Depends on the players.
Generally speaking, I start the group off with their back stories as a center point so I know the angles that the players are coming from. From there, I note the Runs they do, how they do them, and how well they pull them off all factor into the overall narrative that gets dispensed to them by way of the News and rumors. With a healthy dose of Corps-speak.
I guess it also depends on how much "leading by the nose" your players are willing to handle. Some groups expect to be hand held from objective to objective and never do anything more then the baseline required for the run, with little attempt to interact with the NPCs and environment. Meaning, they go to the meet, get the job - no trying to get anything extra or fishing for more info. they do the job exactly as told to/ put minimum effort into planning, then got directly to the Johnson to get paid. They never try to interact with anyone beyond "I call my bartender contact at rating 3/3 and ask if he knows anything about our run"
Other players seem to take extreme and perverse pleasure in trying to predict what the GM wants, and doing the exact opposite. "hmmm... The GM seems to be really pushing the narrative as this Johnson is super important... I SHOOT HIM IN THE FACE! - JUST THE SHITS AND GIGGLES!!"
And yet other players, get really engrossed into the atmosphere, and focus on the NPCs and their little pet personalities. "Yea, no, I really don't care what that Johnson wants, he not interesting at all. But Frankie that Gay Troll hooker who we talked to last week for that other job, he was interesting! We blow off the Johnson and go talk to Frankie, see what he's up to, and see if he has anything we can help him out with. Yea, he may not be able to put cash in our pockets, but he was really cool man!"
So my advice is: Learn your players before you decide on just what and how your game is going to be about. No sense in wasting all your time plotting out a campaign about moral justice and sticking to the MegaCorps if your players are going to play a bunch of Amoral assholes that think rape is a weekend pastime and will shoot their own mothers for recycling rights on her organs.
My next piece of advice is: Plan out every possible avenue ahead of time as best you can, and have multiple runs lined up and ready to go that each have their own flavor to them. If you only expect your players to make "right handed turns" every time, you are going to be lost when they turn left... so to speak. Also, by having multiple runs ready to go, you have something to pull out if your player surprise you and don't agree to your silent data theft run because they want to shoot people with very big, loud mini-guns.
Generally speaking, I have found that the best Campaigns are the ones that grow organically through the player involvement. I usually start with a small premise for the Campaign, such as "Taking out the Mayor's office" or "removing the Yakuza from the neighbourhood", then design the first few runs to be rather open ended. usually touching of the true nature of the campaign by about the third or fourth run. However, what usually happens is the players get engrossed into the world, and follow their own path based off of their character's motivations and outlooks, which takes the Campaign down paths I never expected, and sometimes totally changing the campaign from what I originally envisioned!
Being able to adapt to what your players are directly and indirectly telling you throughout your campaign will be what makes are breaks your game, so having as many options planned out ahead of time can help you keep ahead of your players' expectations and make for successful campaign.
