Buying Operations Software in Somalia: A Straight Guide

Most software buying guides are lists of features to check. This one is four questions, because in this market the feature lists are not where vendors differ — what separates them is which parts of the country they have understood, and you can establish that in a single meeting if you ask the right things.

One caveat before the questions, and it applies to us as much as to anybody: we are a Nairobi-built product with no office here and no implementation partner. Treat everything below as questions to put to us with the same bluntness you put them to anyone else.

Question one: if I receive the same item through two ports, what does your system hold?

This is the question with the highest information per second in this market, and most vendors have never been asked it.

Berbera, Bosaso and Mogadishu are administered separately, so what a consignment cost to bring ashore depends on which one cleared it. Two containers of one product, same supplier, same day, different total cost. A system holding one cost per item averages the two and from then on every margin you report is measured against a figure that describes neither consignment.

Our own answer: cost is held per consignment and stays open after receipt, so late clearing and handling invoices reach the goods they belong to and two landings keep two costs. Making the port itself a dimension you can group and compare margins by is a defined build rather than a live feature. The full argument, with the arithmetic, is in one item, two landed costs.

Question two: which revenue authority do you file with here?

This is the trap question, and it is deliberately placed second.

There is more than one administration collecting revenue in Somalia. What your organisation owes, where, and to whom is a question your own adviser and your own forwarder deal with per consignment and per port — they will know, because they do it. A software vendor who answers this crisply in a sales meeting is describing a country simpler than the one you operate in.

A related one worth asking everybody: what rate do you have configured for this country, and where did you get it? Ours read 10% for years. It has been 5% since 18 August 2024, we corrected it in August 2026, and it was the first time we had found one of our own rates wrong in the direction that overcharges. It survived that long because the row was unreachable by the code that resolves a rate, so nobody ever saw a figure derived from it and nothing contradicted it. A vendor's country list is not evidence — including ours. Confirm the figure with your own adviser.

Separately, and worth knowing before you configure anything: this is a sales tax with no input credit, so tax paid on a purchase is part of what the purchase cost rather than something you reclaim. We have argued that mechanism at length elsewhere and will not repeat it here — see the tax you paid is part of what the goods cost, written for a neighbouring market that levies the same kind of instrument.

Question three: how does money actually leave your system?

Mobile money penetration here is around 73%, the services are dollar-denominated, more than half the country is paid that way and most of those never cash out. So the rail your money genuinely moves on is a wallet — EVC Plus, Zaad, Sahal, e-Dahab — and the honest question is not whether a vendor supports "mobile money" in the abstract but which provider, which direction, and what you still do by hand.

Our answer, stated as the absence it is: we connect to none of them. Our one collection integration is M-Pesa and it is Kenya-only. What runs today is that the employee or supplier record holds the wallet as a destination, the disbursement is keyed and coded to a project, grant or consignment, and the register gives you one place to reconcile against the telco statement. Initiation and reconciliation are manual. This is on the roadmap and commissionable, and it is the item we would expect to be asked for first — the precedent is not hypothetical, because M-Pesa exists in this product only because Kenyan clients needed it and paid for it.

Question four: do you screen my suppliers?

Grant funding and correspondent banking here bring real screening obligations, and this is the question where a reassuring answer is the dangerous one.

Our answer is no, and it is the most important no we publish. We hold the supplier record, the documents, what was proved about a counterparty and by whom, and the date any of it expires — so the evidence is retrievable and its staleness is visible. We check no list, we consult no register, and we are not a screening service. This is a boundary rather than a backlog, and the reason is specific: the failure mode of a half-built screening feature is a user who stops looking. Buy screening from someone whose business it is and use an operations system for the evidence trail underneath it.

Two of these four questions are ones where a confident vendor answer should reduce your confidence rather than increase it. That is unusual, and it is the most useful thing to know about buying software for this market.

The things nobody asks and everybody should

1. What language will the people keying transactions be working in?

   Somali and Arabic are the official languages and English is widely used in business — which is not the same as everyone on your finance floor working in it comfortably. Our interface is English only. Test that with the person who will use it daily, not the person signing the contract, and establish it in week one rather than month three.

2. What happens when the person who implemented this leaves?

   The single best question in software buying and it works everywhere. Ask it of every vendor including us. The answer reveals whether the configuration is documented and self-serviceable or whether it lives with one consultant.

3. Can we capture where there is no signal?

   A yard or a holding station is not where the network is. Ask which operations work offline — specifically which, because "mobile app" and "works offline" are different claims. Ours covers four field operations; counting, receiving a purchase order and job time are not among them.

4. Does the working day actually overlap?

   Mogadishu and Nairobi are on the same clock, so a Nairobi vendor gives you a full shared working day. A European or Gulf vendor gives you a partial one. This is small, real, and easy to check against whoever is selling to you.

Where we are the wrong choice

Worth stating plainly, because a guide that concludes "buy from us" is not a guide.

If your decision turns on a live wallet integration, buy from someone who already ships one — they are starting where we would be finishing. If it turns on filing, or on anybody telling you authoritatively what your obligation is and to whom, we are not that and neither is any software. If your finance team's working language is not English, establish that before anything else. And if you need screening, buy screening.

Where we are worth talking to is narrower and we would rather be believed about it: keeping two landed costs true when the port is the reason they differ, holding stock that is yours and not yet shippable as its own state, and being able to answer a funder or an auditor from a report instead of from somebody's recollection.

What is built, what is on the roadmap with a price on it, and what we would decline from a paying customer is set out in three columns on the Somalia market page. The season's stock problem is in stock that is yours, counted, and cannot ship.