It's Like Summer Camp for User Experience Types! | Main | I was in a car accident

July 28, 2005

More from "The Don" Norman -- Activity-Based Design

The hottest thread on the SIGIA-L mailing list concerns Don Norman's recent essay, "Human-Centered Design Considered Harmful." In it, he advocates shifting away from human-centered design (what most of us call "user-centered design") towards activity-centered design.

His central point is not to focus on "the user," but on "the activity." User-centered design emphasizes understanding the user, their context, their circumstances, and what they're trying to get done. This is usually achieved through some form of field research, such as contextual inquiry. However, such field research only gives us a snapshot of a person at a particular point in time. If we were to come back a week later, the circumstances could be wildly different. Also, it only gives us a snapshot of *that* person... And users are nebulous, various, with a host of different strategies and approaches. Designing to serve "the user" often leads to lowest-common denominator design.

Don has a tendency to "discover" things that are not new, and then proclaim, "Hey! Look what I found!" and talk about it as if it were the best thing ever. He recently did this with the subject of Emotional Design (where designers and researchers have been aware for a long time, possibly forever, that appreciating emotional concerns in design is crucial), and he's doing it with his refutation of user-centered design. 9 months ago I wrote a post, "Pity the Poor User," inspired by Clay Spinuzzi's book Tracing Genres Through Organizations, where Clay calls into question the user-as-victim trope, and the fact that UCD tends towards a one-size-fits-all approach.

Don posits activity-based design as a potentially superior approach -- focus less on the people (who are many, various, and changing) and focus on the activity (which is comparatively stable). Also, a focus on the activity requires taking greater care to understand the capabilities of the technology. Though members of SIGIA-L seem too dense to comprehend it, Don readily acknowledges that the methods in activity-centered design are pretty much the same as in UCD; the difference is in the mindset, the approach. Instead of putting The User at the center of discussion, you put The Activity. You then recognize that the user is simply a component of that.

I'm disappointed that Don mentions "activity theory" in only a throwaway sentence, with no citations -- "The hierarchical structure comes from my own brand of "activity theory," heavily motivated by early Russian and Scandinavian research." Interested parties have no clue as to where to turn. I can't claim a full understanding, but Activity Theory has emerged as perhaps the second most important theoretical platform in human-computer interaction research, after cognitive psychology. A couple years ago I read the work of Yrjö Engeström, who has done a lot to make activity theory actionable. Unfortunately, the best paper of his that I read is locked up behind a publisher's wall...

So, while Don isn't doing right by his colleagues, and has a bit of a habit of taking credit for ideas that have been around, he deserves credit for shedding light on an overlooked approach. And for continuing to cast doubt on what has become the rhetorical dogma of user-centered design.

I'm surprised Anne Galloway hasn't chimed in yet (maybe she hasn't seen the hub-bub.) A year ago we found ourselves in a large group discussion, where she about her dissatisfaction with designers being obsessed with "types" of users, because an individual can be many different things, and want many different things, depending on their particular contexts. She instead advocated a notion of designing for practice -- what people are doing, not who they are.

At the time, I considered this a radical shift in how designers approach people. Judging from the response to Don's article, it remains so.

Posted by peterme at July 28, 2005 09:22 PM


So, my first question is how is Don's "Activity-based design" any different than Cooper's "Goal-based design"? It seems to me that activity is a sub-set of the users' goal.

Since Don is high on the hierarchy in his essay, it seems to me that it should be: Goal>Activity>Tasks.

For instance, I want to see how my stocks are performing today. That's my goal. While designing for that goal, I'm going to design for specific tasks (activities) that the user performs in order to accomplish that goal.

There are several areas of Don's essay that I don't agree with.

To the Human-Centered Design community, the tool should be invisible, it should not get in the way. With Activity-Centered Design, the tool is the way.

I don't think that the mantra for HCD (UCD) is that the tool should be invisible. I think it's that "we" shouldn't get in the way. In other words, we should be enabling users, not disabling them through the tools we provide them to do whatever it is they need to do.

Another thing he saysl, which I think is incorrect:

One concern is that the focus upon individual people (or groups) might improve things for them at the cost of making it worse for others. The more something is tailored for the particular likes, dislikes, skills, and needs of a particular target population, the less likely it will be appropriate for others[...]

Well, yeah, if it's done poorly. The fact is a good design takes into account that there are different audiences (typically). And you design for those audiences. So, there's no reason why a proper UCD process can't produce a product that grows with the user. In fact, that's what it should be doing.

Don't blame the process, Don, blame the execution of the process.

Posted by: Todd Warfel [TypeKey Profile Page] at July 29, 2005 07:30 AM

I started a comment here, but it turned into a post of its own.

Posted by: heyotwell [TypeKey Profile Page] at July 29, 2005 09:17 AM

I think it always comes down to recognizing that things fluctuate over time. Both people and the tasks they perform. To design based only on the data you get at a certain point and time is pretty silly. It is a good starting point for thought and direction, but not the "whole picture". When working on a large telecom site I focused on both user "type" and "state". They are stereotypes, but they are useful as a base to design upon and to act as a check and balance. But it is not a simple box, it is fluid. People are not constants and so identifying "states" helps in that regard.

Focusing solely on activities or solely on users is simply too narrow. Look at both. Try both approaches. If you don't, you're very likely to miss the bigger picture. If you focus only on the activities you miss how the person's personal state effects efficiency. If someone is in a state of distress the system's response should change accordingly.

Posted by: sloanNYC [TypeKey Profile Page] at July 29, 2005 12:44 PM

and the next step (after user / activity) would be... the context with an emphasis on both the artefacts (Hutchins) or the environment (Suchman)... as if something new has been discovered. At last, it's going to be new for some designers.

Posted by: nicolas Nova [TypeKey Profile Page] at July 29, 2005 09:12 PM

Seems to me that the 'context of use' is key in and this to some extent is based on analyses of 'the user' as well as the activity itself. The former is broader in its understanding of context of use and can have certain benefits when, for instance, looking at editorial tone and 'usability', though as you mention creating one-size-fits-all experiences is a major failing of UCD. Either approach fails if it is moves toward a behaviourialist or functionalist approach where the user or the activity is 'reified'.

Tension undoutedly exists in the *attitude* of design practice if, as Don Norman states, the designer using AT 'leads' and rarely listens to users whereas the opposite is true of the UCD designer. However, from my experience it is invariably more complex that that simple polarisation suggests. Apple, for instance 'listen' to their users a lot even it is in terms of their activity - look at the application of the 'shuffle'!

Interestingly, I've also found in marketing there is a parallel move toward developing CRM based activity and new products based on *frequency* of use rather than traditional demographic or attitudinal 'segmenation'.

Posted by: jimmyb [TypeKey Profile Page] at August 2, 2005 05:54 AM

I appreciate where y'all are going in this...and if the audience for this is the HCI/UX/Whatever community, it's worthwhile.

The danger is when this jargon escapes our bailiwick and enters the world our clients, our managers, our customers inhabit. I'm afraid they'll hear, "So, just worry about the tool & wht it does, right? So we just center development on the application we're building? So we just cut out worrying about people and just worry about technology...? Greaaaat. We'll go back 20 years and do tech-centered development. That way, we're focusing on the tool."

I'm not saying that's what you and Don ARE saying, I'm just afraid that's what customers will hear. So from my viewpoint, I still feel UX makes sense.

Posted by: jsokohl [TypeKey Profile Page] at August 2, 2005 05:55 AM

Re-reading Don's essay, I just had to support something Peter sez: not only does Don seem to skim over references, he seems to forget his onw writing. In his section "Tools Define the Activity: People Really Do Adapt to Technology," I hoped he would have then admitted being wrong. In "Things That Make Us Smart" he writes, "'People Propose, Science Studies, Technology Conforms'
My person-centered motto for the 21st century." I guess he was wrong. Have I been wrong to read and be inspired by him? Will another 10 years find that ACD is just so much B.S.?

Posted by: jsokohl [TypeKey Profile Page] at August 2, 2005 07:04 AM

Every guru his own label...
activity-centered design, goal oriented design, human centered design, user-centered design, humane design, listening labs, ajax, web2.0 ...
Mostly saying the same, just different nuances and labels. One of the first things I learned was to study the user AND his tasks AND his tools and environment. I forget what label was put on this methodology and I don't feel bad about that...

Posted by: pantomimecow [TypeKey Profile Page] at August 3, 2005 02:48 AM