Choosing boring technology for a brand-new startup
A company gets maybe three innovation tokens. Spend them all on infrastructure and you have none left for the thing the customer can actually see.
We're building a company. After years of client work, a school management platform — the idea I've been circling since I watched a clerk hand-total attendance registers in 2012 — is finally a real product with real users. Which means the tech stack decision is no longer a Friday-night debate; it's a bet with other people's money.
Innovation tokens are real
There's an essay I keep sending people — "Choose Boring Technology" — whose core idea is that a company gets maybe three innovation tokens to spend. Every shiny choice costs one: the hip database, the six-month-old framework, the architecture from that conference talk. Spend them all on infrastructure and you have none left for the actual product.
So we chose boring, deliberately
- PHP and MySQL — not because they're fashionable (they aggressively aren't) but because we know their failure modes by heart, hosting is a commodity, and every hire in this market can be productive in week one.
- A monolith — one deployable, one log file, one thing to reason about at 2 AM. Microservices solve problems we don't have yet, at a price we can't afford yet.
- jQuery on the front — I've written admiringly about React, and when our interactivity earns it, we'll adopt it. An attendance form does not earn it.
Boring technology isn't the absence of ambition. It's spending your ambition where the customer can see it.
The uncomfortable truth about early-stage products: your technical risk is almost never "will it scale?" It's "will anyone want it?" Schools don't care what's under the hood — they care that fee vouchers print correctly on the first of the month. Boring under the hood is how we make sure they do.