The social life of humans and machines: How to design for the social interaction

Solitary designers, lonely geniuses, and isolated creative teams: it’s time for you to think about social interaction for your user interfaces. No, I am not talking about social software, user-created content, or Web 2.0. I am talking about any old-fashion user interface: because for us human beings any interaction is a social interaction–even when we are interacting with “a stupid machine.”

We all know this: we tend to treat objects that display intelligent behavior as humans. We curse at our computers and at frustrating websites, we trust or distrust them, we talk to them, we find them cute or insufferable; we get angry, happy, frustrated, outraged, even if there is no living creature on the other side. (Well, there is always a living creature on the other side: somebody has designed and built the application. And human qualities bleed into inanimate objects very easily.)

Interfaces can persuade, make us uncomfortable, happy, or unhappy. The personality of an interface emerges by the interactions with humans: the style of the content, the choice of words, the behavior, how well it responds to our expectations and needs, and the appearance blend together to trigger our all too human feelings. Interactions between humans and machines are conversations (or arguments) in which information is exchanged, connections are made, needs are satisfied or frustrated, and relationships are built. When we design applications we need to think of them as social entities.

Do you remember the first and roughest attempts to add human qualities on the interface? They were literal translations of human characteristics. We put avatars, faces, voices, and personalities in our applications. We failed. Literal translations are usually the first to be tried and they almost always fail.

Literal translations fail because they don’t understand the experience they are trying to replicate; they are lazy copy-and-paste exercises that only capture the most superficial characteristics of the experience. Literal translations don’t account for the unique opportunities and limitations of the medium we are working with.

We don’t need an interface to have a voice, a face, or a personality. Interfaces already have them by default. Voices, faces, and personalities are create in our minds by the interaction with the interface, in the same way we create images in our mind when we read a book. We tend to anthropomorphize whatever we interact with because interactions with other human beings are the model for all our interactions and are complex and usually effective models.

Translating a non-human interaction to a human interaction allows us to categorize it, understand it, and react to it. One of my computers, for example, is vain and temperamental. The other one is practical and good natured. The computer I use at work is stubborn and not very smart. None of them has a face, a voice, or an added personality. It’s my day-to-day interaction with them that assigns human attributes to machines and shapes my own behavior and attitude towards them.

Building socially-aware interfaces requires us to understand the important characteristics of the interaction and to filter out superfluous and annoying features. It requires us to understand opportunities and limitations that our technology offers to model social interactions. When we test our interfaces with people we need to pay attention to how they are perceived and why. Is our application paternalistic, arrogant, or too intellectual? Or is it supportive, friendly, and caring? Does it satisfy the social needs of that particular task?

Recently, I worked on an application that provided purchase suggestions for products. We realized that we needed to do some work to convince people who our recommendations were in fact appropriate for them. The people we interviewed described the process of recommending a product as a conversation. They wanted to answer some questions about themselves, receive suggestions appropriate to their situation, understand how the suggestions mapped to their situation, have options, ask questions, and be able to negotiate the results. These were the qualities of that particular social interaction that we needed to capture.

Designing an interface for a successful social interaction is not about adding faces, voices, and injected personalities. It’s not about using chat, videoconferencing, or other fancy technological tools. It’s about being able to recreate the social experience that satisfies people’s need to be listened to, guided, persuaded, cheered, and reassured. When we started paying attention to the qualities of the social experience, a miracle happened: people started to love our design.

Tags: design, social interaction, software, persuasion, web.