Is ethnographic research worth it?
The design blog, Core77, has published an interview with Don Norman, author of a series of books on user-centred design (the best known of which, The Design of Everyday Things, was the first, and possibly only, book on user experience to become a bestseller).
Norman always makes interesting listening. I loved his recommendation to his interviewer, Bruce Tharp, that he should ‘question authority, even if that authority is me’ so was frustrated that Bruce didn’t go off-script a little more to pick Norman up on some of what he said. So I’ll do so here on his behalf.
The ‘question authority’ comment came about a third-way through the interview where Norman criticises the techniques used by companies who are committed to user-centred design. His comments focus particularly on ethnographic research and the creation of personas. Norman complains that, interesting as these processes are, they fail to connect to the engineers and designers creating products and services and so are, in effect, a waste of time. What Norman wants, he says, is processes that focus on the task, rather than on people. I know what he means, but worry that he’s throwing the baby out with the bathwater.
So let’s take a short step back from Norman’s comments and remember what preceded the typical user-centred processes many companies now use. Often, nothing. Or, if you were lucky, focus group research on a product that was in the final stages of development, when it was too late to have any impact on its fundamental design. Products were shaped by engineering dictat and the styling of the designer (think back to VCR controls of the 1980s). As Norman mentions, a typical company reaction to problems users encountered with their products was that the users were stupid. It’s certainly true that people feel stupid when dealing with products they can’t work. And that has implications for the company selling the product. Would you buy again from a company that had made you feel stupid? Probably not.
So how do companies take that step from thinking ‘right product, wrong customer’ to thinking that their customers have a right to products and services that meet their needs and are genuinely easy to use? I think it’s through empathy with the user (so does Norman). And how do you get that empathy? By meeting your users and seeing how they use products and services and by hearing what their concerns and preoccupations are; that is, through observing people in context, precisely the ethnographic research that Norman dismisses. Then you need to find ways to keep those end-users in mind so that, even when you’re not talking with them, something of their attitudes and concerns remains with you while you design. Not an easy task, but one in which the personas Norman writes off can be a useful tool.
These research tools rarely answer detailed design questions. That’s what iterative prototyping and user feedback cycles do. But they help focus design thinking on what people want, they support understanding of gaps in the market and decisions about what to design, and sometimes, yes, give insights into design detail too.
What I think Norman is criticising is the unthinking application of research techniques, and the failure to connect them to design process. That isn’t a problem with the techniques themselves, but inflexibility in some organisations in how they are used. There was a similar rigidity about 15 years ago when many companies established ‘usability labs’ to test products during development. Over-scientised in their approach, over-staffed and unwieldy, they couldn’t provide nimble enough input to be useful. In most cases they were abandoned in favour of a leaner approach, often involving designers directly in testing their own products.
As with usability testing you need to understand ‘ethnographic research’ as a portfolio of techniques to get close to users; for example, shadowing them as they go about their daily tasks, talking with them in context, asking them to create diaries, to take part in workshops with designers, finding out what they’re saying in their blogs or on forums (no, this isn’t ethnography, in its academic sense, which is why I prefer to call these techniques observational research). There’s no single true way. And since many real-life interactions are private and it’s hard to predict when or where’ll they’ll take place, finding a good way to research them requires some creativity.
Norman comments that the best design is often found in gardening, farm and workshop tools. But these have evolved over scores, sometimes hundreds, of years into the familiar implements used today. Not so for new technology tools, where there are no long-standing traditions for what they do and how they do it, where there often isn’t a readable relationship between form and function, and where, as I’ve commented before, the people designing the tools may not be those who will use them. That’s where understanding the people as well as the task is so important.
A rhetorical (I assume) question in a review by Alan Cane of improvements to text entry for mobile phones (FT, 24 November) highlights the importance of understanding both people and task. Cane questions whether developing better text entry techniques than we have presently is worthwhile: ‘…surely inputting text through keys is simply a passing phase before voice input takes over?’. I suspect not, and think Cane would be convinced too if he were to observe the detail of how text is being used. Voice input may take over some tasks, or even facilitate new forms of messaging, but text is so embedded in people’s lives, I’ve no doubt it’s here to stay. Take a look at the many contexts where text is used because talking wouldn’t be practical or acceptable; look how text allows editing in a way voice input might not. Text is not just a cheap substitute for talking, it’s a channel with a role of its own. By looking at the task alone, you might miss that. But you’ll know that it’s worth the effort of developing more usable text entry systems, if you look at the people as well as the task.
1 comment November 29th, 2007