The Last Mile Is Not a Delivery Problem

Why Traditional Logistics Models Structurally Fail in Africa — And What Must Replace Them

The standard framing of Africa's logistics challenge goes something like this: the continent has poor roads, insufficient warehousing, and a lack of efficient delivery networks. The solution, in this framing, is to build African versions of companies that solved last-mile delivery elsewhere — a DHL for Africa, an Amazon Logistics for Lagos, a Grab for Nairobi. Pour capital into fleet management, route optimisation, and warehouse construction, and the problem will yield to the same operational playbook that worked in Southeast Asia, India, or Latin America.

This framing is not just incomplete. It is fundamentally wrong. And the companies and investors who continue to operate within it will continue to burn capital on solutions that are structurally mismatched to the environment they are trying to serve.

Africa's last-mile problem is not a delivery problem. It is an economic geography problem. The spatial distribution of demand, the structure of commerce, the patterns of settlement, and the nature of the goods being moved are so different from any other market that has been "solved" by logistics technology that the entire conceptual framework needs to be rebuilt from first principles.

The Geography That Defies the Playbook

Start with the physical reality. Sub-Saharan Africa has approximately 204 kilometres of road per 1,000 square kilometres of land area — one of the lowest road densities in the world. The OECD average is over 1,500 kilometres. Of the roads that exist, only 25 percent are paved. During rainy seasons in East and West Africa, unpaved roads become impassable for conventional delivery vehicles for weeks or months at a time. This is not a temporary condition. It is a permanent feature of the operating environment that any logistics model must accommodate.

But the infrastructure deficit alone does not explain why conventional logistics fails. India has comparable road quality in many regions and has produced functioning last-mile delivery networks. The deeper issue is the spatial distribution of economic activity. In developed markets and most emerging ones, economic demand is concentrated in identifiable clusters — residential neighbourhoods, commercial districts, industrial zones — that can be efficiently served by hub-and-spoke delivery networks with predictable route densities. The economics of last-mile delivery depend on route density: the more deliveries you can make per kilometre of driving, the lower the cost per delivery.

In Africa, this concentration does not exist in the same way. Economic activity is dispersed across a vast geography of smallholder farms, informal markets, peri-urban settlements, and secondary towns connected by poor roads. In Nigeria alone, there are over 72,000 settlements. In Kenya, over 60 percent of the population lives in rural areas dispersed across hundreds of thousands of square kilometres. The demand is there — people need goods, businesses need supplies, agricultural producers need inputs and market access — but it is spread so thinly that conventional hub-and-spoke models cannot achieve the route density required to operate profitably.

This is not a problem that more capital solves. You cannot optimise your way to profitability on a route where there are three deliveries across 40 kilometres of unpaved road. The unit economics are broken by the geography itself.

The Commerce Structure That Breaks Assumptions

The second structural mismatch is the nature of African commerce. Global logistics technology was designed for formal retail — for businesses with fixed addresses, predictable order patterns, standardised inventory, and digital payment capabilities. In Africa, formal retail accounts for less than 10 percent of consumer goods distribution. The remaining 90 percent moves through informal channels — open-air markets, roadside stalls, itinerant traders, and micro-retailers operating from structures that do not have formal addresses, do not appear on maps, and transact primarily in cash.

This informal commercial infrastructure is not a market failure to be corrected. It is an economic system with its own logic, its own efficiencies, and its own distribution networks. The typical FMCG supply chain in West Africa moves goods from manufacturer to regional distributor to sub-distributor to wholesaler to retailer through a chain of intermediaries, each of whom adds a margin but also adds local knowledge, credit relationships, and physical distribution capability that no technology platform has yet replicated at scale.

The logistics challenge, properly understood, is not to replace this system. It is to make it more efficient without destroying the social and economic relationships that make it work. A motorcycle delivery network that bypasses the local wholesaler may reduce the cost of a single transaction. But if it eliminates the wholesaler's credit function — the informal lending that allows micro-retailers to stock their shelves without upfront capital — the net effect on the ecosystem may be negative. This kind of systems-level thinking is almost entirely absent from the logistics investment thesis in Africa.

The Vehicle Problem

Consider something as basic as the delivery vehicle. The global logistics industry runs on vans, trucks, and increasingly, electric vehicles designed for paved roads, fixed addresses, and consolidated deliveries. In African last-mile logistics, the most effective delivery vehicle in many markets is the motorcycle — not because it is optimal in any engineering sense, but because it is the only vehicle that can navigate unpaved roads, navigate through congested markets, and reach locations that are inaccessible to larger vehicles.

Motorcycle logistics in Africa is a massive, largely informal industry. In Nigeria alone, there are an estimated 6 to 8 million commercial motorcycles (okadas). In East Africa, the boda-boda motorcycle taxi industry is one of the largest private-sector employers. These motorcycles already carry goods — agricultural inputs to farms, manufactured goods to rural markets, pharmaceutical supplies to clinics. They do so through informal networks of riders with local knowledge, personal relationships with merchants, and familiarity with road conditions that no routing algorithm can replicate.

The companies that will win in African last-mile logistics are not those building fleets of vans. They are those building platforms that aggregate, formalise, and optimise the motorcycle logistics networks that already exist — turning an informal workforce of millions into a coordinated distribution system without losing the local knowledge and flexibility that make these networks effective. This is a fundamentally different operational model from anything that has been built in global logistics, and it requires fundamentally different technology, different unit economics, and different management approaches.

The Data Desert

Logistics optimisation depends on data — on knowing where demand is, what road conditions look like, where inventory sits, and how long it takes to move goods between points. In developed markets, this data is abundant: GPS-equipped fleets generate real-time location data, point-of-sale systems capture demand signals, road networks are mapped in detail, and weather data provides accurate forecasts of conditions that affect delivery times.

In Africa, most of this data does not exist. Road conditions change seasonally and are not systematically recorded. Demand data from informal markets is not captured by any digital system. Inventory levels at small retailers are tracked mentally, not electronically. Delivery addresses, in many cases, do not exist in any conventional sense — directions are given relative to landmarks, not as coordinates or street numbers.

This data deficit has a compounding effect. Without accurate demand data, you cannot forecast efficiently. Without accurate road data, you cannot route efficiently. Without accurate address data, you cannot deliver efficiently. And without delivery data, you cannot generate the feedback loops that allow logistics operations to improve over time. The data problem is not a secondary challenge to be addressed after the logistics network is built. It is a primary challenge that must be solved simultaneously, because the logistics network cannot function without it.

The companies that recognise this are building logistics and data infrastructure simultaneously — using their delivery operations to generate the demand data, road condition data, and address data that make those operations progressively more efficient. This is a flywheel that takes years to build, but once built, creates a data moat that is nearly impossible for competitors to replicate.

The Agricultural Last Mile

The discussion of last-mile logistics in Africa overwhelmingly focuses on consumer goods delivery in urban areas. This is where the market most closely resembles the use cases that investors understand. But the largest and most consequential last-mile problem on the continent is agricultural — getting farm inputs (seeds, fertiliser, crop protection) to smallholder farmers, and getting agricultural output (harvested crops) from farms to markets.

Africa has over 33 million smallholder farms, accounting for approximately 80 percent of all farms on the continent. These farms produce the majority of the continent's food. But post-harvest losses — crops that rot, spoil, or lose value between farm and market — account for an estimated 30 to 40 percent of production. In dollar terms, this represents $48 billion in annual losses. The primary cause is not poor farming practices. It is poor logistics — the inability to move perishable goods from dispersed farms to buyers quickly enough to prevent spoilage.

Solving agricultural last-mile logistics requires capabilities that are largely absent from the consumer delivery playbook: cold chain management for perishable goods, aggregation infrastructure that collects small quantities from many farms into consolidated loads, quality grading and sorting at collection points, and market linkage platforms that connect fragmented supply to fragmented demand. The economics are different (margins on agricultural goods are thin, so logistics must be extraordinarily efficient), the seasonality is extreme (demand for logistics services spikes during harvest periods and drops to near zero between them), and the geographic distribution is the most challenging of any last-mile use case on the continent.

But the opportunity is also the largest. Reducing post-harvest losses by even a third would be equivalent to increasing agricultural output by 10 to 15 percent without planting a single additional hectare. The economic impact would dwarf any consumer delivery opportunity on the continent. And the companies that solve agricultural last-mile logistics will, by necessity, have built the infrastructure — the collection networks, the route knowledge, the vehicle fleets, the data systems — that can be extended to other last-mile use cases.

What This Means for the Next Decade

The African logistics market is projected to exceed $180 billion by 2030. But the companies that capture this value will not be those that imported a global logistics model and tried to make it work in Africa. They will be those that started with the African reality — the dispersed demand, the informal commerce, the motorcycle-based distribution, the data desert, the agricultural imperative — and built solutions that are native to these conditions.

For founders, the implication is that the competitive advantage in African logistics is not technology per se. It is local operational knowledge — understanding road conditions, market rhythms, trader relationships, and seasonal patterns at a granular level — encoded into technology that makes that knowledge scalable and repeatable. The best logistics companies in Africa will be those that look like technology companies from the outside but operate like deeply embedded local businesses on the inside. This is a much harder thing to build than a routing algorithm, and it is correspondingly much harder to compete with once built.

For investors, the implication is that African logistics requires different evaluation frameworks than global logistics. The relevant metrics are not deliveries per hour or fleet utilisation rates calculated on paved-road assumptions. They are cost per delivery in low-density rural settings, post-harvest loss reduction as a measure of agricultural logistics effectiveness, and informal retailer retention as a measure of commercial last-mile stickiness. Investors who apply global logistics benchmarks to African operations will systematically misprice both the challenges and the opportunities.

For policymakers, the implication is that logistics infrastructure is not just a roads-and-ports problem. It is an economic connectivity problem that requires investment in data infrastructure (accurate mapping, demand tracking, road condition monitoring), regulatory frameworks for motorcycle logistics (which remains unregulated or over-regulated in most jurisdictions), and market infrastructure (aggregation centres, cold chain facilities, quality grading systems) that connects dispersed producers to markets. The $48 billion annual post-harvest loss is not an agricultural problem. It is a logistics policy failure with agricultural consequences.

Africa does not need a better delivery app. It needs a new theory of logistics — one built for dispersed demand, informal commerce, challenging terrain, and the world's largest smallholder agricultural system. The last mile in Africa is not the last mile anywhere else. And the sooner the ecosystem acknowledges that, the sooner it can start building what actually works.