You have a task to finish by a given deadline.
One strategy: start now, work on for a while, and come back to refine it over and over until the deadline. If it's an essay, you write a rough draft and then make a bunch of revisions. If it's a sculpture, then you first make the big cuts to get the shape of the stone roughly right, then shave and sculpt the remaining until it is what you want it to be. If it's a big party, you get the venue and the food figured out, then come back to decide the playlist and centerpieces and so on.
We might call this the Zeno strategy, named for the 5th century BC Greek philosopher whose paradoxes you probably have heard of. Basically, we imagine that you might do 1/2 of the work at first, then later come back and do 1/4 of the work, then 1/8, then 1/16, until your sum converges (i.e. the deadline arrives) and you submit/finalize/have the party. This is usually what is recommended in e.g. college classes: "Start on your final projects now, so you don't have to rush it later!"
There are a number of advantages to the Zeno strategy. First, you give yourself extra time to think through your vision, change your mind, or even start over altogether. Second, you can notice and solve problems as they arise; sometimes when I'm writing a long piece of code, I can't find the bug the day I'm working on it and only figure out the problem the next day in the shower.
But there are disadvantages too. You will spend a lot of time reading+rereading, or painting+repainting, or whatever. If you're particularly flighty, you might change your mind 10 times and do 10x the work before you finally settle on one of them. Maybe you'll spend a lot of time calling every DJ in the city to find the "best" one, or save an extra $5, when it is not worth your time and you would have been totally fine to just go with the first Google search result (see: maximizers vs satisficers). In the worst case, you might make an irrecoverable change that would not have been possible if you hadn't had so much extra time to overthink things.
A second strategy: wait until the deadline is relatively near, and then kick into high gear and get the whole thing done "all at once" (in this context, this could mean a day or a month). If it's an essay, you write it the day before it's due. If it's a piece of art, you get in the zone and finish it up all at once, with no danger that you'll come back in a different mindset and ruin it. If it's a casual get-together with friends, you call the restaurant the day-of and pick up a six-pack, and figure that's all the planning you need.
I'll call this as the Cunctator strategy, the nickname of Quintus Fabius Maximus Verrucosus*, the Roman general of the 3rd century BC who famously invented the novel strategy of not sending all your troops into a head-on, all-out war against your enemy (as was the style at the time). He instead opted to wait, strike only when necessary and against advantageous targets, and won many battles against Carthage.
One advantage, as discovered by Cunctator himself, is you can step back and take a global view of the situation, and only engage in those sub-tasks (battles?) deemed necessary at a given time; action is a cost and should be paid only when necessary. On the other hand, including this advantage may seem unfair; this isn't what my professors were warning against when they told me to avoid procrastination!
Even if you literally don't engage with the task until you deem it necessary (the day or the week before it's due, maybe), there are still advantages. Partly, the advantage is to avoid the failure modes of the Zeno strategy. You can't overthink the task because you simply don't have time to do so. You just pick a DJ that will do a great job, without agonizing over which is the absolute top one in the city. And sometimes, all you need is a B or a C (not an A+), and so you can 80/20 Rule yourself into a quick result that's חֲפִיף and get back to more important tasks (see: half-assing it with everything you've got).
I used to play Chess in person with my brother, and when we moved apart we tried playing online in a non-real-time platform. I found it took so much longer, not just in actual time but in game time, and we both played worse, because each time one of us opened the app to play a move we had to re-examine the board and re-figure out our strategy. The Cunctator suggests you just play a whole game in real-time whenever possible, or more generally, watch out for fixed-cost attention penalties incurred when you bounce between different tasks; just get the damn thing done and (only then!) move on to the next.
Of course, if you're not good at managing your time and anticipating how long things will take, a Cunctator strategy risks a hurried and poorly-executed final project. When unexpected problems arise, you may not be able to think them through carefully, and may have to live with mistakes. Fatigue is also an issue: I've wasted hours trying to work through a problem on little sleep / a fatigued brain, only to solve the problem easily the next day.
\\
When should you be a Zeno and when should you be a Cunctator? I feel like I was told all my life why I should be a Zeno always and with everything, because Cunctator = procrastinator and procrastinator = bad, and yet I can't help but notice that nobody thinks you should take your garbage all the way to the dumpster every time. So there must be proper uses for the Cunctator strategy outside of Carthaginian wars.
As with everything, there's a spectrum here and the advice that could help you is an arrow, not a destination. I don't have a full model of this, but I'll outline some things I think are relevant.
On one hand:
If a task has many moving parts or parties that need to be coordinated (think a wedding or other large event with many participants), then you should think more like a Zeno;
(Related) If a collaborator is waiting for you to complete your part of the task in order for them to begin theirs, think more like a Zeno;
If a task has known unknowns/failure modes that are difficult to anticipate, think more like a Zeno;
If a task has a very high bar for success (think an important job interview or writing a book), think more like a Zeno.
Note that for a large project with multiple moving parts, the project itself may lend itself to a Zeno strategy, whereas any particular sub-task of the full project may be better handled by a Cunctator.
On the other hand:
If it's a well-understood and not-so-critical task and you can confidently predict how long it will take, you should think more like a Cunctator (though, beware!);
If you need important input from someone else before you can do your part of the task, think more like a Cunctator (else you might have to go back to the beginning if your collaborator surprises you);
If working on a task incurs a fixed cost every time (think walking to the dumpster every time), think more like a Cunctator;
(Related) If a task is highly technical and incurs a fixed mental cost when you start (think non-RT chess or writing a difficult piece of code), think more like a Cunctator.
Finally, note that this isn't completely up to you; I think some people are natural Zenos and some are natural Cunctators, and where they should move on the margin may be dependent on them as much as it is on the specifics of the task.
For example, most of the time I employ the Cunctator strategy. I work best when I have things on my calendar and due dates set out in advance, and I work on the most pressing one as it comes up. As a result, I try to work on one thing at any given time. (I basically don't believe in multi-tasking, because ~evidence but mostly because it feels identical to distraction to me on the inside.) Sometimes I'll get an email and change gears, but when I do I switch fully; I advance my other project to a natural stopping point and close it before starting the new one.
My wife is much more of a Zeno. When her boss gives her a task, even if it's not urgent, she can't help but work on it right away. But she works several part-time jobs (mostly remote/at-home), so there are many things coming up all the time, and she ends up working on many things at once. She tells me things like, "So I started working on the e-book for [her boss], then flipped over to research clients for [her other boss], then started making graphics, then went back to the e-book, then called the embassy, then..." It sounds like craziness to me, but she insists that this works much better for her than doing one thing until it's done. Either she's wrong, or (as I've come to think) her mind just works differently from mine.
I expect that one can derive the proper advice-direction by looking at their recent failures. Do you tend to get everything done way in advance, but over-write and constantly re-edit your work? Maybe it makes you anxious and overwhelmed all the time? Then it's probably time for more Cunctator. Or, do you fail to hit your deadlines, feeling anxious at the last minute because you didn't give yourself enough time to do a good job? Then let Daddy Zeno help you out. Thinking carefully about how to balance these Greek and Roman styles can improve your quality and quantity of output in the long run.
--------------------------------------------------------------------------------------------------------------------------
* Yes, this is the same guy whose tactics led to the term Fabian strategy, which I am intentionally not using, because my focus is on the postponement rather than the strategic strikes; in English today, Cunctator is a synonym of procrastinator.