The Trouble with Scenarios and Personas

Persona from IVREA web site

Scenario-writing is the scariest part of my job. Writing scenarios is a lot like writing fiction the way it’s taught in introductory creative writing workshops. You quickly learn that if you keep the story too close to reality, it doesn’t make for very good reading; a plot that stays true to lived events simply sounds wrong. On the other hand, careful plotting with a few well-chosen details can engender the desired engagement with the story, the willing suspension of disbelief, the unquestioned sympathy for the beleaguered hero.

So it is with scenarios. The best-received scenarios have little to do with people and their observed interactions with technology. In scenarios, e-books are docked snugly in their battery chargers at night; laptops are never forgotten on the roofs of Audi station wagons; and the stars of these convincing vignettes are captivated by your elegant software, not hung up on their stupid complicated messy little lives. They never walk away from the system mid-transaction to answer the phone and forget to come back. Nor do scenario stars doodle in meetings. You won’t find them playing solitaire in class either. A well-loved scenario thrives on deception.

What about personas then? Personas are enjoying great popularity in the software design world of late. They give a friendly face to cold demographic data, add some spice to the blandness of that most generic of terms “the user.” No longer does the UX (User eXperience) team tell the PM (Program Manager), “63% of our users don’t read junk mail” or “79% of our users register on some Web site.” Instead, it’s “Penny Henderson doesn’t read junk mail” or “In the last 6 months, Penny Henderson has decided it’s okay to buy books and clothes over the Internet.” It’s easy to picture Penny screwing up her nose when she receives Teen Sluts spam in her morning email or pre-ordering the latest J.K. Rowling novel on for her Harry Potter-obsessed tween daughter. That Penny’s a good egg.

My colleagues argue that these design devices – scenarios and personas – are most certainly better than nothing. Without them, the story goes, pop-tart munching developers will just pull designs from their Aeron-cradled butts. But I counter, do scenarios and personas actually help, or do they just create a warm illusion of user-centered design?

When personas first caught on in my group, I noticed that one of the other architects had posted some attention-grabbing photos and thumbnail biographies outside his office, thumbtacked to an expanse of bare wall. Who’s that attractive lady trucker, blonde and ponytailed? And what about that stylish insurance salesman – he’s obviously been reading the February GQ.

“Who are those people?” I asked him.

“Those” my co-worker answered, “are our personas. They’re going to use the system.”

Oh. That’s who they are. So. What kind of documents are on that stereotype-busting babe’s clipboard? Do independent truckers like her have the same paperwork as drivers for the big shipping lines? What was her day like yesterday? What’s in that tarp-covered big rig she’s driving?

My co-worker – a good egg himself – couldn’t answer.

Of course that was an extreme case, but I do have a real persona on hand, smiling at me from the window behind this one. You can’t see her, but she’s good, developed by a talented team of User Experience professionals. She’s an accountant, a devoted soccer mom who cooks dinner for her family every night, an avid home gardener, a member of Oprah’s book club. She is mindful of her husband’s high blood pressure and her kids’ homework assignments. And – to top it all off – she’s got a Hispanic surname; she’s the very soul of diversity. The family shares a single computer, a desktop-style PC that’s located in the den.

And we do know what her fictive day is like: Up early, out of the house by 7:30 am, the three kids in tow to drop off at school, a busy 5-hour work day, home to cook dinner, and perhaps spend some time on the computer in the evening, sharing digital photos with family and friends. Of course, she cedes the computer to the kids if they need to go online to do their homework. Bless them – they come first. She knows how to surf the Web, but her husband helps her if she decides to search; she’s an expert Excel user and quite competent at Word. And she’s dismayed by the possibility that her kids might encounter naughty Web sites.

Oh, there are plenty more details: by the time we’ve read the introductory material, we know the dog’s name, which National Park is her favorite hiking spot, and how she investigated her husband’s blood pressure medication on the Internet. In fact, we know her well enough to make up our own stories about her. Those digital photos she takes? She sends them to her mom, who has just now learned to use her own AOL account. We can bend her to our will as long as she continues to fit the profile of the enthused user (41% of whom have had online access for 12-18 months).

Want to make personal security and encryption a necessary feature for her email application? Let’s just give her a secret life in which she’s rekindled a dormant college romance via email. Surely she’ll want to keep those torrid messages to herself. Want to increase her bandwidth requirement? She and her lover can begin using IM to talk in low, breathy voices to one another after their respective families have turned in for the night. That would be frustrating without broadband. Want to make an argument for an external file store? Our persona decides to leave her perfect family – no more dinners taken from the pages of Cooking Light – for what she thinks is her last chance at love, but she still needs to access her files on the family computer. She does, after all, keep her day job in accounting; most enthused users (76%) love their work. She’s like the next-door neighbor who turns out to be Jeffrey Dahmer: I guess we didn’t know her that well.

So scenarios and personas are dangerous. What now? What’s the antidote? Should we go back to simply designing software we ourselves would like to use? I’d hate to see that happen. What I’m suggesting may sound simple-minded, but it might just work. Don’t just use scenarios and personas. Pay attention. Observe. Reflect. We’re all out in the world we design for. Notice, for example, that people use other communication channels besides their computers. Like the telephone. That if their shopping list is on their PDA, they won’t have two hands free to push the cart. That their homes and offices are full of paper, reminding them to do this and that without requiring them to produce any explicit representation of a “to-do” list. Of course, informal observation is no substitute for real ethnographically-derived inquiry, but it’s a start. Perhaps it should be called “feral ethnography.” I’m convinced we can save ourselves from many of our most pernicious design mistakes by simply tempering our lovely technological visions with conscious attention to the world beyond our screens. And it’s fun. People do the darnedest things.

next entry table of contents

copyright 2003 Catherine C. Marshall