The objectives and usefulness of personas are disputed topics in professional HCI contexts. I’d like to reopen the debate, by suggesting that personas are based on at least one flawed assumption.
First, though, a quick definition of persona. Personas are narratives that describe hypothetical users of a website, typically through details about their background, their values, their overall information needs, and their specific objectives on a site.
The best descriptions of personas focus on the way that they can function as communication devices, and drive home to business decision-makers that their choices will affect real users. While this might be true to some extent, I feel that some critique is necessary to clarify that personas are not fact, but instead are a pretty nebulous methodology, and not ’scientific’ tools.
Major philosophical problem 1:
This problem is more a clarification of what personas are not: objective, or able to subject to considerations of proof of concept. Personas are impossible to refute as a scientific technique, because they are based on the construction of hypothetical worlds. As a result, any arguments of whether or not they are good or useful, or bad and not useful, is meaningless. Is astrology useful? If you believe in it, you can use it to validate certain perceived characteristics of your self, others, or situations. But you will never have any evidence as to whether it is valid or not. Personas, I argue, are a similar form of ‘pseudo science’, because there is nothing that can’t be explained through a persona, and there is no single, or even identifiable subset, of correct personas for a given website affliction. Let’s say you are a designer asked to evaluate a website. Can it be said whether there is any truth, in terms of real-world correspondence, between the type of person described and the aspect of the site? Is the persona true? That is impossible to say, regardless of how much sense it may seem to make as one
reads the tasks, needs, etc, associated with the person. If personas live in the world of fiction, how can we possibly tie them to actual aspects of a website and types of people?
Major philosophical problem 2:
Personas assume that there is a real connection between desires (i.e, information needs), actions, and aspects of a website. But these things are not at all necessarily contingent, and I argue that the specific connections personas imply of these sorts are probably quite often completely false, if not verified through other, less subjective techniques. The problem is that, stated in terms of information beliefs and personal values, personas are only a re-description of the action that they are thought to be predicting, and that that action is only a re-description of the aspects of the website that it is thought to be linked to. But in reality, the connections personas imply are logical, not contingent — they are created in the mind of the designer who writes the persona.
In sum, a persona is in my opinion best described as a label. A complicated, completely arbitrary label, used to direct focus to an aspect of a website, but without any foundation in fact.
If you are a huge persona fan, feel free to weight in. My intention in this blog post was not re-hash obvious facts of personas, but instead to touch on some of the more tacit assumptions that go along with them.
6 Comments
I’d say you’re actually rehashing something that is VERY obvious to anyone knowledgeable about personas: of course they are not objective! They are not meant to be – are use cases objective? Are state diagrams? In what way are they so? Objectivity (or subjectivity, for that matter) is not the point here – the point is if they are good at what they do – personas can filter, give insight, focus, whatever, depending on how you use them – it’s a working tool. Ethnography or working with other forms of anthro methods are not “objective” either – they are observations from the outside, since we cannot look into the users minds. Hence, they’re interpretations of behavior, needs, experience, and such. They’re basically stories we tell (and write) to gain insight into the other. How can that form of insight be objective? If we’re only interested in use analytics (clicks, bounces, whatever…), I’d say we miss a great deal of knowledge.
You also seem to claim that “the designer make up fanciful ideas about users he or she thinks are gonna use the site”…for most part, I would argue that personas are not necessarily to be made by the “designer” as such, but rather by some form of stakeholder collaboration. Personas are not brainchilds of the designer, but a collaborative, cross team achievement that focuses on specific aspects of the users lives.
Just my thoughts on this…
you can use a methodology on personas that is reaasonable, maybe not the scientific method persay. I mean scientist’s explanations for atoms and the like are all creations of the imagination used to explain observed phenomenon. Likewise you can create pages that appeal to problems of personas that exist and pages that don’t and see how they perform. If one does better than the other, the best explanation is that personas account for the difference in performance . . . and you can refute astrology.
Hi steve hi Mads
I am sticking by my point that there are assumptions that are not considered around personas. Namely, they assume (and thus) create a construct that there is no evidence for, between user characteristics and aspects of a site. My thought is that before an entire company gloms onto these connections as ‘truth’ they need to be verified by a more empirical method. You can address problems that you know to be relevant to actual people, but I don’t believe in creating ideas of problems as connected to personas until you have some evidence. Humans are very good at reading causal associations into phenomena when those associations don’t actually exist, so I see not doing this as dangerous to good web design methods.
And I argue that while scientists (like physicists) must create definitions and theories around what they think they observe, the work of say, relativity theory, has led to the discovery of numerous other novel facts. In ‘degenerating research programs’ (Lakatos) on the other hand, theories are fabricated only in order to accomodate known facts. Personas don’t qualify for even this, until the connections between users and aspects of website have been established. I have no problem with personas, as long as some facts are brought into the picture.
You can refute astrology, but skeptics will never be able to prove its lack of worth given a goal like ‘unearth previously unknown insights into user behavior’. On the other hand, believers will never be able to prove its worth. It is not falsifiable. Given our talents for seeing nonexistent patterns, I think that if major time is going to be invested in a method, we better have some evidence that it works to reveal verifiable truths, not just to construct them.
between user characteristics and aspects of a site. -> I guess it’s common sense Jessica. Businesses understand the characteristics of their clients, and at least what drives them there, through work experience and knowledge of the industry. Which can all probably be found by simple research. The connection in the site is the stuff you put in that appeals to that characteristic. For example, the way I describe what I do to someone who is in my field is different from the way I describe what I do to someone who has no experience in it. If you reject that there is any relationship between a site’s characteristics and how it appeals to people’s characteristics, you should probably have a reason for it because it goes against a very appealing and simple explanation.
With astrology, I think the way to falsify it is just through occam’s razor. i.e. it doesn’t provide anymore explanatory power than what it overrides, so it’s not true.
I don’t at all reject that there are relationships between a site’s characteristics and how they appeal to people’s characterstics. What I am saying is that personas are not a tool for identifying them. From what I have seen of HCI methods, personas have a place within a larger qualitative framework (interviewing users, etc), all of which are used to inform new designs. But this is very often all qualitative work. HCI is known for very small sample sizes (like, under 5 people is not at all uncommon). The time it takes to create personas, explain them etc, not to mention the interviewing or other inquiry methods that feed into their creation, is all invested when there is still no empirically sound evidence for what they ultimately end up creating.
Honestly, I have no real problem with personas; I think that they can be useful in some circumstances. I just think that the objectives and level of validity should be made clear, especially when they are being created for people who are not so familiar with methods for user research.
personas are not a tool for identifying them -> I agree. I don’t think it’s always necessary to do interviews of people and the such – you might if you feel like you don’t know your target audience. If you already are in a place where you can identify characteristics of your audience, then personas are just a summation of those qualities. It is usually put around a lot of rosey language, but you can certainly strip it down to the same thing you might get from interviews and other stuff.
When I went to school, chemistry teachers used to talk about atoms and stuff like they were people. Personification was common to help concepts sink in. I think this is sort of like that, creating something a little more elaborate is entertaining and helps the list of qualities sink in.
I definitely think it is possible to come up with emperical evidence as to whether or not personas work – this is where I disagree with you. I haven’t seen any. But that’s outside my field. It might be accepted based on common sense. But definitely would be nice to have studies about – I’d read it.
I think it’s easy to make a website without thinking about who might be, who’s supposed to be, and who you want to be reading it. This gives me a little bit of interest in personas. I’m sure this leads into all sorts of things about using emotions and empathy in marketing – something that is probably really important for people to identify.
Post a Comment