At the core of advertising is the "idea" - almost similar to a startup. Neither number crunching nor some psychological tricks are able to replace this. In fact their purpose is in my opinion just to evaluate or optimize towards better follow up investments.
Now this might be a different story when talking about (direct)sales where you can try to quantify action-reactions related to a specific product. And for most businesses the role of marketing ("all processes for creating, communicating, delivering, and exchanging value") is far more complex than just Tweets, Facebook and Foursquare.
Perhaps, but that doesn't negate the strategic fundamentals of either marketing, advertising, or sales. The emergence of breakthrough tactics -- even disruptive tactics -- does not necessarily upend strategy. In some cases, it simply allows the marketer (or advertiser) to pursue his/her strategy more effectively.
"The product sells itself" is, to some extent, a misperception. Even in the coming state of things. No product ever truly sells itself -- even if it is found in search, or recommended by a friend, or stumbled upon. The art and science of building, distributing, pricing, and positioning that product -- so that is is recommended by the right people to the right people, stumbled upon under the right circumstances, or searched for in the right context -- still draws upon all the knowledge built up in the marketing field to date.
Psychology is as relevant to digital advertising as it ever was in analog advertising, and it'll be massively more important in social advertising. Understanding how the human mind works, why it chooses what it chooses, and how people are influenced and, in turn, influence others -- these things are still the bedrock of advertising. Media types may rise and fall, but the human brain is still the human brain. Group behavior is still group behavior. Psychology and sociology will be increasingly important tools in the marketer's arsenal as we move into the media of the future.
It's not an either/or proposition between psychological theory and real-time data analysis. Savvy marketers should combine the two. Both have their place. (As you've pointed out, the piece tends toward too many "either/or" implications that needn't be).