Beyond Persona-based UX Design

For quite a while now, UX designers have been referencing and relying heavily on user personas to design software.

A persona is usually defined as one would define a character in a book, then placed in various scenarios and given various tasks to complete using the software. In this way, we try to envision our users at work and better understand their intentions and goals. Then we begin designing and create concepts within the context of our software infrastructure and user interface to offer our users tools, support and capabilities for what they want and need to do.

But a user persona is ultimately just an abstraction, a generalization based on a collection of users that in real-life fits no one user perfectly or exactly. As designers, we use these personas as a way to help us get to know our users better and to consider and anticipate what they may be trying to do with our software. Successful UX design is often measured by how well it supports the needs and goals of our defined personas.

The personas we use aren’t completely fictional. Ideally, they’re based on real users who are shadowed in the workplace to discover typical job roles and work patterns within an industry. A process is used to take this information, analyze it and form it into categories of needs and goals to develop a user profile to embody within prototypical personas. We know that in the real world of work that people assume many roles and have blended responsibilities in their jobs that cross over into related (and sometimes unrelated) areas. So no one persona or collection of personas will fit any person perfectly. Persona ambiguity and over-generalization is something that designers have accepted as an intrinsic part of software design. And we have worked with this shortcoming for quite some time now.

Now imagine that your software is smart enough to engage with you in an adaptive, interactive way to give you a much richer user experience. Imagine that it can really learn about you, your roles, your traits and tendencies at work. Imagine that it can remember what you did yesterday and last week at work, and can offer you suggestions and tips on how to optimize everyday or challenging tasks to save you time and trouble. Augmented intelligence has come a long way in recent years. We’re just about to turn the corner into a world of much smarter software for everyday users within the enterprise. And the more you use the software, the faster it will get to know you and better understand what you're trying to do. It may even be able to anticipate with impressive accuracy what you are likely to do next. And all of this happening while conversing with the software in natural language, with the opportunity to train it, teach it, and correct it, if you want. 

For decades now, users have had to learn how to use software. Designers work hard to clarify task flows and user journeys and to make successful task completion obvious, clear and intuitive in software. And yet we know that what is intuitive for one user may not be at all intuitive for another user. A popular designer fallback is to make features in software discoverable. This means the user can click on something in the software application and discover its capabilities through use and observation. And once discovered, it's then known and remembered for the next time.

Perhaps very soon, software will be smart enough to learn the user. Using a conversational user interface with augmented intelligence to identify semantic patterns would be a great start. A smart entity can query a person about their work and how they best like to do it. It can learn from the user the nuances and specific idiosyncracies of modern job responsibilities. Chances are good we will move beyond persona-based design into an entirely new way of designing software for users when we are able to learn from our users in an intelligent way that parses, categorizes and summarizes unique personality characteristics, skills, interests and responsibilities that are most suitable and effective for a particular position into discernible and predictable patterns that can be used as a basis for software design.

But the question of how we design for learnable, adaptive software for our users is still open and needs to be answered. What will this mean to UX designers? How will this change how we design software? UX researchers, what do you think? If you were to represent yourself as an intelligent entity within a conversational interaction with a user, what would that look like? We really do need to figure this out.

Because moving beyond personas could change everything.