Four ways of working together, depending on what is broken
The expensive technical decisions are made at the start, when you know the least. Stack choice, architecture, what to build first and what not to build at all.
From validating the idea to a product in production with active subscriptions: MVP, multi-tenant architecture, billing.
ERP, e-commerce, payments, CRM. Stop retyping the same data in three different places.
Technical SEO on the things you can measure: indexing, Core Web Vitals, structured data. Plus extractability for whoever answers instead of Google.
Four steps, and the first one also tells you whether the second makes sense
Working out what is broken and whether it is a technical problem. Sometimes it isn't, and it's better to say so straight away.
What we will do, in how long, with which verifiable outcome. Before starting.
Short iterations, with something to look at at every step instead of a single delivery at the end.
We measure what we said we would measure. If it hasn't moved, we say so.
Where to start, how long it takes, who works on the project
It starts with a short conversation, which may well end with "this isn't worth doing".