TL;DR — Across the six markets we work in outside South Africa, the regulators' own reporting puts water losses between a third and a half of everything produced: Kenya 48% and rising, Ghana around 54%, Botswana around 45%, Tanzania averaging 33% with regions worse, Zambia rising despite near-universal metering (the meters themselves are the suspect), and Nigeria fighting the battle at the billing-accuracy front. Different causes, one shared gap: almost none of these networks can say which zone loses the water. That answer costs a handful of telemetered zone meters per scheme — and it's where every successful reduction programme on the continent starts.
The scoreboard, from public reporting
| Country | Headline figure | Source context | The local twist |
|---|
| Kenya | 48% national, up from 44% | WASREB 18th Impact Report (FY2024/25), 94 utilities — ≈242 million m³/year lost | Regulator publishes every utility's number; Nairobi's utility is answering with a KSh 2.57bn ultrasonic metering programme |
| Ghana | ~54% of production | Utility and industry reporting on the national water company's 88 urban systems | A telemetry control room and metering lab now operate in Accra — the direction is set, the zone layer is the gap |
| Tanzania | ~33% average | EWURA performance reviews; regional utilities reach 40%+; Dar's utility lost tens of billions of shillings in past cycles | The prepaid-water wave is fixing collections — making the residual gap increasingly a physical loss story |
| Zambia | Rising, despite ~100% metering | National NRW Strategy 2022–2026; audits found large shares of meters faulty | The instrumentation itself lies — verification before replacement is the rational sequence |
| Botswana | ~45% of production | Sector reporting on the national utility | Every lost megalitre travelled the North-South Carrier or came off a wellfield — the most expensive water on this list |
| Nigeria | State-by-state; billing accuracy is the battleground | Lagos alone has deployed tens of thousands of meters at ₦200–350/m³; concession debates centre on provable billing | Borehole-first supply means the loss story starts at private sources, not treatment works |
Global good practice sits near 20%. Figures are the latest public numbers as at August 2026; each country page linked below carries the fuller context.
Three patterns hiding inside the percentages
1. The number everyone quotes is the number nobody can act on. A national percentage has no address. Crews can't be dispatched to "48%". The unit of action is the zone — a district metered area with a telemetered inlet — and the difference between utilities that reduce losses and utilities that report them is almost always that the first group knows its worst zone by name.
2. Commercial and physical losses need different armies. Kenya's rise, Ghana's half-lost production and Botswana's carrier losses blend leaks (send repair crews) with metering and theft (send inspectors and the metering lab). Zone data splits the two cleanly: high night flow means pipes; a wide billing gap with normal night flow means meters and connections. Zambia is the cautionary tale — near-universal metering that under-registers is worse than honest under-metering, because it hides the problem inside the instrument.
3. Prepaid and smart metering don't answer this question. Consumer metering — Tanzania's LUKU-style wave, Botswana's 35,000-meter rollout, Lagos's programme — fixes collections and reading. None of it can see a trunk leak or a bypass. The utilities getting ahead pair consumer metering with independent zone metering, so vends-versus-inlet becomes a daily loss statement.
What the successful programmes share
Having built this data layer on networks from municipal systems to estates, the working pattern is consistent and unglamorous: meter the zone inlets (a handful of telemetered bulk meters explains more than thousands of consumer meters), watch minimum night flow (the 2–4am inflow is the strongest leak signal a utility owns, and it's invisible to manual reading), split the loss buckets per zone, fix the worst zone first, and re-measure to prove the recovery — because the documented recovery is what funds the next zone. The instrumentation survives outages on solar and buffering, which matters precisely because the grid events that stress these networks are also when records go missing.
Country-specific versions of this programme, with each market's regulators, utilities and numbers: Kenya · Ghana · Nigeria · Tanzania · Zambia · Botswana. For South Africa — where the same story runs through Blue/Green Drop regulation — see addanode.co.za.
Frequently asked questions
Which African country has the highest non-revenue water?
Among markets with recent public reporting, Ghana (~54% of production) and Kenya (48% nationally, per WASREB's FY2024/25 report) sit at the top of our working set, with Botswana around 45%. League tables shift with each reporting cycle — the more useful comparison is a utility against its own zones, month by month.
What is an acceptable NRW level?
Global good practice is commonly benchmarked near 20%, with world-class utilities below 10%. For most African networks the realistic near-term win isn't the benchmark — it's naming and fixing the worst two or three zones, which routinely recovers water worth multiples of the monitoring that found it.
Does prepaid water metering reduce NRW?
It reduces commercial losses — collections and reading — and Tanzania's national experience shows that working at scale. It cannot see leaks, bursts or bypasses. Paired with independent zone metering, though, prepaid data becomes powerful: vends versus zone inflow is the cleanest physical/commercial loss split available.
Where should a utility with a small budget start?
One or two telemetered bulk meters on the worst-suspected zone's inlets, plus night-flow trending. That's the smallest deployment that produces an actionable answer, and its documented recovery typically funds the rollout — the worst-zone-first pattern every programme here converges on.