One’s destination is never a place but rather a new way of looking at things.— Henry Miller
I was thrilled to be able to share my design-focused narrative taxonomy concept and process at the Austin UXPA, hosted by Rackspace. Great crowd, great discussion, great experience. Thanks to organizer (and user research guru) Candice McFarland for organizing this!
Originally posted on former personal blog UXtraordinary.com.
Cross-posted and expanded from my LinkedIn account.
Everyone demonstrates the fundamental attribution error—a variation of correspondence bias (pdf)—to some extent. We look at the action and assume it’s the character. Even when we know there are extenuating circumstances we do it. The defense lawyer, doing his duty to provide the best defense possible, is seen as supporting crime. The debate student, assigned to defend a certain position, is seen as believing it; no matter the usefulness of the role or the purity of intent, every devil’s advocate runs the risk of being seen as devilish. And of course, the criminally negligent incompetent person driving the car that just cut us off.
In the workplace this can create misunderstandings, usually small but sometimes project-killing or even career-destroying. It’s a problem because the only way to overcome correspondence bias and not commit the fundamental attribution error is to constantly question your assumptions and opinions, looking for the larger context.
Since we’re all story-driven creatures, sometimes an anecdote can help. This is a story of a time a life was on the line, and it’s the best example of correspondence bias I know.
My mother’s uncle was a man named Jara. He was my grandmother’s brother, an artist when he could be (I saw beautiful sculptures and drawings in his widow’s home). His best friend, whose name I don’t know, was a professional artist.
During Hitler’s occupation of Prague, Jara and his best friend were sent to a labor camp. At the camp worked another man whose name I don’t know. Let’s call him Karel. Karel worked as an overseer, managing his fellow citizens for the Nazis. Karel was hated. He treated everyone “like a dog,” Jara said, swearing at them and driving them mercilessly, generally making the labor camp experience every bit as awful as you imagine it to be.
Watching Karel’s behavior, day in, day out, Jara and his friend eventually realized Karel could not be allowed to live. It was obvious to them. Karel was a traitor, a collaborator with the enemy, and responsible for much misery. They were young, and passionate about their country. They made a pact together, that if all three of them survived the war, Jara and his friend would hunt down Karel and kill him. They viewed it as an execution.
The war ended, the labor camp closed, and life continued for all three men. Jara and his friend discreetly found out where Karel lived. They obtained a gun, and one day they set out to his home.
Karel lived outside Prague, in a somewhat rural area. When Jara and his friend arrived, Karel’s wife was outside, hanging laundry. When they said they’d known Karel at the labor camp, she smiled and invited them in, calling to Karel that friends from the camp had arrived. They followed her to the kitchen, where they found the monster they sought.
Karel was sitting by the table with a large tub of water and baking soda in front of him, soaking his feet. He was wearing rolled-up pants, suspenders, and a collarless, button shirt, the kind you could put different collars with under a jacket. He greeted them with a broad smile, immediately calling them by name and introducing them to his wife. Jara said Karel was so happy, he had tears in his eyes. He asked his wife to give them coffee, and she brought out pastries, and all sat down to talk about old times.
Jara and his friend were dumbfounded, but did not show it. During the conversation they realized that Karel had not thought of himself as collaborating with the Nazis, but as mitigating their presence. He was stepping in so no one worse could. His harshness was protective; the Germans could not easily accuse the workers of under-producing when Karel pushed his fellow Czechs so hard.
They stayed several hours with Karel and his wife, reminiscing and privately realizing no one was getting shot that day, then took their leave. On the way back they threw the gun into a pond. Jara went on to work at the Barrandov film studios, where he met his wife, Alena (she was an accountant). They married, lived a long life, and were happy more often than not.
Jara was transformed by this experience. Never again would he take any person’s actions as the sum of their character. And I do my best to see things in context and not judge, in part because of the man my great-uncle didn’t kill.
Leadership is the big skill of getting groups of people to complete the right work.— Clark Aldrich
At SXSW I had the pleasure of sitting in on Andy Barr’s and Sarada Peri’s 37 Practical Tips to Help You Write & Speak Better. Much of these focused on simplicity of grammar and content, and reminded me of the writer’s vow of chastity my husband conceived back in 2007. Since then, my first drafts attempt to follow the below rules as much as possible. Only then do I go back and add anything more, trying to restrain myself to choices that add clarity.
So here it is.
Just as Lars Von Trier and Thomas Vinterberg wrote the film maker’s vow of chastity (also know as Dogme 95), so Bart has proposed a writer’s vow of chastity.
A draft of the rules as of August 4, 2007:
The writer’s vow of chastity
The writer will use no modifiers.
- No adverbs.
- No adjectives.
The writer should act as a behaviorist.
- No words describing emotion.
- The writer will not make the reader directly privy to a character’s thoughts (no interior dialogue or interior monologue).
The writer may break these rules only when it is unavoidable.
The above may be summarized as, “Not doing the reader’s work for them*.”
*The summary references advice from C.S. Lewis to his students:
Don’t say it was “delightful;” make us say “delightful” when we’ve read the description. You see, all those words (horrifying, wonderful, hideous, exquisite) are only like saying to your readers “Please will you do the job for me.”
Barr and Peri also provided a page with the 37 tips, on which I took copious notes. Here it is—enjoy!
Originally posted on alexfiles.com.
Simulations are powerful when students need to be engaged more than they are. Clearly, this is an area in which distributed classrooms have suffered, as death by PowerPoint has not just been refined in many programs but almost weaponized to military specifications.— Clark Aldrich
Love data visualization? Love pies? So do I.
Last Tuesday we had a Thanksgiving potluck at work, and at the instigation of a coworker I made a Venn diagram pie. Here’s how I did it, if you want to try it yourself.
What you’ll need
You can use whatever flavors you want, but remember they have to combine into a pleasing flavor profile. The flavors of the three ingredients I used were pumpkin, pecan, and apple crumble.
The recipes were fairly straightforward. Since the undertaking was complex I went simple with the recipes. For pumpkin and pecan filling I adapted About.com’s Southern Food Classic Pumpkin Pecan Pie recipe. For the pumpkin filling I bumped up the ginger and added nutmeg. For the pecan filling I used dark corn syrup, replacing about 1/3 of the dark syrup with maple syrup and a little molasses.
For the apple pie I used Cortland, Gala, and Honey Crisp apples. (I recommend The Apple Works for good information on what apples work best in what contexts.) Once again the About.com Southern Food section provided a good Apple Crumble Pie recipe.
I’d never used it before, but Pillsbury’s rolled up pie crust did pretty well!
So, to make this happen you need the following:
- Two sets of aluminum cake or pie pans (4-6 pans). Cake may allow you to overlap the three pans and the middle section more easily, but pie works, too. You’ll need three pans for baking, and at least one extra to make the middle crust.
- Aluminum foil to cover the pie plates and prevent leaks. You’ll also need it to protect the crust while baking.
- Enough pie crust for the bottom of four pies. This will cover the three-plate section, provide crust for the middle, a base to hold the middle pie crust in place, and a little extra in case you want to add decorations.
- Pumpkin pie filling to taste.
- Pecan pie filling to taste.
- Apple pie filling to taste.
- Crumble mix (no pecans to start).
- Pecan halves for topping and to add to crumble mid-way.
- A cookie sheet to support the pie plates, which will not be structurally sound enough to support the weight of the pie.
- Two six-packs of graham crust mini-pies (for excess filling).
Making your Venn pie plate
Here’s how the three pans overlapped. Note how corners are folded over. I used an ancient pizza pan for support instead of a cookie sheet (most cookie sheets don’t fit in our tiny oven).
A detail from the bottom. Cut your flattened sides into sections so they lie flat and don’t warp your pan.
Line your completed pan with aluminum foil to cover the sharp edges you’ve cut and prevent leaks.
Making your crust
Preheat your oven to 350°.
Unroll your pie crust and lay it out. Cut away excess and press edges together so you have a continuous bottom that isn’t too thick.
What not to do
The image below, with raw dough and supporting aluminum, does not work. The crust melts and doesn’t hold its shape. Thin metal dividers also do not work: they leak abominably. I learned this making my first “pie chart” pie; metal dividers were much more trouble than the crust below (for one thing, I had to tilt the pie in the oven until the pecan filling set).
Partially bake your pie crust edges and bottom about 10 minutes at 350°. Also bake sections for the middle crust, separately (see below). Note that the middle crust has a slightly tighter curve, to make the overlapping areas slightly egg-shaped instead of pure circles. This will give you more space in the middle sections. Don’t forget to puncture the crust with a fork to avoid bubbles!
Important: reserve extra unbaked pie crust. You’ll need it to make the middle section work properly.
These are the crust pieces you’ll need to shape the middle of your Venn pie-agram. I used six crust lengths: one long curve, one not-so-long, three short ones, and one tiny one.
To make the crust stay in place and reduce leaks, use unbaked dough to hold the partly-baked middle sections in place. It doesn’t need to be perfect. Tip: Use a bread knife to gently saw the pie crust sections. Pie crust is crumbly. Try to make them line up naturally with the outside edge.
Filling and baking
Tip: Start with the firmest filling first. It will fill up any weak spots a more liquid filling might break through, keeping your sections more compartmentalized. Here, the apple pie section has been filled and covered with crumble; apple has been layered in the bottom of its three overlapping sections in the middle. To the right, pumpkin filling (my next step). The pecan filling is in the center; pecans have yet to be added. I used crumbled pecans for a thicker mixture. Pecan filling went in last.
Here’s how I did the fillings, in order:
- Apple crumble
- Apple covered with pumpkin filling with normal crumble (no pecans)
- Apple covered with pecan filling
- Center: Apple covered with pumpkin filling with pecan crumble (I added pecans into the food processor with some of the crumble mix)
- Pumpkin covered with pecan
- Cover the pecan area with pecan halves; place pecan halves in the three pecan-containing middle sections
Ready to bake! Note the aluminum foil protecting the edges from over-baking. You can see some of the mini-pies I made using the excess filling.
The pie baked about an hour before a knife came out cleanly from the pumpkin filling. The mini pies, which I baked after the pie, took about 35 minutes without a cookie sheet. Oven heats vary, so check your pie at around 45 minutes, and your mini-pies at 25 minutes.
Originally posted on UXtraordinary.com.
I’ve submitted a proposal for SXSW 2014! Vote here.
User experience and storytelling go hand in hand. UX professionals consciously apply personas, use case scenarios, underlying narratives, content strategy, and visual elements to provide a stage on which users play. But there is a key element missing in this drama: taxonomy. Much more than mere collections of categories, hierarchies, facets, and navigation elements, taxonomies “narrate the natural relationships between concepts.”
That quote is from a 2011 article in the journal Data & Knowledge Engineering, in which three computer scientists explored narrative taxonomy from the perspective of data algorithms in effective, adaptive content retrieval. This session will discuss the implications of narrative taxonomy for user experience design; specific kinds of taxonomy stories; how to recognize, analyze, and apply narrative taxonomies; and the results of user testing against different narrative taxonomies in different contexts.
- How does a taxonomy tell a story, and what are the narrative elements of that story? How can a UX professional recognize and analyze them without the help of sophisticated database algorithms?
- Why and when is analyzing taxonomy from a narrative perspective helpful to UX design? How does it compare to more typical approaches?
- In UX, taxonomy is primarily addressed by information architects and content strategists, but not always by both at the same time. Can narrative taxonomy bridge gaps between these different specialties?
- What does narrative taxonomy mean from an interactive perspective? What are concrete examples of this? Can users tell their own stories?
- How can narrative taxonomies be tested, and what are the results of such tests?
Pabini Gabriel-Petit approached me for an article in UXmatters in May, 2012, and in July published Intention-Focused Design: Applying Perceptual Control Theory to Discover User Intent. Below is the article as it appeared.
At this point in the development of the field of user experience, I’m assuming that most good UX professionals know how to tailor sites or applications to user profiles, create personas, and tell a compelling story that drives user process flows. But sometimes we encounter a situation that’s a bit more challenging: we’re asked to design one product for very different users—or even users with seemingly conflicting goals.
Without a unifying narrative, such challenges can result in compromised user experiences. A client, or even a UX designer, may find it simpler to either target the most valuable or common user profile or to design very different process flows and interactions for different users. These approaches aren’t necessarily bad, but integrating them gracefully is difficult without a shared context. Intention-focused design is a specific UX strategy that can help you to discover hidden and shared user narratives.
Perceptual Control Theory and User Intentions
You know you’ve got a good piece of software when people use it for purposes for which the designers never intended or designed.
Psychology is the UX designer’s friend. We use it all the time. Our process flows and interfaces apply learning, perception, and cognition theory. Decision-making research, gamification, and emotional appeals evoke the appropriate response to the stimuli we provide. But the most useful framework for understanding users that I’ve encountered is a little-known system called Perceptual Control Theory (PCT). The invention of William Powers, an engineer with degrees in physics and psychology, PCT is based equally in cybernetics and psychology. It has recently been gaining ground as a useful framework in areas such as education and psychotherapy.
The basic premise of PCT is that human behavior is not about the behavior itself, but about reinforcing desired perception. William James said, “My experience is what I agree to attend to.” William Powers suggests that behavior is not just about agreement, but about constantly refining our experience to achieve an agreeable level of perception.
For example, a bagger at a grocery store who packs items carefully—cold with cold, fragile on top—may be reinforcing a self-perception as a caring, careful person. An executive who signs the order for a layoff may be reinforcing a self-perception as a strong, dominant person. A voter voting against his own interests may be reinforcing a perception of himself as a moral person. A programmer who writes elegant code may be reinforcing a self-perception as being a particularly rational human.
PCT suggests that a person starts off with a desired reference point for her experience. (This may change over time.) She compares her environment to that reference point and takes action to bring it closer to her desired perception. She compares the result to her reference point and acts accordingly. If something external, called a disturbance, interacts with her environment, the same comparison of input and reference point recurs, resulting in the appropriate output, or behavior. This feedback loop is the core of PCT. As Powers wrote, “Control is a process by which a person can maintain some controlled variable near a reference condition by varying actions that oppose the effects of disturbances.”
In UX terms, users actively work to optimize their own experience as much as possible. Let’s take a look at how a user-focused PCT feedback loop works, as shown in Figure 1.
PCT goes further: it reminds us that what we think a person is doing and what they think they’re doing can sometimes be very different. Powers demonstrated this very simply at a presentation to the American Education Research Association in the 1990’s.
Using a large, knotted rubber band, a chalk board, and a piece of chalk, he asked a volunteer to maintain the knot over a dot, while also holding a piece of chalk there. Powers then pulled the other end of the rubber band around in a large circle. Although Powers took care to move deliberately and not too fast, the volunteer’s chalk described a smaller circle on the board. Powers then asked the audience what the volunteer had done, and they replied that he had drawn a circle. But when Powers asked the volunteer what he’d done, he said he had held the knot over the dot.
The volunteer neither expected nor intended to draw a circle, but worked to maintain the knot’s placement. However, despite hearing the instructions and watching the demonstration, the audience had seen the visible evidence of the volunteer’s behavior, the circle, as the primary object. An analyst looking at the behavior without having heard the instructions might easily take the circle to be the purpose of the exercise.
PCT is valuable to UX designers because it helps them discover what users actually think rather than what we think they think.
Several design challenges benefit directly from intention-focused design. Next, I’ll discuss how to apply intention-focused design when we’re asked to transform a site or application.
Occasionally, a UX designer or creative team receives a request to completely re-envision a site or application. Perhaps the company’s business value is changing, or the site has to adapt to the mobile world. Yet the site may have a strong following of users who are content with the site as it is, so resist change. Often UX designers offer users a more usable interaction, then are surprised by a vehement negative response. It’s a truism that users resist change. The PCT insight that people desire equilibrium at their personal reference point goes to the heart of this.
A person’s desired reference states operate in a hierarchy of importance and awareness. People work to maintain equilibrium at different levels of being. I’ve adapted this hierarchy for a compassionate animal lover from a 1982 diagram from psychologists Charles Carver and Michael Scheier:
- System concept: Be a compassionate person.
- Principle: Be kind to animals.
- Program: Walk the dog.
- Relationship: “Dog walkingness” sequence: Keep the dog from running into traffic by pulling on his leash, tightening his choke collar. (I know, I know. Wait for it.)
- Transition: Pull on the leash.
- Configuration: Fingers around the leash.
- Sensation: Friction of the fabric against your fingers.
- Intensity: Muscle tension.
The animal lover simultaneously compares all of these perceptions to an expected, desired reference point, but those that are lower in the hierarchy can change in response to new information—provided they match the deeper expectations of system and principle. But I know that choke collar is bothering you, so let’s provide our dog owner with new information: choke collars are painful to dogs. The animal lover’s self-perception as a compassionate person and the principle of kindness to animals trumps maintaining equilibrium in the “Walk the dog” program or “dog walkingness” relationship. A new program appears: Buy a walking harness.
You have to address not just the level of environmental equilibrium, but the deeper levels of purpose and self-perception. Not doing this can result in debacles like the Netflix separation of streaming and DVD viewing. Had anyone at an empowered level said, “Customers don’t want DVDs or streaming video, they want movies and entertainment,” they might have averted the loss of customers. Great design actively engages purpose and self-perception to change the user’s environmental expectations, effectively resetting their equilibrium to a new state.
The Hidden User
Sometimes a site’s users are not the only participants in its narrative. For example, the medical software company I work for provides software for quite a few highly specific user personas: nurses, therapists, schedulers, billers, marketers, and more. Some of these are not only very different, but initially seem to have conflicting interests.
I struggled with this when I first began—looking for the underlying hook that would tie a good, user-centered design together; chunking out different user activities; getting data from customers and coworkers who knew them deeply. But it was a PCT analysis of their intentions that exposed their shared narrative. Part of every persona’s internal script was a hidden user the persona focused on every day: the patient. From then on, patient first became the design mantra, and I’ve set aside all design that doesn’t center on that hidden, absent user.
A retail site offering educational toys and tools for children offers another example of a hidden user. Parents, educators, and others are on the site, but the purpose they all have in common is a better-educated child who enjoys learning. That child is the absent user, providing the context for a taxonomy based on age, special needs, specific curricula, and other factors driving customer intention.
When designing solutions for user activities, it is very easy to fall into the common-sense trap. Anyone trained in experimental methodology knows that common-sense assumptions are often proven wrong. For example, in a study of the likelihood that drivers would call 911 or otherwise act in a dangerous situation, the results were surprising.
Someone looking like a small child walked beside a rural road, a suburban road, and a busy highway. Common sense might make people expect more 911 calls from the busy road, because more people were there. But the opposite occurred: the more people were present, the fewer calls or offers of help they made. Analysis revealed that drivers on the little-used road figured there might not be anyone else to help, so they took action. Drivers on the busy highway figured someone else must’ve already called, so took no action. The response on the suburban road fell in between.
My favorite example of a common-sense design mistake is the typical business mental model for sharing user-generated content. Viewing users as single units in a much larger pool of users, who touch only some of that pool, businesses place a user at the center of an ever-expanding circle of sharing: user, friends, site members, public, as shown in Figure 2.
But users don’t perceive themselves that way. They experience themselves as the horizon between their self and their world, and they adjust how they share according to a greatly varying series of needs. With a friend, they might share dreams and romantic details, but not the physical details of a medical problem. With a doctor, the opposite holds true. Both kinds of data are intimate details, but shared intimacy is not the same across the board. Different groups of people receive different kinds and levels of information.
Here, the user’s system concept might be a desire to be likable to as many people as possible. As a result, the user might share only information that is appropriate, to avoid putting people off by sharing too much. Of course, people vary, and context should always drive analysis of user intentions. LiveJournal got this years ago, offering highly customizable, user-created groups years before other social networks—but the social circles of Google+ are hands-down the best design implementation of it.
Intention-focused design deliberately empowers users as active, goal-driven participants in shaping their experience. In this article, my goal was to introduce the concept of Perceptual Control Theory at a high level, showing why it redefines our basic understanding of user analysis beyond site interactions and immediate goals to deeper levels of intent. As examples, I showed three strategic uses of intention-focused design: getting a handle on site transformation, discovering hidden users, and empowering users through design.
Carver, Charles S., and Michael F. Scheier. “Control Theory: A Useful Conceptual Framework for Personality—Social, Clinical, and Health Psychology.” APA Psychological Bulletin, 1982, Volume 92, Number 1.1.
Powers, William T. Behavior: The Control of Perception. 2nd ed. Montclair, NJ: Benchmark Publications, Inc., 2005.
Paul Bryan, of the LinkedIn UX Strategy and Planning group, contributed There is no such thing as UX strategy, on UXmatters. Bryan’s clearly got a handle on the subject, but some of the user responses (“This UX Strategist role should be a skill of a PO;” “I thought we decided there was no such thing as UX Strategy…and that UX was strategy?”) revealed a widespread lack of understanding on what it is and who should do it (and no, it’s probably not product owners. A PO truly gifted in UX is not only extremely rare, but has many non-UX roles to fulfill. Adding this to their plate is the wrong call.)
While I agree with Bryan’s thesis that there should be better understanding and use of UX strategy, but in the article he behaves as though this is a goal. Having seen UX strategy happen and consciously done it myself for the better part of a decade, I submitted the following comment:
It is real, and it’s happening in some places. I think the reason for the confusion lies in lack of definition about not simply what UX means, but what strategy means.
For me – and I’ve been doing user experience under one title or another for over 14 years now – UX is “everything that is the case;” it’s everything the user experiences in the context of your brand. Designing good UX is not possible without understanding product strategy; designing great UX is not possible unless product strategy integrates UX strategy. Frequently the only person who can do that is the UX designer, unless you’ve hired product people with design backgrounds, which is rare. User experience rests on the three pillars of user research, usable IA/UI/IxD, and purpose-driven vision. You have to understand your users’ goals, your client’s goals, and be able to bridge them.
So there are different flavors of UX strategy, and a good UX strategist uses them all at different times.
- Brand-integrated user experience design that is not only usable and delightful, but actively furthers the brand. For example, it’s not enough to simply provide a space for a promo on a page; the UX designer should help drive which promos will not hurt the purpose of the page, and may even increase user value and enjoyment. You have to integrate the web, print, TV, off-site advertising, enewsletter, and other items to have an integrated UX strategy. (Yes, this does actually happen at times.)
- UX evangelism strategy. Figure out how to get people thinking in user terms. At a highly numbers-driven social network, I introduced them to the concept of measuring not just user-generated content (UGC) but user-generated activity (UGA), lumping it all under user-generated experience (UGX). Product owners and others measured UGA on their own, which forced them to think from the user’s perspective.
- Research-driven UX strategy. In the example in the second bullet, UGX became a strong driver of overall UX strategy – we consciously presented activities to users in a particular order, based on user research and testing, designed to both optimize their experience and increase the ROI on their activities. We also studied communication patterns of UGA and UGC, determining where the best user value and ROI lay there as well.
- Road map strategy. As user advocates and researchers, UX strategists can contribute significantly to road map work. For example, putting on our analytic hats we can show product strategists how to objectively measure concepts they tend to consider intangibles, such as competitiveness. We can also show how UX focused strategies such as the ones above can be integrated into their road map for the benefit of both user and company.
- Last but certainly not least, there is perspective-drive UX strategy. Here, the underlying narrative/perspective of the users on the site should drive UX strategy. For example, examining user personas recently to get the unifying “hook” behind a software app, I realized that while the users themselves were very different in many key ways, they were all concerned with the same ultimate goal. It’s actually not in the app itself, but putting that goal first in my design immediately became the underlying theme/narrative behind all my UX choices. If a design choice doesn’t further that goal, it’s probably the wrong choice, and it’s out.
These are some of the many aspects of UX strategy I’ve used, and I’m sure I’m not the only one. While which ones work for you depend on your role, good UX designers should probably consider all of them as much as possible in our work.
Thanks for the great conversation-starter.
Originally posted on UXtraordinary.com.