* Avoid personal drama. I suspect that in more cases than we realize, management-level changes happen simply because people stop working productively with each other. In particular, you need to be prepared to DISAGREE AND COMMIT; to be able to signal that even in cases where you disagree about the direction of a decision, you aren't plotting behind the scenes for its failure.
* Stay close to the money. Most of you are tech people. As a tech person in a moving company, you can basically be doing three things and still be perceived as someone who is executing: you can be shipping product improvements, you can be doing "marketing engineering" (ask Patrick what this means, but, briefly: "things that improve customer acquisition or customer LTV"), or you can be talking to customers/the market. There seems to be a huge trap for techs in startups in the "CTO" and "Chief Scientist" and "Architect" roles; AVOID THEM.
* Be strategic about roles and hiring. This stuff about palling around with the board may be important, but from what I can tell, the game is won or lost on the org chart. Don't hire people who are going to ladder-climb around you. This is one aspect of Startup Soap Opera Drama that is not overhyped: the market is lousy with people who, for good reasons or bad, have a primary goal of being one of the key people on the m-team. A simple trap to avoid: if you're the Dir/E, be very careful about who the VP/E is; if she can't be you, aim for your kid's godmother. Similarly, if you're in Product Management, even as VP/PM, be very careful about VP/Marketing.
* Be careful about demotion. Clarifying special case of the previous note: in a lot of biz cultures, there's basically no such thing as a "demotion"; usually there's only termination, lateral moves, or "constructive" demotion (hiring SVP/M to oversee and eventually consume raw the VP/M, then EVP/M for the SVP/M, etc). If you allow yourself to be demoted, and your m-team is a bunch of assholes (not an infrequent occurrance), you can be perceived as weak.
* Keep zeroes out of the m-team. Human resources doesn't belong in the m-team. The Dir/M in a single-product company where a VP/PM handles 99% of marketing doesn't need a seat at the table just to represent "Marketing". The less vital a role is, the more likely they are to be a magnet for bullshit politics.
* Be customer facing. I think I said this already: if you aren't committing code that the company will sink without, you should be meeting customers or partners constantly. Be very careful about getting sucked into the conference circuit. It's easy to convince yourself that a particular conference is important to your business but be perceived by everyone else in your company as a tourist.
* Be on the same page with your board. In two of the cases I've been directly involved in, the conflict underlying a planned coup was at the level of "are we going to grow the business or are we going to position the company to get bought". It seems like you need to have an eye for the kinds of decisions that involve liquidity. Early on, your board may be worried that you're tilting the game for an early exit; later on, especially if you aren't on the original founding team, you may be perceived as an obstacle to bizdev when your VC just wants to get rid of a board seat gracefully. You don't have to agree on every decision, but it seems like you really do have to be perceived as having aligned interests.
* Be extremely cautious with metrics. Anyone who's ever managed a sales team knows about "sandbagging", where the guy running a region lowballs the numbers so he can sail over them and collect bonus accelerators. A lot of devs have a natural habit of being optimistic --- about schedule, about product adoption, about support costs, about COGS --- all of which gives the management team ammunition down the road.