—Process6 — One person, six stages
The reason this is worth writing down is that most of it usually happens at a handover between two companies. Here it does not happen at a handover at all.
- 01
Discover
Deciding what the thing is, and what it refuses to be.
- Your problem gets written down in one paragraph, in your words, and agreed.
- The work is cut: what ships first, what waits, what is not being built at all.
- The hard constraints surface early — the integration nobody mentioned, the data that does not exist yet, the thing legal will ask about.
- What comes out of it
- A scope, a sequence and a price, in writing.
- What you do
- An hour or two of your time and honest answers about what already exists.
- 02
Design
The product map first, then the screens it implies.
- The objects get named — what a record is, what states it moves through, who may touch it.
- Interface design follows from that rather than preceding it, so the screens cannot promise something the data will not support.
- The main flows are drawn end to end before anything is built.
- What comes out of it
- A data model and the screens, both agreed before code exists.
- What you do
- Reading it and saying where it is wrong. That is cheapest here.
- 03
Build
Client, server, database and the admin, as one job.
- It is built in vertical slices — one whole working feature at a time, front to back, rather than a finished frontend waiting on a backend.
- You get a running URL early and it stays running, so progress is something you can open rather than something you are told.
- The admin is built alongside the product, not bolted on at the end when there is no budget left for it.
- What comes out of it
- Working software you can use before it is finished.
- What you do
- Looking at it weekly and reacting. Nothing else.
- 04
Connect
Payments, mail, messaging, accounting, whatever already exists.
- The systems you already pay for are wired in rather than replaced.
- The failure cases are handled explicitly — the timeout, the duplicate, the webhook that arrives twice — because that is where integrations are actually lost.
- Anything that has to be reconciled later is designed to be reconcilable now.
- What comes out of it
- One system instead of several that disagree.
- What you do
- Access, and the name of whoever owns each existing system.
- 05
Launch
Live, with the boring parts done.
- Environments, domains, certificates, backups and the accounts that own them — in your name, not mine.
- Whoever will operate the admin is walked through it until they are not asking questions.
- Analytics goes in before launch, not after, so the first week is measurable.
- What comes out of it
- A running product, and every account belonging to you.
- What you do
- Deciding when. The technical readiness is my problem.
- 06
Iterate
What the first weeks show, acted on.
- The measurement that went in before launch starts answering questions, and the answers are usually not the ones anyone expected.
- Changes are made against evidence rather than against opinion, including mine.
- If the right answer is that nothing more should be built yet, that is said out loud.
- What comes out of it
- A product that gets better for a reason you can point at.
- What you do
- Deciding what is worth the next round.
—StartOpen for new work
Tell me what you are trying to build and what is making it hard. If there is a fit I will say so quickly; if there is not, I will say that too.
- A reply
- Usually within two working days, from me.
- A conversation
- Scope, constraints, budget, timing. No pitch deck.
- A straight answer
- Including when I am the wrong person for it.
01
02
03