Meeting the Psychological Need

Personalized service is already ubiquitous. We don’t think of it as a separate capability, we just expect it. No major consumer retail site can be viable any longer without providing for the delivery of personalized suggestions. And because of that, it is also a business model imperative for many companies.

Amazon derives over 30 percent of its revenues from personalized recommendations. Seventy-five percent of Netflix’s rentals are driven by personalized suggestions. Google’s search results are increasingly personalized. The modern digital assistance technology in smartphones such as Siri and Cortana are not only becoming ever more intelligent, but also more personally responsive. Are these personalization technologies perfect? No. But these are the technologies for which much of the world’s expertise in data science and machine learning are devoted to, so it is inevitable that they will keep getting better and better.

Will they ever be better than people at providing customer services? That depends on the who, what and when we are talking about. For providing handy and timely information that fits with personal circumstances, automated approaches are often more efficient and effective. But if a customer is unhappy about something and wants to discuss her issue (or just vent) a real person should do the listening and interacting. The emotional/social aspect to this case calls for a person-to-person interaction, even if a computer-based system is capable of providing identical interactions. Psychologically, we want to feel we are heard and empathized with when our emotions are running high, and only real, live people can fulfill that psychological need.

The best practice seems to be for the intelligent, personalized system to handle the general and mundane, and for real live customer service people to handle the exceptional cases. That requires the machine to do the appropriate triage up-front to identify the exceptional case, and in a much more intelligent and interactive way than press “1” for this, “2” for that and so forth which we have all experienced. If the system can interact in an intelligent enough way by asking the right questions as the customer provides answers, that’s the best case.

And that may be more often the best case approach than we might otherwise think, not only because of advances in the technology, but again, based on psychology. Because just as there are psychological reasons that make interactions with live people more appropriate in emotionally-charged situations, there are other times in which automation, perhaps surprisingly, can be more effective than a human with similar capabilities.

Systems Can Alleviate Social Baggage

This counter-intuitive situation has been recently confirmed in the area of personalized learning conducted at the New Jersey Institute of Technology. People are more accepting of being corrected by a system, and therefore learn more effectively, because of the lack of the counter-productive social baggage that comes with being corrected by another person -- being embarrassed, feeling judged, etc.

Learning Opportunities

I would speculate the same thing may apply to customer service as systems become more sophisticate. Given relatively similar capabilities to people, in some situations the emotional connection that a person can provide will give them the edge, in other cases the reduction of the emotional and social overtones that come with not interacting with a human may give the automated system the edge. Finding the right balance for best customer service will require continuous tuning as technology progresses, while keeping this psychological dynamic in mind.

Another example of the personalized system serving to overcome psychological impediments is illustrated in the book, "The Learning Layer," authored by my colleague Steve Flinn. Flinn noted that in connecting people, the intelligent, personalized system (termed a “learning layer” in the book) can become a trusted intermediary that can take the “blame” by either party if something goes wrong so that neither party “loses face”:

High-quality people recommendations delivered by the learning layer also have a subtle but important psychological advantage. If a person needs some expert advice or just wants to make contact with another person because of what seem to be common interests, there can often by a bit of reluctance to reach out .... The beauty of credible recommendations of other people made by the learning layer is that it significantly reduces the hesitancy in making connections because the system can take the 'blame' if there is any misfire. 'The system suggested I contact you,' is a very powerful phrase for both of the parties of a people recommendation. For the person needing assistance, it provides cover. For the person whose assistance is recommended by the system, if he is not the best fit to address the issue, or simply is too busy to help, he will more likely be able to honestly communicate that back to the person needing help. The embarrassment factor inherent to such communications for both parties is vastly reduced. The system as an intelligent but not necessarily infallible intermediary enables a 'no harm, no foul' level of interactions. Together with the learning layer uncovering valuable connections that might not otherwise be found, this positive psychological aspect of people recommendations promotes a much healthier flow of expertise throughout the organization."

Advances in system-based personalization will inevitably continue apace. However, no matter how effective such personalization is, there will always be a place for the human touch, if for no other reason than the idiosyncrasies of human psychology. But perhaps less intuitively obvious, other elements of our psychological baggage will drive us to prefer, or at least attain better results from, automated personalization.

Title image by AdRikTa (Flickr) via a CC BY-SA 2.0 license