[diegoeche]

It seems is all about "context". A user might like an article about current election. But what does that "Like" mean? Does he actually like the article? Or maybe he liked it because it's clearly ridiculous (almost a joke in his opinion?).

Knowing that a user like let's say "Book X" but no context of "What he liked about Book X" sounds less powerful. But at the same time: Easy to quantify.

The "Having a multi-dimensional representation of people's tastes can give a ton of information" needs to be leveraged against the noise the insignificant contexts of people's tastes introduce. That IMO is half what the pitch should be about.

How to quantify reliably these "contexts" of people tastes. How to exploit them in advertisements. And how this approach is better than let's say, FB's powerful ad creation tool based on "the graph" is only part of it. You'd also need to tell, how are you planning to get your critical mass. Your approach needs to be: novel enough, pretty enough, useful enough, and viral enough, for users to get attracted to start using your tool.

[fbuilesv]

I also like the idea of exploiting the multiple levels of people's tastes as the selling point of this so I will be basing my explanation on it. Thanks for the feedback.