I can't speak to your situation, per se. I can only tell you that in my experience (managing and marketing in both the for-profit and NPO sectors), comprehensive demographic information was very valuable. Since I don't know the agenda of LessWrong in detail, I can't say if, for instance, knowing the income distribution and the markers for charitable giving amongst LessWrongers would be of use. These typesof data help you to define the kinds of projects you can reasonably hope to fund, and thus reasonably hope to market to your demographic. Markers for giving were very reliable in my experience - today, given the economy, I don't know.

Beyond money, a well constructed survey will almost invariably reveal all kinds of insights, not just about your members, customers or readers, but about your own operation - how it is perceived, what people like but aren't getting, and sometimes, insights into your own psychology and approach that you didn't previously have. The key words here are "well designed," because it is surprisingly hard to do a comprehensive survey and get most of the questions you that you want answered, answered. And to avoid bias in the way the questions are phrased, or even in the order in which they appear in your survey. While I can't prove it, I think it likely that intelligent use of survey information, gathered in the 1980s, was in part responsible for the brief period when Alcor membership growth was nearly exponential. While such data will never do that for you absent many other things being done "right," they can, IMHO, amplify the effect of good management and marketing.