As every programmer knows, God was able to make the world in only six days because he didn’t have an installed base. Programmers rarely have that luxury.

New managers have a different kind of installed base to worry about. While the difficulties they face are not as technically daunting as creating a backward-compatible operating system upgrade, the social engineering issues faced by a manager taking over an existing organization present their own set of significant challenges.

When you take over a department, whether it’s through a promotion or a job change, you don’t get the luxury of designing your operation from scratch. You’re inheriting an installed base — an existing team, well-worn processes and ways of doing things, and an entrenched culture. But where programmers usually have a test environment in which they can safely find and fix mistakes, managers have to do their testing in the production environment of an ongoing operation. Missteps are very public, and hard to unmake.

The social engineering starts before you take the job. If at all possible, find out whether you’re walking into a problem area or not. If it isn’t a problem area, try to get a mandate for change from the reporting manager to create a problem where none existed before. Failing that, let some other victim take this no-win job.

Coming into a smoothly running organization is much harder than taking over a disaster area. How are you to succeed? Your chances of further improving the situation and having the team look to you for leadership are low. If your charter is to maintain the status quo, your predecessor will get the credit if you succeed; you’ll get the blame for any deterioration.

Compare this to the desk o’ death. The department is in shambles. The team is demoralized, productivity is low, waste is high, service levels aren’t. Whenever possible, choose the desk of death, especially if you’re the third or fourth manager to get the job — expectations will be so low that your success is virtually guaranteed.

So long as you follow a few simple rules.

The first is to keep your yap shut. Beyond the usual pleasantries of how delighted you are to have the opportunity, say as little as you can. Listen to everyone, in group settings and one-on-one. Neither agree nor disagree with anything beyond broad philosophical concepts, and above all, don’t choose sides or make any commitments. Offer no ideas of your own. Listen and make note of who says what.

In a desk o’ death, everyone has a private agenda and is trying to recruit you. Assume everything you’re told is biased. You have to piece together an accurate assessment jigsaw puzzle fashion out of bits and pieces. The moment you accept any individual as a preferred or unquestioned source of information, you lose your ability to lead — your preferred source will have established his perspective as your own.

So the first rule is to take time to size up the situation. Then you can decide what needs to be changed — processes, technology, reporting relationships, team members (chances are, if it’s the desk o’ death not everyone is a great employee), attitudes, or what have you. And, you can choose your priorities.

That’s the first rule. The second will have to wait until next week.

Until then, trust nobody.