Security camera hanging off building.
PHOTO: Maksim Shmeljov

The first time I heard the phrase "creep factor" was while doing some consulting for a large pharmaceutical company around their personalization project. They were struggling with just how much of the customer data that they held that they could expose back to a customer during a transaction without it feeling that they knew too much.

And they did know an awful lot about their customers, not just the doctors and medical facilities they dealt with, but also the individuals who had been prescribed one or more of their products. The conundrum they faced, and it’s a common one for any company that captures information about the customer, was how could they use that information to deliver a personalized experience without it coming across as creepy? When did personalization become too much?

Know Your Audience

The first rule of creating any sort of customer experience, from marketing, sales, all the way through to support, is to “know your audience.” We need to be able to empathize with our audience, to understand what it is they are trying to achieve when they interact with us.

What problems do they have that we can solve for them? What can we do to make their lives easier, to make them more successful? Companies that understand their audiences have greater success and more repeat business. And to do that we need to collect information about who the audience is; otherwise, we are just acting on assumptions.

The problem is, what do we do with all that data, especially if it is data that allows us to build a picture of an individual’s behavior? Is it Personally Identifiable Information (PII)? There are rules and legislation in place, as well as good business practices to follow around the handling of such data; but the question remains of just how do we use it in communicating with our customers?

Related Article: 3 Tales From the Creepy Personalization Vault, and 1 Important Lesson

Creating Barriers to Engagement

Working with the pharma company in question we determined that their customers had a tolerance for voluntarily sharing four pieces of data when they interacted with the company, and around the same for when the company communicated to them. It turned out that engagement was better if those four pieces of data were spread over a series of communications. Showing that the company knew four things about the customer in a single email did in-fact cross that "creep factor" threshold.

We also discovered that the tone of any personalized communication also mattered when determining customer reactions. Overfamiliarity can instantly cause a negative reaction.

Interestingly this came up in conversation when my wife and I were out for dinner with several friends the other week. And in my very unscientific round-the-table survey, everyone agreed that if they receive an email that uses their first name in an over-friendly, overfamiliar manner from someone they have never interacted with before, then it will be immediately deleted. Overfamiliarity in personalization can be viewed as not only creepy, but as pushy and creates a barrier to engagement.

The Illusion of Personalization

Perhaps the answer to using customer data is not personalization on the individual level, but the illusion of personalization. At a recent conference I heard the following quote from the IDC analyst Marci Maddox, and I think it makes a very valid point:

“Personalization is not about putting names on 10,000 pieces of content; it’s about delivering the right valued content to 10 people.”

Wouldn’t it be better to be able to deliver the right content and experiences to a group of people with similar interests and problems to solve? It’s an approach that shows empathy without being creepy, but still makes the customer feel that we do know them and their problems, interest, desires and goals. 

Instead of trying to manage the "creep factor" while trying to reach the mythical 1:1 personalization goal, maybe we should focus on the tasks at hand? Just because we can collect data about individuals, maybe we need to consider if we should be doing it.