They certainly wouldn’t work for months to get everything “right” before launch.
The goal is to try and avoid this scenario :
Breaking that upfront budget into smaller chunks and spending it based on information that you actually find out about your users/customers means that you are at market faster and not wasting resources on things that you don’t need yet.
This is THE way to build new sites and also a handy approach to refreshing existing ones. The only reason it seems unfamiliar to clients is because this approach is almost impossible in the physical world that they are often used to. If houses could be built this way (cost effective iterations based on factual information and changing needs) they would be. Maybe that isn’t too far off either.