Reports of Persona's Death Have Been Greatly Exaggerated
User: How much wood would a woodchuck chuck if a woodchuck could chuck wood?
Siri: Forty-two cords, to be exact. Everyone knows that.
Many practitioners consider persona a dirty word in speech interface design, as well as in the larger world of visual interface design. Yet in these two worlds, the definition of the word persona differs greatly. In speech interfaces, a persona is the character represented by the speaking application: It is comprised of the vocabulary, phrasing, and tone of scripted language combined with the voice, fluency, and articulation of a professional voice talent. GUI designers have used persona to refer to the personal characteristics of a user; persona is a tool to help designers create an experience for a specific (if hypothetical or composite) person. Although both concepts have been fraught with controversy, in recent years, VUI persona has mostly met its demise, a victim of the heated debate about the humanness of machines.
I've never been a big fan of persona. While the concept does resonate with some clients, I haven't found documenting a persona to be a needed or necessary step in my own design process. I've also been troubled that persona is a red herring that has favored theoretical argument over the advancement of high-quality design practice. Here's what the psychological literature on social perception has demonstrated:
Speech and language are social cues that provide information about a speaker.
Automatic cognitive processes (or those not within our willful control) cause us to perceive and make inferences about personality in the presence of social cues (within as little as 250 milliseconds of receiving a cue).
Studies of spontaneous trait inference have demonstrated that even inanimate objects are perceived to have humanlike qualities in the presence of social cues.
A wealth of communicative-psychological research indicates that we understand enough about human perception that there is no question about whether speech and language precipitate personality perception. They do.
With that issue out of the way, the designers' job is to control the language (script) and speech (voice talent) of an application, meaning the designer's choices about scripting, talent selection, and vocal coaching are responsible for users' impressions of a system. However, in speech applications used for customer service, satisfaction is the result of user impressions compared against a narrowly defined set of service provider expectations. My own research has revealed that usable, satisfying IVRs must focus on user needs (efficiency), helpfulness, friendliness, and use of succinct, everyday language in communicating messages. The persona of most common speech applications is highly constrained because it must adhere to customer service expectations.
Enter Siri. As speech technology moves further into the realm of mobile devices, designers may be on the cusp of a new creative renaissance. Mobile device assistants provide new opportunities for speech and language usage, free of the rigid behavioral standards that user expectations impose on IVR design.
These differences open up voice design to a broader range of vocabulary selection and phrasing than is usually preferred in IVR self-service. In addition, humor, idioms, metaphorical and inferential language, and a host of speech presentation styles may be not only appropriate but critical to user delight with mobile interactions. I'm not suggesting designers abandon efficiency and conciseness in mobile speech design. However, unlike IVRs, where interaction efficiency alone can predict user satisfaction, it seems reasonable that mobile assistants are allowed to be more creative, talkative, unique, and, yes, fun.
Of course, user-centered design practices remain paramount for emerging technologies, but good mobile design also means we have space to play. After all, who wouldn't be delighted with a mobile assistant that knows the meaning of life or the unladen flight speed of a swallow?
Melanie Polkosky, Ph.D., is a social-cognitive psychologist and speech language pathologist who has researched and designed speech, graphics, and multimedia user experiences for almost 15 years. She is currently a human factors psychologist and senior consultant at IBM. She can be reached at polkosky@comcast.net.
Related Articles
When it comes to IVR design, are you seeing the forest for the trees?
05 May 2014
A window into the future of speech technology.
10 Sep 2012