Website Architecture
Information architecture, URL structure and content flow decided before design starts, because retrofitting it is how sites end up with five pages competing for one query.
What does architecture decide?
What exists, what it is called, where it lives, and what links to it. Which is really a set of commercial decisions wearing a technical hat: whether services split by what you do or by who you do it for, whether locations get their own pages, whether the blog is a category or a dumping ground.
What goes wrong without it?
Cannibalisation. Two pages targeting the same intent, splitting the signal between them so neither ranks. Or an orphan section nobody links to, discovered by nobody. Or a URL structure that cannot accommodate the next service without a migration.
How is it documented?
A sitemap with intent attached to every page: the query it answers, the audience it serves, the action it asks for. A page that cannot fill in all three does not need to exist.