Context
I started Saby Infotech in 2023 for a reason that sounds unfashionable in technology: I wanted to build a company that was calm. Not a company that raised money, hired fast, and burned out. A company that did a small number of things well, wrote everything down, and could be handed to a new engineer without three months of lost momentum.
The market I was building into is East African — Tanzanian businesses and institutions that have been told for years that software is either a cheap one-off website or an expensive imported system that does not fit how they actually work. Most of them had been burned by one or the other. The opportunity was not novelty. It was reliability: being the firm that ships, documents, and stays.
Saby today runs two workstreams. The first is modernising the operations of existing East African businesses — the unglamorous, high-value work of replacing spreadsheets, WhatsApp threads, and paper with systems that hold. The second is building the Smart School Management System, our own product for African education. Both are built on the same discipline.
A company that loses three months every time someone leaves is not a company. It is a group of people holding context in their heads. We write the context down.
Decisions
01
Decision
One stack, deliberately.
Most small agencies take whatever the client asks for and end up maintaining six stacks badly. We chose one — TypeScript end to end, Next.js, Node, PostgreSQL — and got disciplined about it. The cost is turning away the occasional job that wants something exotic. The benefit is that any engineer can move between any project in a day, every codebase looks familiar, and we are not the only people on earth who understand our own work.
02
Decision
Document decisions, not just code.
Every architectural choice that matters gets a written reason that outlives the conversation that produced it. Why this database, why this boundary, why we said no to that feature. It is slower in the first week and dramatically faster in every week after. It is the single practice that lets us onboard someone in days rather than months — and it is the thing clients notice when they leave a previous vendor for us.
03
Decision
Service-first, product-funded.
The consulting work is not a distraction from the product — it funds it, and more importantly it keeps us honest. Working inside real East African businesses every week is how we learn what software here actually needs to survive: unreliable power, expensive data, intermittent connectivity. That knowledge goes straight into the product.
04
Decision
Build for the real operating environment.
We design for the conditions our clients live in, not the conditions Silicon Valley assumes. Offline-tolerant flows, low-bandwidth payloads, mid-range Android as the primary device, mobile money as a first-class payment rail. The default assumptions of imported software are exactly where it fails here.
Outcome
Saby has grown the way I wanted it to — slowly, profitably, and without drama. The work compounds: documented systems get cheaper to extend, clients stay because the lights stay on, and the discipline that makes a project maintainable is the same discipline that makes the company calm.
2023
Founded, Dar es Salaam
1
Stack, end to end
30yr
The goal this funds
The metrics above are deliberately structural rather than vanity figures. Replace or extend them with your own verified numbers — clients served, retention, team size — before treating this as final.
- TypeScript
- Next.js
- Node
- PostgreSQL
- Vercel
What I’d do differently
-
I would have written the company's internal documentation earlier. We preached documentation to clients before we fully practised it on our own operations. The handbook, the decision log, the onboarding doc — those came a year later than they should have.
-
I underpriced the first year. Service-first does not mean cheap. Some of the earliest engagements were priced to win trust rather than to reflect the value delivered, and re-pricing an existing relationship is harder than pricing it correctly at the start.
-
I should have hired for documentation discipline, not just engineering skill. The best hires are not the fastest coders — they are the people who leave a trail. I learned to interview for that habit later than I would have liked.