After you divide the work up into small bites, I know in some cases, those small bites have to be done in a particular order, but on some jobs (or multiple jobs) you can decide for yourself which to do first. So how do you decide.
Personally, I always liked to do the worst jobs first, so wherever I was in the process, I could always figure the worst is behind me.

But I remember when I just a kid that my Dad said he liked to start with the easiest job, so he could always figure he was doing the easiest job available.
After sub dividing a large or complex task into sub tasks and maybe some of those sub tasks into sub sub tasks (lather-rinse-repeat as needed) you may find some tasks are independent and can be done at any time while others have a rigid sequence . For example: you don't install the shingles before you build the roof but you can build the fence around the house any time.
Decomposition or successive decomposition is the professional term for breaking a task into its component parts. With experience you learn how far to go in the decomposition of tasks into ever smaller tasks. Once a task is decomposed into small enough pieces that you and or your compadres can wrap their minds around it then don't simplify any further. For example: lunch break shouldn't have to be broken down into simpler pieces for anyone.
A very important part of this systematic approach is to know what the required inputs are to a process and what the output(s) is(are) you get out. This relates directly to Bird's question of how do you decide. If a process requires say three inputs... You need electricity to the job site area of this sub task, you need water to this location too, and you need one particular non load bearing stud wall framed up to accept the installation of some thingy. So before you uncrate the thingy and try to install it you first have to have electricity and water and the stud wall.
Experienced contractors sort of employ this process casually and informally to build a simple house or whatever. Architects, engineers, and tradesmen erecting a sky scraper tend to get a bit more formal with their plans and have detailed written descriptions of the broken down tasks.
Not everyone can intuit the flow of work in a complex job, especially if it requires coordinating the efforts of various groups of workers or even applying your own labor in the proper sequence of events. I find that I NEED to think about a larger or more complex job in these terms when the consequences of "messing up" are not trivial.
I am currently in the process of moving a 35x70 foot metal bld with 18 foot side walls and 14x14 foot roll up doors at opposite ends. I am moving it about 1/4 of a mile across one of my fields to a more convenient location. One of the early sub tasks was to empty the bld. But... what do you do with the contents in the mean time? So a precursor task to emptying the building was to build some 4x4x8 ft boxes to store stuff. I also emptied out 2 each 20 ft steel shipping containers, dragged them out of the bld and refilled them with stuff.
I also had a pad constructed to accept the bld. Now I need to have piers dug by my driller who can do 16 inch piers up to 24 ft deep. I need to fabricate weld plates to set in the concrete of the cast piers. I need to build "boots" to raise the pipe columns of the bld to make up for the fact that I will not be pouring a floor or a foundation or stem wall. The current situation has a ledge in the poured floor to accomodate a few inches of overlap of the metal siding. If I just sit the cut off columns on the weld plates I will crush a few inches of the bottom of the siding all the way around the building. So I will make "boots" out of the next larger size pipe to raise the building 6-8 inches. I will then be able to see "daylight" all around the building. I can use pressure treated wood set into the soil an inch or two and be screwed to the metal siding. Thus no earth metal contact, just wood to dirt and metal to wood.
This is my first building move of this sort so I had to break it down into bites I could easily understand and chew one at a time from start to finish to be sure everything was going to flow smoothly. Now it is raining and I can't get the driller on site. If it rains enough with insufficient time to dry between rains I may be "rained out" and the next window of opportunity will be after the ground freezes enough to take the load (if the winter is cold enough.)
Anyone have a discount ticket for a heavy lift helicopter?
Oh by the way there are lots of different pieces of competing software that simplifies decomposing a job into tasks and subtasks and tracks the input requirements and outputs. If estimated times to do the various tasks are input then the software will help you estimate several things rather well, including: the "critical path", lay times, total time to complete and such.
These software applications typically support PERT charts, a powerful management tool. For info, Google on PERT chart.
Pat