The math forces VCs to work through referrals: any individual VC can fund only a few companies a year, meets with 15–20 for each one funded, and has hundreds more who want meetings. She has to rely on her network to screen the hundreds down to the 15–20.

Why cold outreach almost never works.

Submitting a business plan 'over the transom' to a venture firm is likely to amount to just as much as submitting a screenplay over the transom to a Hollywood talent agency—that is, precisely nothing.

Don't bother with a long detailed written business plan. Most VCs will either fund a startup based on a fleshed out PowerPoint of about 20 slides, or they won't fund it at all. Corollary: any VC who requires a long detailed written business plan is probably not the right VC to be working with.

The best thing you can walk in with is a working product. Failing that, a beta, prototype, or mockup—and ideally some traction. Rule of thumb: when in doubt, work on the product.

On how to make a strong first impression before you even need contacts.

Qualify, qualify, qualify. It is completely counterproductive to pitch a health care VC on a consumer Internet startup, or vice versa. Individual VCs are usually quite focused, and screening out the wrong ones is absolutely key.

If you're a programmer, create or contribute to a meaningful open source project. Emailing a VC saying 'I'm the creator of open source program X with 50,000 users worldwide, and I want to tell you about my new startup' is far more effective than a normal cold pitch.

When a VC is exploring a new communication medium—Facebook, Twitter, blog comments, audio pitches—she's often more open to interacting with strangers through it than she otherwise would be. Jump on new mediums early and engage there.

Andreessen notes Fred Wilson once solicited audio pitches entrepreneurs could upload for him to play on his iPod.

Ranked alternatives if you can't get to VCs: angel funding first (good angels know good VCs and will introduce you), then bootstrapping via consulting or early customers, then keeping your day job and moonlighting, and finally credit card debt. Only angel funding is actively encouraged.