Saturday, October 10, 2015

The Good, the Bad, and the Ugly: A Language of Critique for Information Architecture

A dried maple leaf on a distressed background with lines of barbed wire in the foreground
At the 2014 Language of Critique roundtable, Marsha Haverty said “If we don’t have a way to describe what we do, we’ll be limited to… being wireframe monkeys.”

IA is more than wireframes. But we’re confined by the mindset that thinks IA is a box to check off on a project plan. If you find this a problem, you’ll want a way to change the discourse.

A language of critique is going to help you become a better, more influential UX professional.  We can all use that.

Furthermore, maybe you’re at a stage in your career where you’re elevating the practice. Are you teaching, researching, or publishing? Then you might be interested in the project to shape a language of critique for IA.  In order for us to develop IA as a discipline, we need a framework for evaluating the goodness of information architecture, both as a whole, and in specific cases.

So what’s a language of critique for IA?  And what’s wrong with whatever we’re using now?

 Introduction to the M3 model

We can talk about our work as taking place on three levels. 
  • The bottom level is the level of applied work.
  • The middle level concerns theories and models.
  • The top level is the level of paradigms. 

This is called the Meta-Modeling Methodology or M3. It was developed by John Van Gigch, an organizational theorist, in 1991 as a way to look at how scientific disciplines ideally operate back and forth across different levels of inquiry.  In the book Reframing Information Architecture, Lacerda and Lima-Marques proposed using the M3 model as a means to develop IA as a discipline.

Reframe IA history

The work of Lacerda, Lima-Marques, and many others is part of an effort now underway to bridge information architecture research and practice.

Think about psychology, literature, physics, economics. All mature fields have a framework and a vocabulary for making sense of what they do.  IA emerged in the 90s as a practice, along with the emergence of the web. But we still lack a foundational framework that gets taught, that is continually developed and refined, and against which our work is measured. We lack a shared understanding of what we do.

A seminal moment for IA occurred six years ago in Memphis, when Jesse James Garrett delivered a scathing closing plenary at the IA Summit. Jesse challenged the IA community to move beyond being a practice, and figure out how to become a discipline. He called for us to develop a language of critique.

Keith Instone, Andrea Resmini, and other leaders in our field are engaging the community right now to shape the future of IA.  This effort has brought academics and practitioners together in workshops and roundtables. It’s developed the book Reframing Information Architecture, published last year.  And it’s now engaging the broader UX community in an ongoing Reframe IA conversation. 

A practical example

To get our heads around the M3 model, and see how it can help our work and our practice, let’s look at a few examples.

I was approached by a client at work to answer the question: does the website have a good information architecture? This is the kind of thing we’re all asked to do on a regular basis – design a good IA, or evaluate an IA for goodness to make a product better. This is work at the applied level.

It so happens that Smokefree is a cross channel program. It includes websites for various audiences, a text messaging program in different languages, a handful of native apps, and it uses several social media platforms.  It’s a complex, maturing program that’s designed to help people quit smoking.

I suggested a heuristic evaluation.  Where would heuristic evaluation be on this model? It’s a research-based tool. So it lives at the level of theories and methods.

We tend to use Nielsen’s heuristics for user interface design, or something adapted from that set.  But as I observed to the client: you’re not operating under the old web-interface-only world.  Your program works through multiple channels. You’re acting within a new paradigm. You’re aiming to deliver a connected experience to your users. So we really should be evaluating how well the IAs work across channels.

There was cautious enthusiasm for this at first. It can be uncomfortable to think or to commit resources outside an old, well-known paradigm.  But a cross channel evaluation really did fit in with delivering a connected experience, and so they agreed.

Interestingly, when we went looking for heuristics that would let us assess IA across channels, we couldn’t really find them. So we began developing a set on our own.  And a member of my team, Dilini Abeywarna, presented “Cross Channel Heuristics” at Mobile UX last month, contributing to the theories and methods layer.

A comparative example

Does this make your brain hurt? Let’s apply the model to a different discipline to see if we can get some insights.  Consider the field of comedy.

The SOLUTIONS LEVEL. Here we think about the devices and conventions used to create comedy. They include techniques – like sight gags, repetition, hyperbole, slapstick.

We can think about comedy at the THEORY LEVEL.  What makes something funny? Some people favor:
·      The Superiority Theory. “Here Comes Honey Boo Boo” is funny.  Why? Because ridicule, and feeling superior to others, is one of humor’s primary uses.
·      There’s the Benign Violation Theory.  We laugh when some line is crossed, but the outcome isn’t really threatening. According to this theory, tickling make us laugh because it seems like an attack, but it’s actually harmless. Jerry Seinfeld is funny because he points out the outragous in everyday life, which is benign.
·      The Incongruity-Resolution Theory.  A man at the dinnertable dipped his hands in the mayonnaise and then ran them through his hair. When his neighbor looked astonished, the man apologized: “I'm so sorry. I thought it was spinach.”

And there’s the PARADIGM LEVEL. Here we address questions that explore comedy as a field of study. These include:
·      Why do only humans seem to have humor?
·      Why does timing matter?
·      What are the necessary and sufficient conditions for a thing to be funny?

Well, a cognitive scientist, a philosopher, and a psychologist walk into a bar to discuss a grand unified theory of humor.  These examples I’ve shared come from their book, Inside Jokes: Using Humor to Reverse-Engineer the Mind.

More about M3

Back to user experience.  We can talk about our work as taking place at different levels. The problem is, we don’t do that. We tend to collapse all questions of goodness into one packed layer, mostly the bottom one. 

A client is unhappy with your design. But why? Is it poorly executed according to current best practices?  Does it fail to take into account an important dynamic – say the growing importance of the 50-something market, or people’s expectations that things will work smoothly across mobile and desktop?

Or conversely, is your design based on your firm grasp of these shifts, which she’s not aware of?  It’s important to be able to talk clearly and constructively about these different layers.  So let’s parse this out a little further.

The bottom level, the level of applied work. Anything that involves designing or evaluating an artifact, such as a website or a feature, is an example of this. You start with a practical problem. You need… what?  A search results page. A series of input screens. A campaign minisite. You need whatever the user story, or project plan, or your discovery process says you need. The output of work at this level is a solution to a practical problem.

Now, how do you know how to do this work? How do you evaluate whether you’ve done a good job? You have your practices.  And your practices are informed by your science, which comes from the next layer up.

Level 2, the level of theories and models. It includes concepts that help us understand the problem space. Some examples: Designing for accessibility. Content strategy. Theories of behavior change. Embodied cognition. Responsive design.  At this level we also see research that investigates whether specific practices work. The HCI literature falls here, and publications from the Nielsen/Norman group. Is your client hesitant about parallax scrolling? Or is she asking whether we need to banish the hamburger menu?  This is where you say: Let me show you the science!

Theories and models inform how we carry out the practice.  Conversely, the practice tells us what research is needed, and helps new concepts emerge.  But there’s more. Theories and models change – or need to change – based on what’s happening on the next layer up. When our world shifts, and you’re using methods from the wrong paradigm, you’re going to run into trouble.

Here’s an example of a challenge at the theories and models level.

On the Smokefree project, the team wanted to know if specific information architectures can be used to support people trying to change their behaviors. It seems to me the answer would be yes; there’s an IA for that. Our work enables and persuades people to do things. Like renew, contribute, buy, share.  We design the IAs for apps and devices to help you lose weight, get exercise, become better at saving money. There’s going to be research on IA and behavior change, right?

We looked for empirical research on the effects of different information architectures on health outcomes, behavioral outcomes, or website engagement.

Out of an initial 688 candidate papers, only 1 both looked at IA and controlled solely for IA. So only one paper could perhaps contribute some light on the question of whether specific IAs support people trying to change their behaviors.

We know by experience and annecdotal evidence that some IAs are better for informing and persuading people than others. But we can't prove it, because there's a gap in our science. The language for asking the right questions is missing. Researchers aren't looking to fill that gap. And they don't know that they're not looking.

We are going to get traction on this research. It’s one of my quests this year.

On  to Level 3, the level of paradigms.  This is philosophy.  It’s how we CAN know what we know.  Paradigms are like containers, articulating the shape of what they hold. Understanding the paradigm gives great power to your endeavors. It helps you see the possibilities, and lets you bring meaningful things to life.  Being oblivious to the prevailing paradigm will lead to your work being ineffective and irrelevant.  Which is tragic.

We can think about paradigms by looking at what changed and what’s now possible, and then giving a name to all that.  For instance, Uber exists in a paradigm we can call “the Sharing Culture.” That paradigm changes how we think – and how we CAN think – about transportation, as well as commerce and other domains.

The 90’s saw the birth of an epoch-making paradigm called the World Wide Web, where digital pages were connected with other digital pages.  This has shifted to a paradigm called “the Internet of Things,” where the objects I use and the very life I lead are intertwined.  Things and information are connected to one another and are tied to my goals, my behaviors, and my location in space and time. 

The increasing focus on customer experience in the government sector is a result of the widening socialization of the Internet of Things paradigm. It’s also a result of another, related paradigm shift.

I don’t know what it’s called, I haven’t heard a good name for it yet. Digimodernism. Selfie Realization. Here, the individual is at center stage.

In the old paradigm, culture is a spectacle before which we sit.  The author, the creator is primary, while we, the audience, watch and listen.

In the new paradigm, the spectacle does not and cannot exist unless the individual intervenes. There is no Facebook unless we all write stupid stuff in it.  Anyone who reads a post can, and is expected to, become a co-author. The interesting stuff is in the comment section.

Look at what we’re doing right now. You have a presenter standing up on a slightly elevated stage, an audience seated in a neat pattern below, spoken words, pictures on a glowing screen, and earlier some music. On the one hand, it looks like theater in a classic sense. But everyone here can co-create this act. The interesting stuff in this performance, I think, is in the twitter stream of reactions.

We’re operating in a new paradigm of individual engagement and co-authoring. This is why, if we critique Facebook for being banal, we’re missing the point. If we judge it against a beautiful, controlled website from the 90s, we’re not seeing the potential. Also, we’re not going to understand where we can actually improve things.

To be effective, we have to understand the paradigm we’re working in. We need good theories and methods to work with. And then we can do that thing – make improvements or create something new.

Join the Reframe IA discussion

If you’d like to continue the conversation on developing a language of critique, you’re invited to join the Reframe IA LinkedIn group. Next steps for the group include publishing the results of the roundtable this year and planning a workshop and other activities for the 2016 IA Summit. Join at


Hurley, Matthew, Daniel Dennett, and Reginald Adams Jr., InsideJokes: Using Humor to Reverse-Engineer the Mind, MIT Press, 2011

IA Summit, 2013, Reframing IA Roundtable

IA Summit, 2014, Teaching IA Roundtable

Lacerda, Flavia and Mamede Lima-Marques, Information Architecture as a Discipline — A Methodological Approach, in Reframing Information Architecture, Springer International Publishing, 2014

Malone, Erin, “Defining a language of critique,” 2009

Resmini, Andrea, Reframing Information Architecture, ibid

van Gigch, John P., System Design Modeling and Metamodeling, Plenom Press, 1991

No comments:

Post a Comment