This is a very valid question, since the semantic Web world will IMO mark the end of the “destination” Web and the advent of the “me” Web: I don’t need to go to any Web site to get my information, it gets instead aggregated and filtered out of the junk I don’t want to see in the format I enjoy to consume it the most, one of the filters being the reputation I have of the author of the content as weighted by each member of my social network.

I agree with Scott that good semantic marketing will start with good, and accurate data and metadata about products/services, good distribution in particular via compliance with various established publication standards (microformats, RDFa, etc.).

If I’m correct on how data relevancy will be established by leveraging your social network, SEO will essentially become the old name for “Reputation Engineering” or “Reputation Hacking”. The more reputation your company has in a given social network, the more it will top query results.

In terms of ads, since we are left with a single destination that one can see as the next-generation RSS reader, I see two models:

• Contextual ads next to the content/data my collection of queries is typically retrieving. This is similar to the ads you have in Google Reader.
• Suggestions (similar to Google Suggests) as I’m creating my query in my semantic query builder, companies pay for certain categories of products or names to show up higher in the list (imagine a Google Suggests where companies can bid to push words higher in the list of suggested items).