I've started a few: a hosting company I eventually sold, a subscription website service, a video calling platform, an app. Some worked and some didn't. What follows is what I'd do today, with what I learnt from the ones that went wrong.
1. Validate before building
The most expensive mistake is falling in love with the solution before confirming the problem. Before writing a line of code:
- Talk to ten people who'd have the problem. Don't ask whether they like your idea (they'll say yes); ask what they do today and what it costs them.
- Find who already solves it. Competition is a good sign: it means there's a market.
- Try to charge before you have it. A deposit, a pre-order, a letter of intent. It's the only validation that doesn't lie.
2. Start with the smallest sellable thing
Not the complete version: the smallest one that genuinely solves the problem for someone. With Videomeet.me what mattered was that a video call worked from a link, not the admin panel I had in my head.
And use what already exists. Almost everything you think you have to build can be assembled from existing tools for the first few months.
3. The minimum needed to charge
- A site explaining what you do and who for, with a clear way to get in touch or buy.
- A way to take payment that works from the first euro.
- Professional email on your own domain.
- Analytics and Search Console from day one: the data you don't collect today you can't recover.
- The legal minimum: registration, invoicing, legal notice and privacy.
4. Talk about what you're doing while you do it
Don't wait for perfect before explaining it. Writing about the process brings clients before the finished product does, and it forces you to order your thinking. Almost all my projects got their first users that way, not through advertising.
5. Decide early what this is
A side project, a business that has to feed you, or something you want to sell one day. All three are built differently, and mixing them is the fast route to working hard and getting nowhere.
What I wouldn't do
- Spend on branding before validating. The logo can wait; the customer can't.
- Build for six months in silence. If nobody has seen it, you know nothing.
- Raise money to start. With what tools cost today, the first stretch can pay for itself.
- Copy a competitor's price. Their cost structure isn't yours.
Got an idea and not sure it holds up?
Write to [email protected]. I'll tell you what I'd validate first and what would give me pause.