Tom Chi  

Jack and Jill

April 9th, 2004 by Tom Chi :: see related comic

Personas. Love ‘em. Hate ‘em. Struggle to Integrate them. Are they truly useful, or a little hokey?Perhaps one of the more controversial aspects of personas are the colorful narratives which are created around them. Who could possibly care that Persona 2B a.k.a. ‘Ted’ drives a Blue Buick LeSabre? Or that he is allergic to shellfish? Even when the stories are centered around work, there is always that nagging voice which questions the relevance and applicability of the information being invented. Are we really modeling users well, or simply creating stereotypical users from unfounded assumptions?

One thing that can be said for these stories is that they are memorable. This is because narratives are excellent mnemonic devices. They create a temporal and causal framework which allows our brains to store quite a bit of data. Mention one detail about ‘Juanita’ and the rest of the story comes flowing back: Oh. She has two kids and has to work late at her financial services company… and because she stays late she starts tasks which require deep concentration after 5pm. And so on. Of course such stories can also get mangled after a generation of retellings — but for the most part, personas do a good job of creating a shared vocabulary to call up significant detail about user segments.

While having a shared vocabulary is better than having none, it’s quite the tricky endeavor to develop the right vocabulary. Even if you’ve profiled 100 users to develop your set of personas, oftentimes you’ll still find users who don’t quite fit in the categories you’ve created. And even if you do successfully bin them, it’s often not clear how that bin should inform design. For example, most persona sets have a user which is the ‘novice’ user. What does this mean though? Can you assume that said user understands drag and drop? right-click? tabbing through fields? In testing you might see that novice user #1 might understand two of the concepts, and novice user #2 understands the third. They are the same persona in front of the same screen, but the results will be vastly different.

There may also be environmental differences which drastically effect how two people which are the ’same’ persona interact with the software. While Joey and Tina may both be mid-level managers in a large organization, Joey is beset with phone calls every 10 minutes, while Tina’s day is more stable and structured around meetings. As a result, Joey keeps on timing out on his web sessions because important calls come in while he is halfway through a task, while Tina is able to pick good times to approach these tasks and has no timeout problems.

If personas cannot address these very real interaction design problems, at what level are they useful? Do they only become powerful when paired with use cases, or buoyed by a certain approach to testing? Should they be written for the designers, the developers, the executives? I’d like to open the debate up to how successful our readers have been with personas in the past, and comments as to their appropriate role and scope in a project.

14 Responses to “Jack and Jill”
Keith wrote:

Love the comic this week — it’s a classic.

I’ve recently posted about personas as well where I talk about how I feel personas can be useful for developing and writing content.

I also think that they can be useful to illustrate your users/readers/visitors to stakeholders/clients. It’s can be very hard to advocate for users when you’ve got to deal with people who’ve got no concept of usability. I’ve found personas to be extremely helpful in that respect.

I used to question the amount of effort that goes into producing the personas themselves (keep in mind, I’m not really what you’d call an IA or usability specialist) so lately I’ve been scaling that back a bit and finding I’m getting much more out of that process.

It’s not as…scientific, but it gets the job done.

Jake Cressamn wrote:

Really appreciate your insights on personas.

Aside from their usefulness, we’ve had difficulties getting clients to buy into the process of creating them. When it’s their budget and resources that are going into the decision of what car “Sarah McFudd” drives and whether she prefers Eddie Bower or Banana Republic, stakeholders can get a little anxious for conventional results.

Jake Cressman

Tom Chi wrote:

One benefit that I didn’t mention in my article is that it gets everyone thinking about people instead of technologies. Without personas, it’s very easy for developers to spiral into discussions about database schematization during interface design meetings. It also works for marketing too. Instead of thinking of people as market segments from which to derive revenue at a specific conversion rates, they start to consider how the atomic user will react to and use the product.

That said, it may be that personas are not as useful for HCI people in shaping their design decision, as they are a construct to involve a number of groups in user-centered thinking. But, as Jake has pointed out, it is hard to get the positive aspects without a clearer tie to business value.

Ronnie wrote:

Great article. I agree with the previous two posts, as there’s often pushback on the persona creation budget. Ideally, you should be able to get personas ironed out for your interface within a day or two, provided you can get and spend the time analyzing the site, and talking with business owners to understand the current and expected users.

I’m a fan of personas, and there is something to be said for diving into the deep details of a persona (e.g. Quentin likes Raquetball and natural foods), but I prefer to first create the more basic ’shell’ of a user. For example, with a broad swipe of an expected user type, we can create Patty, stay-at-home mother of 2 teenage boys, who drives a minivan, lives on her cellphone and uses the internet mostly to send photos and shop for clothes from her favorite online stores. This is usually enough to use as a starting point for user class analysis. Her proclivity toward purebred English Bulldogs, her excessive facial hair, matching track suits and favorite Soap Opera star might help me paint a better picture of who she is (and what she looks like - in this case, Patty is quite a gem… can’t you just picture her?), but this knowledge rarely tranlates into whether she’s going to buy that Danielle Steele book online.

Rather, I like to position those tidbits of information as parts of Patty that will ‘grow’ with her persona as we test it. It can actually become somewhat of an icebreaker with other UE folks and business owners over the design phase. It’s a fun, bonding exercise to decide Patty’s hair color, how much makeup she wears, and whether she’s deciding the wax that ’stache etc..

Just my 2 cents.

Lynne Duddy wrote:

This article gets people focused on the people - not the technologies - and this is the real benefit of personas. Technology is a wonderful thing and it is becoming more and more woven into our lives; but without understanding the people, their needs, behaviors and motivations, how can be make intelligent design decisions. That’s not to say that personas can provide all that, but it is certainly an integral piece that can be used to help the business stakeholders, developers and designers put a human face of the people that will use the product.

Tina wrote:

I’ve cancelled all my meetings and proposed to Joey, but he’s engaged…

mindful_learner wrote:

I believe the key behind the success (or otherwise) of personas is the degree to which they get you to focus on overall user GOALS rather than specific interaction techniques. I don’t believe they are the correct tool for deciding whether a group of users can do drag-drop or right-click. They ARE the right tool to help you get to the heart of what users are really trying to do rather than the little tasks they need to complete to get to where they want to go. I still believe Cooper gives the best illustration of the power (and the rational behind) Personas in his Inmates book.


Tom Chi wrote:

While I agree in theory that personas should be used to set user goals, there is a point where you need to make those interaction decisions. I might produce 5 designs which all facilitate the same user goal, and personas will not help me there.

If personas are only to be used for establishing user goals, why would a team choose them over say, a series of contextual inquiries? Why not any of the many possible requirements gathering techniques which have user goals as their output? What makes personas more valuable or sucessful than those?

Lisa Hagstrom wrote:

I agree with most of what’s been said so far, but also want to add that I usually point out up front what the personas are used for, to differentiate personas intended for branding, vs, task oriented personas. One can’t apply 3 personas to a myriad different business objectives or purposes. Rather, it’s important to focus on what they are intended for, so they are relevant to the audience that is going to use them. That will cover why it’s important to know what brand of car Mary drives vs. the types of information structures John prefers and finds useful. Providing the lens in which to view the personas helps audiences better understand how to use them and what they are intended for.

Crybabypirate wrote:

In response to Tom C. - I have had extensive training in both Cooper-style persona creation and use and Beyer/Holtzblatt’s Contextual inquiry.

Beyer/Holtzblatt’s techniques are great for people who don’t have much experience with/feel for the human side of the design equation, and also for managing the communication overhead of larger teams in which everybody will participate in the design process. And, while Contextual Inquiry does a great job of orienting the team in the right direction, there are no guarantees the resulting design will be good. In the end, you still are relying on solid design talent. Also, I think Beyer/Holtzblatt’s methodology suffers from lack of means to assess design inflection choices, which is one of the primary reasons personas are so great. Finally, there is just so much overhead using Contextual Inquiry. It burns so many ‘man hours.’ And you’ve got to buy all of those specially sized stickies and inhale blue sharpie stench.

For experienced designers, I think personas are a much a more efficient way to get to a *better* design than Contextual Inquiry. The catch is that there is an art to creating truly representative personas and balanced persona sets. I think a lot of people who claim that personas are an inadequate design tool are in fact using inadequately crafted personas. A persona set is a model of the human system. The model is useful only insofar as it accurately and comprehensively captures the relevant patterns across the human system. And, even if your personas are spot on, you still have to have a very solid design toolkit to get a good result.

Kevin Cheng wrote:

For experienced designers, I think personas are a much a more efficient way to get to a *better* design than Contextual Inquiry.

Interesting perspective on contextual inquiries. I’m curious to find out how one goes about developing accurate personas without tools like contextual inquiries. For me, CI’s are so important in understanding users, especially when I’m learning about a new industry. I don’t really feel like Contextual Inquiries and Personas are mutually exclusive but rather, the latter can only be effective with data from the former.

Crybabypirate wrote:

I think perhaps you are using contextual inquiry in a more colloquial sense than I intended. I was referring specifically to the Contextual Inquiry methodology developed by Beyer/Holtzblatt.

Personas should always be grounded in qualitative research, which may or may not adhere strictly to the methods set forth by B/H. There is no reason why one couldn’t develop personas from the research conducted as part of a B/H style contextual inquiry, but it’s my impression B/H feel their approach obviates the need for personas.

Kevin Cheng wrote:

In an article on USAOnline, Best Buy uses personas to revamp their stores. One persona’s name is “Jill”, by the way.

piskodrocho wrote:

I want to listen good music!

Leave a Reply

OK/Cancel is a comic strip collaboration co-written and co-illustrated by Kevin Cheng and Tom Chi. Our subject matter focuses on interfaces, good and bad and the people behind the industry of building interfaces - usability specialists, interaction designers, human-computer interaction (HCI) experts, industrial designers, etc. (Who Links Here) ?