Uses and Abuses of Personas
Created by Catherine Howell (University of Cambridge) on March 12, 2007
I've been following the debate on the Sakai Pedagogy list, about personas and their shortcomings.
For the benefit of those unfamiliar with this term, 'personas' are generic user profiles, similar in many respects to the consumer profiles used by marketing organisations. They are employed as a tool for systems analysis, the aim being to design and build more usable systems, by understanding the needs and intentions of the people who will use them. The process of creating a set of personas normally involves an iterative process of research / evaluation, whereby individuals' unique 'needs and intentions' are grouped into normative sets.
The problem is, they don't work. Their chief benefit is also their greatest shortcoming: personas are inherently generic. They are not tools for personalisation.
The other issue is that personas encourage systems analysts / software designers to build systems around institutional roles, instead of activities. Why!? Organisations change. Organisational and institutional roles have a tendency to mutate, shift, and/or vanish, and people may change roles within an institution - once, or several times. Changing a customised system once it's built is expensive and time-consuming.
I believe systems should be built around tasks and activities. People don't visit a website or fire up a program in order to perform a role; they want to complete a task. Here's a 5-second test: try looking at the website of a local university. How many institutional websites divide information and hide it away according to roles? ('For Staff'; 'For Students'; 'For Alumni'...) Yet the tasks and activities that each of these groups will need to perform are, in many cases, more alike than they are unalike. This way of structuring information inevitably leads to gross inefficiencies and significant duplication of information across various 'gateway' sites.
Personas can be a useful way to think about the different types of people working within an organisation, and they can be a useful way to group tool / systems users. But I believe their days are numbered, because the type of organisation (and organisational culture) they are designed to support is on the wane. The use of personas ultimately reflects an approach to systems design that is outdated. Organisations are gradually becoming looser and less structured, and their members want the ability to retain control of the tasks they have to perform. They want to have their information delivered in a way that is meaningful to them.
For the benefit of those unfamiliar with this term, 'personas' are generic user profiles, similar in many respects to the consumer profiles used by marketing organisations. They are employed as a tool for systems analysis, the aim being to design and build more usable systems, by understanding the needs and intentions of the people who will use them. The process of creating a set of personas normally involves an iterative process of research / evaluation, whereby individuals' unique 'needs and intentions' are grouped into normative sets.
The problem is, they don't work. Their chief benefit is also their greatest shortcoming: personas are inherently generic. They are not tools for personalisation.
The other issue is that personas encourage systems analysts / software designers to build systems around institutional roles, instead of activities. Why!? Organisations change. Organisational and institutional roles have a tendency to mutate, shift, and/or vanish, and people may change roles within an institution - once, or several times. Changing a customised system once it's built is expensive and time-consuming.
I believe systems should be built around tasks and activities. People don't visit a website or fire up a program in order to perform a role; they want to complete a task. Here's a 5-second test: try looking at the website of a local university. How many institutional websites divide information and hide it away according to roles? ('For Staff'; 'For Students'; 'For Alumni'...) Yet the tasks and activities that each of these groups will need to perform are, in many cases, more alike than they are unalike. This way of structuring information inevitably leads to gross inefficiencies and significant duplication of information across various 'gateway' sites.
Personas can be a useful way to think about the different types of people working within an organisation, and they can be a useful way to group tool / systems users. But I believe their days are numbered, because the type of organisation (and organisational culture) they are designed to support is on the wane. The use of personas ultimately reflects an approach to systems design that is outdated. Organisations are gradually becoming looser and less structured, and their members want the ability to retain control of the tasks they have to perform. They want to have their information delivered in a way that is meaningful to them.
Bookmark/Search this page with:
Identities are important, and complex. Speaking for myself, I would wish to separate my 'identity' from my institutional membership and/or role. I don't use institutional systems in order to reinforce my sense of self. :-)
Even if I restrict myself to thinking about the design of University web portals, dividing up the goals and intentions of site visitors according to role just doesn't make sense to me. Some sample goals: check my webmail; log in to the VLE; look up a phone number; check lab health and safety rules; report a student who's not turning up to class; contact a student society; renew a library book. I could be 'staff' or 'student'.
I disagree with your assertion that this way of structuring information inevitably leads to gross inefficiencies and significant duplication of information. I don't think it's inevitable. In many circumstances it is likely but inevitable is much too strong a word.
Roles are useful as a way of displaying and even categorizing tasks and information. I would argue that roles are closely related to the concept of identity and there are few more powerful forces in one's life than discovering, shaping, and preserving one's sense of identity (a simplification, of course, as we never have just one simple identity - it's always a number of identities or a complex amalgam of identities). Think of it this way: When someone is asked, "What do you do?", the answer is usually "I am a ___." Although the question explicitly asks about "tasks" the response is one of "identity" - I *am* something.
Hence I disagree that the days of "personas" or "roles" are numbered. While they may not answer every question or match very well with particular data systems, they are very closely ingrained in our psyche and fundamental to how we define ourselves. It's certainly interesting to examine these issues and ask these questions, however, as I agree that by challenging our unquestioned and fundamental beliefs we learn and grow and discover better ideas, concepts, and applications of those ideas and concepts. I also agree that the limitations of pre-defined personas are certainly very great and overcoming those limitations is a significant challenge.