On returning to Kaduna to build a software lab.

There is an assumption in Nigerian technology, rarely stated because it does not need to be, that serious software happens in Lagos. If not Lagos then Abuja. If not Abuja then somewhere with a different flag on the passport. Everywhere else is where you are from, not where you build.

I want to write down what changes when you stop believing that, because the answer turned out to be both less and more than I expected.

What does not change

Almost all of the work. This is the part people find deflating and it is the most important thing to say plainly. The architecture is the same architecture. The deployment target is the same. The clients are reachable from anywhere, the code review happens in the same place it would have, and nothing about the difficulty of the actual engineering moves an inch based on which city the engineer is sitting in.

The obvious counter is talent density — that you cannot hire in Kaduna. It is true that there are fewer senior engineers here than in Ikeja. It is also true that a small company does not need density, it needs a handful of specific people, and the market for those people has been national and remote for years. The density argument is real for a company hiring forty engineers. Below that it is mostly borrowed anxiety.

What does change

Proximity to a different set of problems, and this is the whole of it.

The businesses within driving distance of this office are not the ones being built for. They are grain traders, distributors, schools, clinics, logistics operators, family firms in their second generation — businesses running real volume on systems that no product manager in Lagos has ever sat in a room with. Their constraints are specific: a network that goes when the power goes, staff who work in Hausa, customers who will not fill in a web form and will send a voice note, buyers who need to see a registration number before they will trust an invoice.

You can learn those constraints from a distance. But you will learn them as a list of adjustments to a design you have already made, and the design will keep quietly assuming the user it was drawn for. Being here means the constraints arrive first, while the design is still soft. That is not a moral advantage. It is an information one, and it shows up in the work.

The version of this that is not romantic

There are real costs and I would rather name them than write the other kind of essay about this.

Serendipity is thinner. The conversation you would have had in a Lagos co-working space, the introduction that happens because someone overheard you — those do not occur, and they have to be replaced with deliberate effort that is less efficient than luck. Investors are further away, and geography still functions as an unstated filter no matter what anyone says publicly. And a certain kind of client will read the address as a signal about the work, which you answer by shipping rather than by arguing.

None of that is fatal. All of it is a cost, and pretending otherwise would make this piece useless to the person actually weighing the decision.

Why it is the right trade anyway

Because the assumption that the work has to happen elsewhere is the same assumption that produces software built for a Nigeria that mostly does not exist — the one with reliable power, a single language, a customer who fills in forms. Most of the country is not that and most of the businesses in it are not either. Building where those businesses are is not a sacrifice made for sentimental reasons. It is the cheapest way to stop being wrong about who you are building for.

The first client build out of this office was for a grain supplier an hour up the road in Zaria — a business whose two audiences read different languages and whose buyers needed to verify a company registration before committing to a delivery. Those requirements were not gathered. They were obvious, in a way I do not think they would have been from anywhere else.

That is the entire argument, and it is a narrower one than the essays about diaspora return usually make. Not that everyone should come home. Just that the assumption is worth examining, and that when I examined mine it did not hold.