In 2021 I was a web developer taking client work on my own. Five years later DevKey Technologies is a small senior team in Islamabad with 50+ projects shipped for clients in Pakistan, Australia, the United States, Canada and South Africa. This is the short version of what I learned going from zero to one, written for anyone in Pakistan thinking of doing the same.
Start with one thing you can finish
The first mistake I nearly made was offering everything. Web, mobile, desktop, design, marketing. The thing that actually got the company moving was finishing small websites fast and well for local businesses, and being easy to reach. Breadth came later, from the team, not from me.
The lesson: your first product is your reliability. Pick a scope you can deliver on time, and deliver it on time, ten times.
Clients abroad come from clients at home
Our first Australian client did not come from a job board. It came from work done for a Pakistani business that someone's relative in Melbourne noticed. Local work builds the portfolio that overseas clients check. Do not skip it because the rates are lower.
Write everything down
The single biggest upgrade in how we work was moving from calls to writing: written scope, written questions, written sprint updates. It made remote clients comfortable, it made estimates better, and it made the team less dependent on me being awake. If you run a company across time zones, writing is the product.
Hire senior or do not hire
Early on it is tempting to hire juniors because they are cheap. They are not cheap; they are slow and they need someone senior watching them, which was me, which meant I stopped shipping. We now hire when a project pipeline needs people, and we hire people with shipped work. A small senior team beats a large junior one for the kind of clients we want.
Build something you own
Client work pays the bills and teaches you an industry. At some point take what you learned and build a product. For us that was ARTI Commission Shop, an offline desktop ledger for grain-market commission agents, and later Clinic Pharmacy. Neither came from a market report. They came from watching real people in Islamabad do their work by hand.
Say no early
The projects that hurt were the ones I knew were wrong on the first call and took anyway: vague scope, a client who wanted hourly billing with no end, a budget that did not match the ambition. Saying no in the first week costs nothing. Saying it in month three costs everyone.
What I would repeat
Finishing small things reliably. Writing everything down. Being on WhatsApp. Publishing the work. Building a product from a client problem. Hiring late and senior.
What I would skip
Offering every service on day one. Hourly billing. Hiring to look bigger. Spending on a nicer office before a better process.
If you are starting a software company in Pakistan and want to compare notes, I share more of this on the DevKey Tech YouTube channel, in Urdu and English.