Meaningful Encounters of the Second Kind

Although technology still can't mimic complex human relationships, understanding the way your customers think may be the next step in personalization

By Barry Grushkin

Dr. Solomon was an amazing professor. His enthusiasm was contagious. With just the right image, concept, metaphor, or diagram, he made the difficult crystal clear. His questions would nudge the class members toward making amazing realizations on their own. You could utter a few words, and he knew just what you were thinking. Once a student just said, "Hmm," and he said, "No, you are on the wrong track. Check this reference," and he was right!

Clairvoyance? No, the utmost in personalization.

Getting Beyond Semantic Tunnel Vision

Dr. Solomon was constantly attuned to how his students' understandings were evolving. "Personalization" is just one of a number of important terms for which, at our peril, we too easily lose sight of its depth and potential. Others include intelligence, relationship, metaphor, concept, learning, and information. However, if you return to the rich implications of these terms, new technological directions open up. But you must start asking the right questions to move ahead.

Intelligence is the intertwining of many systems and forms of computation far too complex for current methods of understanding. Artificial intelligence (AI) is a grab bag of tools that offer a little bit of problem-specific smarts.

Relationships are between people. Semioticians talk of the phatic channel - the feeling that just being in a person's presence has value. How can a machine reproduce that? Can any machine replace the trust developed from a long-term relationship? I studied because of my relationship with Dr. Solomon. What machine could replace that?

Metaphor: Early AI gurus thought their machines could understand metaphors, but transferring ideas in a parallelism, keeping some key structure of an image, and mapping that into another image or situation - so as to enlighten or color it - has, so far, proved impossible for machines.

Images help people learn and think. No one has discovered how machines might do the same. People develop many associations with pictures: Power relates to size, a smile is uplifting, or landmarks recall directions to visit an old friend.

Computer attempts have relied on turning pictures into a set of logical relationships, as if you could document the many thousands, or more, words a picture is worth - or that the words would even be the same for each viewer.

Concept? No, this term is not a set, a category, or a clustering of texts. This is a socially active, growing, interactive polyvalent nebula of meaning with subtleties that can take people a while to get, and from which human actions, large and small, arise. For example, if you try to diagram the wide-ranging uses of a term, even one as simple as "there," you get a very complex map with differing parts emphasized in differing encounters, with the whole diagram evolving over time.

Learning is not the ingesting of bytes. Words, for example, can conceal as much as they reveal. You also diminish the term to say that neural net algorithms, which merely find forecastable boundary conditions, learn.

Learning is gaining a living experience with material - an apprenticeship with the ideas - so that you can know how to move with them and grow. Information for people is not measured from a formula enumerating patterns in bytes. Information deals with the effect and value within an existing, elaborated context.

In which of the following would you say I am giving out more information: a joke, the inside scoop on a stock, the missing formula for how to build a smaller A-bomb, or something that would change your view of the world forever? Just a few words can offer a key insight that unlocks or draws together complex details.

Personalization? Grouping people with a few million others is a far cry from a person who knows you well and can lead you toward discovery. Dr. Solomon's example hints at a next step for data mining and personalization.

Mining With a Wider Vision

Psychology uses the term "introject" to refer to the internal models that humans have of others. Big relationship blunders occur when internal representations don't square with reality. Feedback is always necessary to update and correct these models. Similarly, when CRM models are off, do customers have a good laugh as they toss aside your misdirected marketing?