When it pays off and when it does not
| Worth building | Not worth building |
|---|---|
| You run a process differently from your competitors | Ordinary admin — accounting, payroll, e-mail |
| An off-the-shelf system would need heavy modification | A ready-made tool already fits 90% of the need |
| Per-user licences grow faster than the investment | You have a handful of users and simple admin |
| Data gets retyped by hand between three places | You do not yet know what the process should look like |
| You need to own the data and have it at hand | The main criterion is the lowest possible price |
What usually gets built
- Internal systems — orders, warehouse and dispatch, job tracking, attendance
- Integrations — so the e-shop, warehouse and accounting pass data to each other
- Automation — work someone does by hand every single day
- Reports and dashboards — numbers you currently wait for until month end
- Work on what already runs — adding the features your current software is missing
An approach that does not end in a half-finished system
Most failed projects die because six months go into building the “big solution” and only at the end does it become clear what should have been done differently. The reverse order works better.
- Find the most expensive annoyance. Not the most visible one — the one that costs the most time and mistakes.
- Solve it on its own. A small working piece people start using within two weeks.
- Let real use do the talking. After a few weeks it is obvious what comes next — and it is usually not what the original plan said.
- Add piece by piece. Each one builds on what already proved itself.
There is a financial upside too: you can stop at any point. You are not left with a half-built system, but with several things that work.
What to check when picking a supplier
- Who owns the code. Ask outright. The repository, servers and access should belong to you, not the supplier.
- Who will actually write it. Agencies often hand the work down to juniors. Ask for a name.
- What happens after delivery. Software without maintenance ages. Find out what running it costs and how fast an outage gets fixed.
- Whether you can start small. A supplier who demands a six-figure analysis first is shifting the risk onto you.
- What it runs on. Exotic technology means nobody will take it over from them in three years.
How I work
I am one person, not an agency. There is no team of juniors the work gets passed down to and no subcontractor in another time zone — which is why I take a limited number of companies at a time. I work on a flat monthly rate rather than by the hour, so there is no estimate to inflate.
Most often I build web applications in JavaScript and TypeScript over a PostgreSQL database, deployed on Cloudflare or Vercel. If you already run something, I build on it instead of rewriting it.
I am based in Czechia and work in English as well as Czech.