I therefore decided to make a campaign in which there would be a lot of powerful factions, but no one faction would be easily identified as the 'bad guys'. All of them would have their separate motivations, and I decided against planning it out at all except for the beginning. I did map out the motivations of the factions in advance.

In order to make it more difficult to read the situation in advance - to further muddy the waters, so to speak - I decided to use more than just two factions. I ended up with four (or, arguably, five) different factions, two of which were otherworldly, one had subverted one another, and two independent factions. None of which were wholly 'good', none of which were wholly 'evil'. Blue-Orange morality.

Go check out TV-Tropes for some more background on Blue-Orange morality if you want.

This duality within the runner crew itself became very clear when the four inexperienced players decided to build Awakened characters - three of which were physical adepts while the last was a hermetic mage - while the two more experienced players decided to build cybernetically augmented characters.

The two cybernetically augmented characters were a Human Face-like character (but with no bioware, oddly enough), and a Troll Street Samurai with SMG and Bo-staff.

The three adepts had different focus - one was a projectile adept (though still with a low-level Killing Hands and some unarmed combat because, you know, spirits) with a Ranger-X composite bow and Biotech skill to make a Hawkeye-style character (if Hawkeye was an elf), another was a human projectile adept with an emphasis on crossbows and grenade launchers (!) (and some skill with demolitions, because anarchism), and the third was a human, more stealth-like adept making an intrusion specialist carrying around a lot of heavy pistols (because... I have no idea. Heavy pistols are cool, sure, and I think the player was going for a Boondock Saints vibe. I still think a 'better' choice - for a certain mileage of 'better - might have been SMGs. That was difficult to predict from the start, though).

The fourth player decided to make a dwarf hermetic mage based on a combat mage. This was partially due to some misguided idea that a mage was 'needed', and partially because a mage is arguably the most powerful and flexible character in Shadowrun. I have perhaps a slight tinge of bad conscience about this character - building a mage 'properly' is a tad tricky (though arguably not so tricky as a cyber- and bio- heavy build), and it soon became apparent that I hadn't thought of - and the player didn't know about - sustaining foci in Shadowrun 3. Which is seriously crippling for a mage, at least in the start.

Standard availability and rating rules
Gear couldn't have availability more than 8, and couldn't have a rating more than 6 - like the core rules suggest. This would curb the more blatant abuses of the system (though Dikote's still availability 6) that had plagued my previously experiences with Shadowrun 2 (and the abuse of ignoring availability while creating characters).

Character creation used BECKS and 450 Karma
BECKS allows the use of much more finely grained character creation as opposed to the 'Five priorities' style of character generation. I also decided to use the 'Staggered Karma Pool' option from Shadowrun Companion, where every 10 Karma gave the eleventh in Karma Pool up to a Karma total of 30 (33), with every 20 Karma after that doing the same. This was to allow the crew to build a Team Karma Pool (and a sense of cohesion) early, because I knew that the campaign would be divisive enough as it was. (More about this in the Campaign Bible)

No Riggers, No Deckers
Quite simply because rigger characters and decker characters aren't really playing the same game as other characters are. In order to again try to foster team cohesion, I didn't want dedicated riggers or deckers. That wasn't the same as having characters without driving skills or computer skills, but the character couldn't have rigging or decking as their primary focus.

Mission Logging
After the first couple of game nights, I decided to retroactively write a mission log detailing the game in a play-by-play fashion. I wasn't sure why I decided to do that, but after a while it became clear that the homogenizing effect on the narrative made that the players had a consensus 'reality' to follow, as the 'official' narrative in the mission log became the 'objective' truth, though it might differ slightly in emphasis and actions as compared to the actual game night. So, essentially, I wrote canon after each game night or couple of game nights.

This wasn't the same as saying that I, as a gamemaster, had full fiat on the narrative - I did get corrections and request for extra detail on some points, but in general if I didn't get any feedback, what I had written was the canon. This canon would have an excellent effect on group cohesion, but has been quite the laborious process.

I also decided to make a campaign sourcebook, and I decided to write mission logs for separate characters that the other players wouldn't receive in order to keep some sense of mystery between the characters themselves.