and governments predict trends. If flu cases begin to rise in one city, officials can prepare hospitals and issue health warnings. Because the patterns are steady, past behaviour is a good guide to the future. This is what data promises when everything is measured and stable: the ability to plan with confidence.
But when you shift the lens to many African economies, the story begins to unravel. Here, the numbers don’t behave as neatly. Household incomes often rise and fall unpredictably. A family might enjoy steady earnings one month, only to lose them the next because of seasonal work, illness, or sudden expenses. A farmer may bring in a large harvest one year but watch the next crop fail due to drought. Traders and small business owners often operate in both formal and informal markets, switching between them depending on what opportunities arise. All this makes spending and production patterns volatile. What looks like a clear trend today can vanish tomorrow.
This volatility means that data collected in such environments doesn’t settle into the steady patterns needed for forecasting. Instead, the numbers swing back and forth, forcing those who rely on them to adjust and adapt constantly. For example, a mobile money provider might see a surge in subscriptions for one month but cannot assume the growth will continue, because customers may cancel when cash flow tightens. A shop owner may notice high sales of imported biscuits in January, only to see demand collapse in February when customers shift back to cheaper local alternatives.
The challenges of data in Africa are not new. They go back to the colonial era, when governments relied heavily on measurement as a tool of control. Colonial administrations carried out censuses, agricultural surveys, and land records, but they did so with narrow goals in mind: to tax, to regulate, and to extract. They focused on what was easiest to count and what served their interests. Cash crops like coffee, tea, and cocoa were carefully recorded because they entered international trade and could generate revenue. By contrast, the maize, beans, and vegetables that fed most households were often ignored.
This pattern of selective counting did not disappear with independence. Newly formed states inherited the same statistical systems, and many continued to focus on export crops and formal jobs because these were easier to measure and report. Even today, when governments release employment data, the numbers are often skewed toward those with formal payslips, leaving out the millions who work informally as casual labourers, street vendors, or family farmers.
Digitisation was supposed to fix these blind spots. With the spread of mobile phones, mobile money, and digital platforms, vast new streams of information became available. Every mobile payment, every e-commerce purchase, and every clinic registration created a new record. For the first time, it seemed possible to capture the “hidden economy”: all those small, everyday transactions that had previously gone unmeasured.
This optimism has not been misplaced. Digitisation has brought visibility to areas that were previously invisible. But visibility is only step one. The question now is how to translate it into a more profound understanding and stronger systems.
Policymakers imagined dashboards showing real-time movements of goods, services, and money. Businesses saw opportunities to tailor products to customers with unprecedented accuracy. Donors and development agencies celebrated the idea of measurable progress, where every dollar spent could be tied to digital evidence of impact.
Yet the reality has proven more complicated. Digital data, like older forms of measurement, still reflects only part of life. A farmer who sells maize through mobile money will appear in the records, but if she trades some of her harvest with a neighbour for school uniforms, that transaction vanishes. A family might register at a clinic once, but if they later turn to traditional healers, their choices remain invisible. Even when data is abundant, it mirrors the fragmented systems through which it flows. Instead of revealing the whole truth, digitisation often highlights the gaps.
It helps to think of these challenges in three groups.
First, there are unmeasured but measurable things. This includes areas such as crop yields, livestock populations, irrigation flows, and electricity usage. These are straightforward to track, and once they are measured, the information can genuinely improve planning because the underlying patterns are relatively stable. For example, knowing how much water flows through irrigation canals can help farmers plan planting schedules more effectively.
Second, there are measured but unstable. Consumer spending is the clearest example. A record of what families bought last month may not predict what they will buy next month, because their income and priorities keep shifting. A supermarket can stock its shelves based on past sales, but it risks miscalculating when those patterns suddenly change. Subscription services, whether for mobile airtime or pay TV, face the same challenge. Customers sign up when money is available and drop out when it runs short. The data exists, but it does not provide certainty. It only captures short-lived fluctuations.
Third, there are measured but biased. This is the most deceptive because it creates the illusion of completeness. Governments and large institutions often measure what passes through official channels, like taxed businesses, exported crops, or formal jobs. These measurements are valuable. They give us a baseline we never had before. But if they are treated as the whole picture, they risk pulling policy away from the areas that carry most of the productivity.
What all this shows is that more data does not automatically lead to better understanding. Scanning produce at collection points, digitising truck routes, or recording farmer transactions may create new streams of information. Still, they do not fix the underlying problems of fragmented systems, unstable markets, or biased measurements. In fact, they can sometimes make matters worse by giving false confidence. Decision-makers may believe they are seeing the whole picture when they are looking at a partial and distorted version.
Digitisation has already proven itself as a powerful tool. The next step is to treat it not as the endpoint, but as part of a larger process. On its own, it cannot deliver stability or insight. What matters is how the data is interpreted and what systems are built around it. The real gains come when visibility is paired with understanding: when numbers are combined with local knowledge, when systems are designed to adapt to change, and when data is used not for prediction but for navigation.
In many African economies, it is more helpful to think of data as a compass rather than a crystal ball. It cannot tell us precisely what will happen next, but it can point us in the right direction, helping us test ideas, adjust quickly, and stay flexible in the face of uncertainty.
Co-authored with Ambassador Professor Bitange Ndemo as part of a series on Value Chain Productivity and Innovation
