Does “hyper-personalization” sometimes (often?) deliver poorer service?
Recently I was reviewing updates to a social media application and a childhood lesson in customer service came back from my subconscious. This service had changed the function of the most commonly used area of its user interface; I assume with the objective of providing a better experience to its consumers. The change resulted in a firestorm of comments and negative user feedback, much of which was posted to that very area of the service. Though most of the feedback was directed toward consumers’ difficulties reproducing the site’s previous behavior it occurred to me that there was another, very significant issue to consider.
The end of the vicarious experience
Essentially, the user interface had changed from a chronological display of all events and updates to a more personalized display. The service would select items it “believed” were likely to be of more interest to the consumer. (There was also much feedback related to those items not being of much interest, though I will not address that here.) While all other items followed, they were no longer in chronological order. You might be thinking “if all of the information is still there, what’s the issue?” That is a question I could not answer until I finally realized what was responsible for the strange feeling I had toward the change and had, until then, been unable to articulate. The vicarious experience was gone.
It then occurred to me that one of the most useful and powerful benefits of the service had been its ability to generate a vicarious experience. That timeline of events had been delivering powerful emotional experiences. It made it simple for its consumer to share the emotions of those posting the events, and to experience these events with the friends and family who had posted them. What was left was a random collection of events. These events were still interesting, though the emotion they invoked was much less powerful in many (most) cases.
Don’t be too smart
As I was using and reviewing this service, as well as some other new applications and updates to others, I began to think of a lesson that I learned as a child. On occasion, I would approach a simple problem in a work setting with what I believed to be a very “elegant” approach. At times (usually) these solutions would require more effort, which I did not mind at all. However, the lesson I often learned was that the “elegant” solutions often did not deliver any more value to my customers or employer than would have a much simpler solution. In fact, these solutions were sometimes less valuable. Furthermore, sometimes the additional effort required would put the timelines at risk.
I will admit those were not the words I would have used in my pre-teens, though I believe they accurately describe the situation. It was at those times where my family would impart some Steel Town wisdom on me: “Sometimes it’s best not to be ‘too smart’”. That is, it is possible to over-think and over-deliver.
Are we getting too personal?
So, a question that I now consider frequently is “are we (as an industry) getting too personal?” Are we getting “too smart”? That is, are we sometimes providing poorer service to our customers by assuming we know better what they want than they do? Ironically, it’s not as simple a question as one might think. With the enormous quantity of data and choice available to consumers it is certainly reasonable to want to offer some “guidance” through personalization, and to want protect consumers from information and services that could harm them. Having stated that, when we personalize we are subject to the possibility of insulating consumers from something that would be of tremendous value to them. Eli Pariser’s TED talk “Beware online ‘filter bubbles’” provides a fantastic explanation of this. (It’s nine minutes very well spent.)
Where do we go from here?
So, I believe that it clearly is possible to have too much personalization, and when that happens significant value and opportunity can be lost. We must certainly be cautious not to over-personalize our applications and services, or at least to make our consumers aware when we have done so; and perhaps offer them the option to opt out in whole or in part. It is also possible to have too little personalization.
Thus, the answer is simple. Perhaps not. What advice would you give to those like Dr. Maria Velez-Rojas who are creating visualization strategies, applications and services with regard to this dilemma?