Nice article today from Jeanne Jennings at ClickZ regarding email best practices and testing. She comments on the use of click-stream data and other analytics to personalize and customize email messages.
She talks about the fine line between the importance and effectiveness of personalization and message relevance and 'hyper-personalization', or a 'big brother-ish' feeling one gets from knowing a company may have too much information about the recipient, where they surfed on a website and how that can be a turn-off. Couldn't agree more.
I will say however that striking the right balance between personalization and intrusiveness is a very important element to test and get right. How much likely are you to respond to an email with your name in the subject line, or with an offer or content from someone or something that you know you recently looked at? Can have a huge affect on conversions.
On the topic -- I've recently received an email from a company (that I've opted in to) that nevertheless went directly into my spam folder in Yahoo. Couple things jumped out at me -- it's important to make certain you have some copy at the top of your emails to ask the recipient to either include your address in their address book/list of accepted email senders or to provide them the option to opt out of future messages (in fact a clear and prompt opt-out procedure is a CAN SPAM requirement).
A quick review of the email also pointed out some potential spam folder issues - it had 'FREE' in the subject line and had quite a few images in the email body. These elements might not automatically doom you, but they don't help.
Many ESP's (definition) have built in spam filters that allow you to test and score your emails to see how likely they are to make it through spam filters and become deliverable into in-boxes. If yours doesn't there are many free spam filters available (SpamAssassin and SpamCheck to name but two) for you to use, I heartily recommend.