It's amazing that everyone here has answers, but no data. If you're going to give an answer, back it up. User-readable? Yes. Documented by Google. No copy? Links only? Works for some sites like CNN, ToysRUs, Walmart that get picked up just because they're huge (observation). But for the majority of the little guys, content plays a role and it would be great to know if the data supports keyword density as still being applicable to G. Tools still measure it (SEOQuake). In natural language, it seems to make sense that a certain percentage of words, on average, are repeated. Google has made it clear that they are trying to master how language is actually used in the real world and providing results based on how humans communicate, not computers. Thus, more people focus, less computer focus. YET, we all know that computers still play a huge role in how SERPs choose winners. We just have to find the balance, right?
