How to persuade people of the business value of usability testing has long been a topic of conversation at any HCI conference. As a field, we have made a great deal of progress in defining the clear business case for usability. But in this article I want to consider the parallel question of how we can communicate the business case for ethnographic design researchethnography, contextual observation, and field studiesto a skeptical audience. Ethnography faces some challenges beyond those of usability testing, partly because usability usually occurs later and has a more immediate connection to product quality. Although people are much more familiar with the concept of ethnography than a few years ago, there are still plenty of questions and quite a few skeptics.
This article will consider these questions briefly and then look at the relationship between design research and market research. This will lead to the argument that design research and market research are based on different, but complementary, world views. Persuading people of the value of ethnography requires that we help them understand both that there are two different perspectives and that these perspectives are complementary. Only then can we really start to address the business case for ethnographic research.
Many practitioners of ethnography have run into the resistance or skepticism I am thinking of. There are some common questions that I am asked by people skeptical of the value of ethnography. Perhaps, once you have introduced the notion of doing ethnographic work, you have also heard questions like these:
- How is this different from what our market researchers are already doing?
- How do you do your sampling to be sure that the people are both representative and interesting?
- How can you possibly learn important and valid information from such a small sample?
- We've been in this business for twenty years, and our field reps have very close relationships with our customers. How could you find out anything we don't already know?
One of the first things one learns is that trying to answer these questions directly can quickly get very messy. After all, you don't know what their market researchers are doing and ethnography is very popular in market research too. I have also found that it is dangerous to argue that ethnography uncovers different information from market research, since that can set up a conflict between the two research methods. Rather than saying anything that could sound like challenging the legitimacy of market research, a better direction to go is to emphasize that the goals of market research and design research are different.
Furthermore, the moment one tries to give examples of what one might learn, one discovers that people start reacting as if the example were meant to illustrate a general truth and it is suddenly even harder to counter the belief that ethnography is little more than armchair theorizing. Even worse, any examples sound somewhat predictable in hindsight, which is not very persuasive. On the other hand, a surprising example challenges one's credibility, unless one can persuasively argue for its validityand one hasn't done any research yet! Thus, the arguments easily become circular.
In the end there is a tendency to emphasize the general benefits that a client will receive from ethnographynamely a design process grounded in the realities of people's lives, not in stereotypes. Jeffrey Veen, writing for New Architect (www.webtechniques.com/archives/2001/06/veen/), says that:
Ultimately, you want your design team to talk about the things you hear from your users, rather than what team members think the users want. It's a significant difference. I'd suggest banning phrases like, "Don't you think users will..." from your working sessions. Much more valuable are phrases like, "Well, when I was following Susan around, she would often...
These sentiments are absolutely true for those of us who are already believers in ethnography, but unfortunately they are not very effective for persuading the skeptic.
There are some good case studies (see Juniper Financial sidebar), but in our efforts to emphasize the contribution of ethnography, we can too easily downplay the contributions of all the other disciplines and perspectives. All success stories seem to lead inevitably to a paradoxif ethnography reveals some previously unidentified latent user need, then this must by definition be also a high risk direction that requires further market research to validate, or if ethnography confirms some existing market research, then it is inevitable to wonder if the ethnography was really necessary.
In this article, I will point to some reasons why people ask the questions that they do and then I shall offer two possible techniques for helping overcome some of the resistance.
Of course, in order to persuade, we have to understand the mindset of our audience. Imagine the client who has already spent $250,000 on market research interviews with 800 potential customershow do we persuade them that their data is not what designers need to really do human-centered design?
One of the biggest challenges is to get across the different roles of market research and ethnography in the design process. It can be difficult for clients to see, for example, that designers need to be inspired by a rich, textured understanding of user needs, whereas most market research provides a list of needs, without much depth or texture. Indeed, much market research is concerned with validating a list of needs and sizing the market associated with each, whereas designers need to understand how the product or service is going to fit into someone's life. Thus, even when similar research methods are used, it is often the case that a report generated from market research is insufficient for design.
This reveals the danger in allowing the discussion of ethnographic research to become a territorial dispute with market research. The critical point is that the two kinds of research are studying the same thing (potential users) in different ways for different purposes, guided by completely different philosophies. Market research is primarily concerned with making business decisions and forecasting sales and quantifying business models. Design research is concerned with enabling design decisions that are rooted in a true understanding of the needs of users rather than in someone's intuitions about what users might need, or in averaged user ratings of desirability of features.
Most skeptics are concerned with typicality and representativeness and have difficulty understanding why good ethnography seeks out extreme and idiosyncratic users. They need to be educated about the importance of identifying boundary cases that the design should strive to handle. I find myself addressing this issue by sharing a few stories like the one of Mary and Don (see sidebar), after which I am often asked how we find people like them and whether I think we'll find some in the current project.
These questions reveal a common misunderstanding: namely that the average person is truly average. Ethnographic stories and observations are clearly not based on representative samples, since there is nothing average or typical about them. But the truth is that there is no average person out there. The answer I usually give to these two questions is that we always find people like Mary and Don. Everyone is interesting and can provide design inspiration if you know how to get to know them.
A colleague of mine was once asked how many people we had studied. Six was the answer and the response from the questioner was that they normally never did research with fewer than 100 respondents. My colleague tentatively continued and presented her research and conclusions and the response that came from the V.P. of Product Development was, "She obviously found the right six people!" This response made her feel a lot better about the meeting, but it is actually as misguided as the original question was. When doing any kind of user research you can study large numbers shallowly or small numbers in depth (which method you need depends upon your questions and the kinds of answers you are looking for).
In essence, there is a philosophical difference between research conducted to inspire design and that conducted to validate designone is about idiosyncracies and little details and the other is about averages and generalities. Overcoming the resistance to ethnographic design research is not about rational arguments about business goals (at least initially); rather it is about enabling someone to see that there is a different perspective and that this approach is not in opposition to market research approaches.
Generally speaking, all the above suggest good rational arguments that make sense and help move the conversation along, but they tend not to persuade. Ultimately the persuasive moment comes when a person feels the rich texture of life present in a piece of ethnographic researcheither through having an experience or through hearing a story about a real person.
Interestingly, this is actually not too different from the world of traditional usability, where the most persuasive technique is to have the client watch the user testing as it happens, or, failing that, watching videos of real users struggling with their system.
What follows is a brief description of two techniques that can quickly and easily give people the experience of an ethnographic approach to human-centered design. They help people come to the realization that there is a different perspective and that it is not only not in opposition to their traditional approaches, but adds to them.
Even when someone has worked in an industry for many years, they may not have paid close attention to their own experiences of their products. In particular they may not notice the detailed context in which the interactions occur. I have found it helpful to focus their attention on their own experiences. This method always feels risky and yet it proves very successful. For example, in talking to a company about a Web site which has downloading printer drivers as one function of its Web site, one might ask people present to remember the last time they needed to download printer drivers and to tell that storythe story-telling is perhaps the key part. Self-observation demands that we focus on a real occasion when we used the product or service and not on how we imagine we might use it.
Using myself as an example, if I recall when I last needed printer drivers, it was just two weeks ago when I was printing a report for a client and the printer kept insisting it had no paper in it even when I had just filled it. When I got exactly the same error on a different printer, I was initially puzzled and then I realized that neither printer showed any paper trays connected. At that moment I realized that it was probably my problem rather than a printer problem, and so I threw away my existing driver and went to the relevant Web site to find new ones. And then the printer worked and the document was printed just five minutes before the client arrived.
It is easy to imagine another person in the room telling a story like Mirka's (see sidebar on page 33) where they had someone trying to navigate them through the Web site over the telephone. With only one or two such stories it is possible to then enter into a discussion about whether it seems such a good idea to place pop-ads and other material on the route from the home page to the printer drivers. These might be the most commonly traveled pages in the Web site, but they are also the ones when the user would appear to be least in the mood or frame of mind for such distractions.
The purpose of self-observation is not to answer questions, but to raise them. By giving the participants a new perspective on behaviors that they thought they understood, this technique can help make them much more receptive to the idea that they should sponsor some research to look more closely at real customers and users.
When time is too limited for the above technique, nothing seems to be more persuasive than sharing simple stories of real people in real contexts. Having a good collection of real people "up your sleeve" is always valuable since one can always share a story that is relevant to the current conversation. Sometimes, the story should point to how particular ethnographic findings had major impacts on design. But, it is also helpful when you can share a simple story and have them discover for themselves the implications for their business and for design. These stories enable an experience that makes it obvious that such insights are not likely to be found without ethnographic design research. Of course, for reasons of confidentiality, these stories often have to be somewhat "genericized," but without losing the nuggets that make them so real.
One powerful story for helping others to discover their faulty assumptions is that of Ilkka (see sidebar). In a recent Internet project, the client was proceeding on the fundamental assumptions that people in Europe didn't shop on the Internet, or use the Internet for as much time as people in the USA. And they also knew that SMS was very common, but weren't too sure what that was or what it meant, but they knew that everyone was talking about it. In a business development meeting with them, I told them Ilkka's story. From this, they saw that the need being met by SMS in Europe was already being met by email in the USA and that the Internet played a different role in European lives. They were able to suddenly see how they had misinterpreted the survey-based data without knowledge of the real-life European perspective.
Sharing stories of real people using real products in real contexts can be very effective in helping people realize how much they need richly textured, individualized information about their own customers, to supplement the averaged data they are used to getting from market research. Having them discover for themselves the implications of these stories for their own business is much more persuasive than stories about the impact of ethnography on someone else.
Inevitably, as people begin to become persuaded, the question arises, "How do you turn these stories about people into great design?"
This is a very understandable and legitimate question. After all, the reason for conducting the research in the first place is to improve design. The tie-in between ethnography and product concept can be relatively clear once initial skepticism is addressed. But the tie in with concrete design can indeed seem like a mystical process. While this topic fully deserves its own article, some of the techniques are quite well known to this audience. The use of scenarios and personas is central and the critical element is ensuring that the important characteristics of the real people visited are reflected in the scenarios and personas written. If scenarios are generated using only those aspects of the real world that are favorable to the product concept, then the value of the ethnographic design research is dramatically limited. If the composite and fictional personas are drawn as stereotypes, then the research might as well not have been conducted.
But the full details on how to achieve this link into design will have to wait for a second article.
Human Factors Design and Research
100 Forest Ave
Palo Alto, CA 94301
Business Column Editors
Susan Dray & David A. Siegel
Dray & Associates, Inc.
2007 Kenwood Parkway
Minneapolis, MN 55405, USA
One good success story is Juniper Financial, an Internet bank which used ethnographic design research to understand its potential customers and which designed not only its Web site, but its whole offering around the needs of one particular psychologically-defined group. Ethnographic research led them to take an approach in designing their customer experience that set them apart from other banks. Our research led us to identify a group called "Onlookers" who had a short-term horizon and low engagement with their finances. These people appeared to be the most closely aligned with Juniper's values and the most likely adopters of Juniper's proposed services. Also, Onlookers' low daily engagement with their money and lack of clear long-term goals made them potentially ideal, loyal bank customers. Based on this research, unlike almost every other bank, Juniper arranged their complete user experience for people in this "Onlooker" category. Thus the outcome of ethnographic research helped to focus Juniper's brand value and provided a framework in which to map the competitive landscape, so that Juniper could differentiate itself in this market. The research not only enabled Juniper to internalize an understanding of some fundamental financial user types and their related behaviors and motivations, it also redirected Juniper's business strategy and service proposition.
Mary and Don live in the mountains of Eastern Oregon. They are, in their own words, "true conservatives," if not survivalists. They live in a rough and ready home at the end of a dirt road and it seemed unlikely that they would be highly sophisticated users of the Internetmy reason for being there. I knew that they do use it, but it became quickly apparent that online shopping is not likelythey would never give out a credit card number online. In fact, they didn't even have a credit card, because that would require them to give someone their social security numbers. And anyway, courier companies do not deliver to the end of their road, so they would have to go 20 miles into town to collect something. However, it turned out that they are so remote that they cannot receive any TV or radio signals and their only connection to the outside world is their telephone line and so they are true Internet enthusiasts. They have two phone lines and one stays connected most of the day. They use the Internet for news and weather, as well as for chat radio and lots of music activity. They play and record their own music and then share it with others. They have little chance of obtaining broadband in the near future, but they are totally committed to their 56k.
Mirka lives in Turku, Finland and she runs her own small business with just three people: Mirka, a colleague and an administrative assistant. They each have a computer, of courselaptops, in fact. And the laptops never go home with them; they are locked into the filing cabinet each night. Mirka bought laptops because her offices were broken into twice in five weeks and her desktop computers stolen each time. When she bought the laptops she needed to install the drivers for her inkjet and laserprinters. She didn't know where the CD was and so she called customer service who gave her a long list of links to follow on their Web site. Mirka wrote them all down and tried, but failed (well, she thinks she failed, since she cannot find anything that looks like a driver on her computer). So she called again and gave a new name and changed her address. She was embarrassed in case the person at customer support recognized her and realized that her previous advice was not helpful. Still they cannot help and they want to charge her $15 to mail her a CD. As a last resort she went down to her basement and hunted until she found the boxes containing the original CD. And she still doesn't really know what drivers are, anyway!
Ilkka lives in Stockholm and runs his own one-man accounting business. He has a DSL connection, but it is metered and he takes considerable care to only use it for essentials. He checks his email two to three times per day and only uses the internet for very specific pieces of information, since it costs him money to do anything else. He was astonished when I suggested e-mail as a means of contacting his clientshe always uses cell-phone based short messaging (SMS) to send a client a message, especially if it is to tell them there is paperwork to be signed and mailed today. He knows that most of his clients also pay per megabyte on their Internet connections and that they are also rarely at their desks. They might not receive an email message for 24-48 hours, whereas he can completely rely on the SMS system to get the message to them right away, wherever they are. And having grown into the habits of SMS, he can't imagine using e-mail even if it were cheaper.
©2002 ACM 1072-5220/02/0300 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2002 ACM, Inc.