Contents
On 1 October 2026 Nigeria gave every addressable building in the country a postcode. The Federal Ministry of Communications, Innovation and Digital Economy announced the National Digital Alphanumeric Postcode System that day, operated by the Nigerian Postal Service. The Postmaster General, Engr. Tola Odeyemi, put it in one line: “Everyone in Nigeria has a location. Now every location has a postcode.”
For a business that delivers anything, that matters more than it looks. Addressing here is not cosmetic: it sits in the cost line, as second attempts, returned stock and riders on the phone. Whether the codes are accurate will settle itself over the next year. The harder question is what a business is permitted to do with a code, and what it should keep in its own records when a code is missing, wrong or not allowed for the purpose intended. NIPOST’s terms of service answer parts of that, more narrowly than the launch coverage suggested.
What is happening?
The system assigns an 11-character alphanumeric code to each addressable building, derived from Geographic Information System data. The NIPOST portal sets out five levels behind it: state, local government area, district, area and building. Individuals look up their code at postcode.gov.ng or in the Android app.
The commercially interesting part is the organisational route: a business can create a developer account, obtain an API key and integrate through the API, SDKs or bulk access. Section 4 of the terms lists lookup, address validation and standardisation, authorised geospatial services, bulk processing and enterprise services; section 10 says premium API access, bulk processing and enterprise services “may be subject to fees”, in naira.
The stated rationale is cost. At a June 2026 workshop reported by BusinessDay on 23 September, Odeyemi estimated the cost of misdeliveries from poor addressing at “anywhere from N50 billion to N80 billion” a year — his estimate, published without a methodology, and best read as an order of magnitude.
The first week was mixed. Nairametrics reported on 2 October that some users found the lookup precise, while others found duplicate codes for neighbouring buildings or no code at all; NIPOST says corrections requested through the app can be completed within 48 hours. Three things remain unknown: how many buildings carry a code, how accurate the register is independently, and what API and bulk access cost.
Why it matters economically
Failed delivery is one of the few operating costs in African e-commerce that is audited and disclosed. In its annual report on Form 20-F for 2025, filed with the US Securities and Exchange Commission on 24 February 2026, Jumia states: “In 2025, failed deliveries represented 25% of the gross orders in Cash on delivery.” Its cancellations, failed deliveries and returns rate rose from 22% of GMV in 2024 to 26%.
That is a group figure across several markets, not a Nigerian rate, and Jumia’s order mix is not a distributor’s. But the filing explains mechanics that generalise: three delivery attempts, then return to the seller, with commission uncollected and fulfilment costs carried by both the platform and its sellers.
The arithmetic is easy to run. An illustrative distributor dispatching 400 orders a week at ₦2,500 per attempt, failing on 15% at the first attempt, spends roughly ₦7.8m a year on second runs before counting the orders abandoned. Halving that rate is worth about ₦3.9m. The assumptions are illustrative; the number itself is computable from a business’s own dispatch records, and most cannot produce it.
What it means for African businesses
The code has no standing outside Nigeria, so a business moving goods across several logistics and trade corridors gains nothing for its Ghanaian or Kenyan deliveries from a Nigerian register. Ghana is the useful comparison, because it has had one for nine years. GhanaPostGPS was unveiled on 18 October 2017 and later required for the national identity card, bank accounts, business registration and passports. In an analysis published in April 2026, the Ghanaian technology writer Joseph-Albert Kuuire argues that the millions of codes generated reflect compliance rather than daily use, in part because the system did not integrate natively with the mapping and ride-hailing applications people actually navigate with. We found no published measurement of how many Ghanaian businesses or couriers use it day to day, and that absence is the finding: adoption of an address standard is recorded nowhere, so no business should assume it.
So the question in Nigeria is not whether the code is official but whether the party at the other end uses it: the courier, the rider, the bank, the field technician. For businesses in e-commerce and retail, distribution and field services, that is a question to ask suppliers this month; for one whose customers collect from a counter, it changes nothing. And in many Nigerian neighbourhoods an address is a landmark plus a phone call — a rational system where street naming is incomplete. A national code does not replace it. It adds a machine-readable anchor beside it, which is the honest claim for it today.
Technology and AI implications
The API is the actual change, not the code. A postcode a customer types into a form is another string that can be mistyped; one your order system resolves and passes downstream to dispatch survives the handoff. That is where the value sits.
AI contributes almost nothing here, and it is worth saying so plainly. The task is not prediction; it is the capture and verification of a field. The familiar pitch — an AI layer that “understands” messy African addresses — describes a parsing problem better solved by having the customer confirm a resolved location. One narrow exception keeps a person in the loop: language models are reasonable at turning free-text addresses you already hold into candidates for someone to confirm in bulk.
Where technology does not help: if a building has no code, no integration creates one; if your courier has no field for it, the code stops at your database. The terms describe the service as “AS IS” and “AS AVAILABLE”, with no guarantee of availability, completeness or accuracy — so a dispatch process that cannot run when an external API is down is worse than the one it replaced.
Practical business implications
Read the licence before designing anything around it. Section 6.1 grants a “limited, non-exclusive, non-transferable and revocable license”, and sections 6.2 and 7.2 transfer no ownership of the database, the APIs or the postcode structures. Section 7.3 says users must not “reproduce, distribute, modify, commercially exploit, publish or create derivative databases” from the materials except as expressly permitted — so building a parallel address database out of your order history is not on the table. Sections 16.3 and 16.4 require users to stop using the data on termination, with NIPOST able to demand secure deletion.
The limits on verification matter more. Section 5.4 says users must not present the information as independent verification of identity, ownership or occupancy; 5.5 says a postcode alone establishes none of those, nor title or legal status. That is materially narrower than the framing in some launch coverage, which presented the system as tightening bank onboarding checks. A postcode tells you a building exists and where it is. It does not tell you your customer lives there. Section 9 also makes anyone processing personal data from the system responsible for their own compliance under Nigerian data protection law.
Four things worth doing, in order:
- Measure before you integrate. Count last quarter’s first-attempt failures, second attempts, returns and the reason recorded against each. If no reason is recorded, start recording one this week. That number decides whether the rest is worth doing.
- Store the resolution, not just the code. Keep the postcode, the coordinates it resolved to, the date, and the customer’s own description of the place: codes get corrected, and your record of what you dispatched against should not change underneath you. Hold all of it beside the phone number in your customer and order system.
- Confirm before dispatch. A message to the customer with the resolved location, answered before the rider leaves, removes more failed deliveries than any code does by itself. Ask your courier what it accepts, too: with no field for the code, it buys nothing downstream this quarter.
- Check the plan against the terms. If it depends on retaining the data indefinitely, redistributing it, or using it to verify who someone is, the licence does not permit it.
RevenueStack perspective
Our view: the economics here are not in the code, but in a measurement most businesses do not have.
First, a national address standard lowers the cost of finding a building. It does nothing about the two costs that dominate failed delivery: the customer who is not there, and the order placed on a payment method that lets them decline at the door. Jumia’s filing makes exactly that distinction, noting that prepaid orders drive much higher delivery success rates than payment on delivery. The address was never the whole problem.
The second point is about ownership. The long-promised fix for African addressing has arrived with a licence saying the data is not yours, that you may not build a derivative database from it, and that you must be able to delete it. That is reasonable for a public register. It is also a design instruction. The durable asset is your own record of where a customer actually received a delivery, keyed to an order or a customer record you control, with the national code held as one verified attribute rather than as the record itself. A business whose customer data has the postcode as its spine is exposed when the terms or the fees change; one holding the code as an attribute is not. It is the same discipline as treating a payment rail as a feed and your own system as the record, and how we approach any external dependency.
Third, on sequence. The register is eight days old and its first week produced duplicate codes and missing buildings — ordinary at this scale, but a reason not to rebuild a dispatch process around it this month. The work above pays whether or not the postcode succeeds, and makes adopting the code cheap on the day it does.
Conclusion
Nigeria now has a building-level address standard with an API, and it is genuine national infrastructure. It will not reduce anyone’s delivery costs on its own. What reduces them is knowing your own failure rate and its causes, confirming destinations before dispatch, and holding address data in a structure you control, with the national code attached as one verified field. Do that and the postcode is a cheap upgrade when coverage and courier support arrive. Skip it, and it is one more string nobody downstream reads.
Sources
- Federal Government Launches National Digital Alphanumeric Postcode System — Federal Ministry of Communications, Innovation and Digital Economy, Nigeria, 2026-10-01
- NIPOST Postcode — National Digital Alphanumeric Postcode System — Nigerian Postal Service (NIPOST), 2026-10-01
- National Digital Alphanumeric Postcode System — Terms of Service (read 9 October 2026) — Nigerian Postal Service (NIPOST)
- Nigeria's new digital postcode targets N80bn delivery losses, tightens bank KYC — BusinessDay, 2026-09-23
- Nigerians report mixed results as new digital postcodes go live — Nairametrics, 2026-10-02
- Annual Report on Form 20-F for the year ended 31 December 2025 — Jumia Technologies AG, filed with the U.S. Securities and Exchange Commission, 2026-02-24
- Ghana Tried To Solve Its Addressing Problems With a Digital Platform. The Results Have Been Subpar — Tech Labari (Joseph-Albert Kuuire), 2026-04-07