Recently, we told you that half of digital marketers believe the ability to personalize web content is fundamental to their online strategy. This Econsultancy report highlighted four steps to perfect personalization: seek consent, gain control, earn trust and deliver incentives to the targeted audience.
According to a white paper by Baynote, these steps speak to the psychology of needs, which drive all of our behaviors, including the relationships we cultivate and the objects we consume.
I Want You to Want Me
The report, "The Human Need for Personalization: Psychology, Technology and Science," challenges the idea that a personalized web experience must deliver unique content to each user. While the goal of personalization is to provide customers with interactive experiences tailored to optimally satisfy their individual needs, Baynote posits that,
Personalization is much like a matchmaking exercise with the ultimate goal of pinpointing the best product or service from your catalog that best satisfies your customers’ needs."
It turns out that we're not as unique as we think we are. Because our interests can be similar to others, it's important to look at community behaviors to deliver relevant content. As such, the strategy for meeting customer needs will look very different depending on the customer’s intent. Baynote introduces a framework that can help us understand the different approaches we can use.

A personalization system not only needs to understand the users’ needs, but also which needs are best filled by which products.

Outputs can consist of the look and feel or organization of a page and typically occur within a particular context or personalization zone, for instance on the homepage or within an email client.
By themselves, inputs and outputs don’t mean much. To make them work to our audiences’ advantage (and our business outcomes) Baynote says that both art and science are required. Neither manual processes nor machine learning are enough by themselves. To help us better understand, Baynote provides this description:
Inputs are the raw data, and the features are how you combine or organize that data. For example, a user’s birthday is an input. However, if you give that to the machine learning algorithm directly, it still has to figure out if the day of the week matters most, and if people with similar birthdays act similarly, etc."
Don't Predict My Behavior, Influence It
Additionally, we often associate the personalization of experience with predicting behaviors. Instead it’s about influence. Rather than predicting what a user’s next action is going to be, it’s more important that a machine be able to predict what a user’s response would be to every possible action it could take, and then to choose the action that changes user behavior that meets both end user and provider.
Learning Opportunities
Much like community management, web personalization is about validating a consumer’s need. We all like to be heard and we take heed in those who get to know usso they can better serve us. Baynote says that user experiences with websites are social interactions. As such, the goals of building relationships and engaging users are the same.
Of course, all of this can be overwhelming. The white paper is definitely cumbersome in its explanation, but worth a read, nonetheless. Yet, to gain a better understanding of what this all means, Baynote also produced an infographic that helps to explain visually what the psychology of personalization means for marketers. The infographic, "I Know What You Did on the Web," details what information popular website are gathering about you and what they do with it all.
