When reading answers to questions like this we should always keep in mind an annoying truth about writing: What works for one will not work for all. Even what works for many will still not work for all. So sadly saying "what objectively works better" is a fool's errand, as the answers are highly subjective.
A solid foundation for projects in general tends to be to start with a core goal, and then begin working towards that goal by breaking the core goal into milestone goals, and re-evaluations of your core goal. Framing things around goals and evaluation [and expected rewriting] allows us to free ourselves from early decision paralysis so that we can begin a flow of steady useful work.
Initially the core goal of a project may be as vague as "I want to write a book", or "I want to build a house", but neither of those goals remains all that useful of an idea if you want to get any actual work done. It gives you nothing to actually go on, there are no details, and we aren't really left with anything to expand off of.
So we invoke the "re-evaluate the core goal", and start better defining it. What kind of book/house do we want? Ask ourselves: What kind of project are we able to tackle? Much like setting out to build a massive skyscraper as our first building project if we've never built so much as a shed, trying to set of writing something to dwarf The Wheel of Time as your first serious writing project is unlikely to be a 'great idea' if we want to finish a project of any reasonable quality. [Not that I totally set off to do just that as a teenager or anything... Nothing stops anyone from trying to start there, and if it is really what someone wants to begin with, even after being warned, then more power to them I guess?]
If we keep running re-evaluate cycles over our core goal, and breaking that out into sub-goals and milestones then we eventually end up at a point where we have focused things down to a point where we can begin writing either supporting documentation or actual content.
Core goal: "I want to write a book" - Too vague.
Re-evaluate: "I want to write a fantasy book" - More defined, but still vague.
Re-evaluate: "I want to write a fantasy book about [concept]" - Better, and we have something we can start splitting out into sub goals.
Sub-goal: "Define/explore [concept]" - We can begin brainstorming while making as few or as many detailed notes as needed. Always looking for new sub-goals/milestones, or points to re-evaluate.
A key point to keep in mind here is to apply this approach in as detailed or flexible way as what works for you. Maybe you run through several cycles of this in your mind while eating breakfast and start writing directly, or maybe you keep detailed journals with links to spreadsheets and mountains of research, plot or dialog fragments, or whatever. [I've written complex model simulation software for sci-fi projects just to help see if a space station setting idea I had actually made sense.]
Focus on what you need, and don't worry too much about what did or didn't work for other writers. Other writer's experiences are useful to consider, but are not laws to follow.
Keep in mind that unlike building physical things writing offers us the flexibility of reuse and reworking of previous efforts to improve on things. If we haven't built something solid enough to use as a foundation before we start putting up a house, then we are at great risk of having to rip everything down [assuming it hasn't fallen down on its own] and begin again from scratch.
But in writing we have the potential to pick some aspect of the project that we truly love and focus our attention there for as long as it pleases us. We can spend as long as we like making 'the perfect kitchen', down to all the finishing details, if we really want, and only then proceed to decide how to build a house around it. [Just be wary of growing 'too attached' to things before you're sure they can work as part of a larger whole. Drill down deep into a single thing if it works for you, but keep in mind it may require major changes or be difficult to fit into a larger scope later on.]
So 'start' with detailed plot overviews, or 'start' with detailed character overviews, or even just 'start' with an especially witty sounding bit of dialog. - Start with something, evaluate things, decide whether it works or not, and expand from there as needed.
Also keep in mind that writing projects typically aren't actually done in stone, and therefore are not immutable... The "Grand high fantasy adventure" I started writing as a teen has been abandoned at least three times, completely reworked from the ground up into a far more mature low fantasy project [aka, I read A Song of Ice and Fire...], directly influenced at least four mobile games, and had a clear impact on half a dozen different books and shorter series that have been finished and published.
In short, work on writing, and stay flexible. Both life and writing can take you to rather unexpected places.