Sam Ladner: Applied Ethnography and The Emic Perspective
Sam Ladner is a sociologist and one of the leading voices among applied ethnographers. She wrote the go-to handbook on the subject, Practical Ethnography: A Guide to Doing Ethnography in the Private Sector. More recently, she wrote Mixed Methods: A Short Guide to Applied Mixed Methods Research. She was the first Senior Principal Researcher at Workday and, before that, the first Principal Researcher at Amazon. In this conversation, she discusses her transition from academia to the private sector, the emic vs. etic approach, and the “luxury of insight” in the age of AI.
Interview by Jonah Ginsburg, Director of DesignLabs.
Words from the interviewer are in bold italics.
JG: I’m curious about your academic training and how it led you to applied ethnography.
SL: I was originally a journalist. I practiced interviewing and writing up observations in a journalistic way for years. I was a technology journalist, so I studied how people use technology, but the deep ethnographic part of it, which I did not even have a word for at that point, escaped my view. When I arrived at grad school for sociology, I figured my research training was going to be mostly quantitative, which it was, but I also took qualitative methods. By the time I finished my PhD, I had more qualitative experience than most of my classmates, partly because of that journalism background.
When I went into the private sector, it was good timing because people were starting to realize that technology was being built without much forethought or user centricity. I was able to bring my qualitative and quantitative skills to help people build technology that worked in both a business sense and a human-centric sense.
The biggest issue I found in the transition was that many people in the private sector were completely untrained in qualitative methods. They thought of research as rigorous only if it was quantitative, and they didn’t understand that qualitative research could actually be very rigorous and systematic, with checks and balances at every stage. Bringing qualitative methods into the private sector was difficult because good training didn’t exist at the time and people thought of it as a time sink.
JG: You then wrote a book on that transition, Practical Ethnography. Why did you write the book?
SL: I knew I wanted to teach this work. I did a couple of workshops and was looking at the literature, but there was literally nothing. There was the Ethnographic Praxis in Industry Community, or EPIC, but they mostly published case studies and papers rather than how-to manuals. I used those as examples, but I needed more. I thought I had to do it myself. I needed something that I could teach, and I also wanted to figure out what I had learned and make sense of my own journey.
JG: For people reading this who might not have any background in the field, how do you define ethnography?
SL: I usually define ethnography (from the Greek “ethno” and “grapho”) as writing about folk. Writing about what they do and how they live. I use the word “folk” on purpose because it’s an everyday, mundane description of what is actually a very complex topic, which includes customs, rituals, and the social reality that they find themselves within that is often invisible to them.
Ethnography involves not just interviews, although that’s a big part of it, but also observation. Observation is the big difference that people have to get used to. They may have heard of focus groups or light versions of research like ride-alongs, but what’s different about ethnography is the observation of everyday life. It’s about deeply understanding the social reality that people live in.
JG: Ethnography is usually done in-person and on-site. Why meet people in their own environment? And what gets lost when doing things remotely or in a lab?
SL: In journalism, there’s a saying: go, don’t phone. If you go to a person’s environment, you get more details. I used to think of it as just details, but after I got proper training, I realized that the details are about revealing everyday rituals, habits, experiences, and mental models. The objects around us say so much about why we believe what we believe and why we do what we do.
There’s also another part of it that took me longer to figure out. The contextual nature of ethnographic work in the private sector is extremely important because it reverses the power dynamic. Typically, the participant is the object you are interrogating to extract information. But if you do ethnographic work and you go to their context, they control everything. It’s their office or their home. They control the temperature, it’s in their language, it’s their place. This subtle reversal is important for people who build technology because they often have a lack of awareness of what matters to the end user. They think they can get a proxy for that by interrogation and pulling information out of users, but when you physically go to those people’s contexts, it completely changes your sense of humility and your approach.
In the book, I label this as “emic” versus “etic” perspectives. An etic position is one where the researcher defines the categories and the importance of everything. On the other hand, the emic position involves meeting the participant where they are and letting them take the lead.
JG: So a researcher with an emic approach doesn’t try to squeeze the participant into predefined categories. They leave things more open-ended and let the participant guide them.
SL: Exactly. Most people understand this from taking a survey where none of the answers fit them. The etic position tells you what categories you’re in and forces you to pick one. I remember doing ethnographic work with a physician. At the end of the day, he explained that he had done many focus groups in the past where researchers would ask him which party hat he liked. He would keep telling them that he didn’t want a party hat, he wanted an elephant. He much preferred talking to us, since we were there to understand what he wanted in the first place. He would never end up with a party hat category in an ethnographic study because he led us to the elephant.
JG: How does letting the participant lead the discussion affect the insights that ultimately come out of the study?
SL: Many people veer away from that open-endedness out of anxiety. Researchers have stakeholders who think they need to know X and Y, but really they need to know Z instead. A lot of researchers get anxious by opening things up because they worry they will not get the insight they promised. With an open-ended approach you can get the insight you planned on getting, though it may be reframed.
If you find during a participant-led interaction that your assumptions are wrong, it can be very stressful. However, it should be a wonderful opportunity to ask how you were so wrong. You can use that opportunity to gain new knowledge about something you thought was a closed case.
For example, let’s say you’re doing a study on radio listeners. You’re going to talk to people and ask, “Do you like talk radio or smooth jazz radio or contemporary hits radio?” And they say, “I don’t listen to the radio at all.” You might initially get worried that you’re in the wrong place. But then you find out the reason they don’t listen to the radio is because none of those categories make any sense to them. They’re far more eclectic in their listening and in their music choices, and they actually listen to music all the time. They just don’t listen to the “radio.” The reason your radio client is not getting this person’s business isn’t because they have too much smooth jazz and not enough contemporary hits. It’s because they have presented the product in a way that’s incompatible with that type of listener. Maybe nine out of ten participants you talked to really enjoyed listening to the radio because they didn’t like making choices. But that tenth person considers making choices core to their music listening. In order to appeal to those people, you need to give them choice first. That assumption busting is priceless. You don’t get that very often.
JG: Do you have some examples of when this kind of approach has been particularly helpful in guiding decision-making or directing a design project?
SL: During my time at Amazon I was researching an Alexa-enabled device. Leadership only wanted to do quantitative research, so I had to prove the method. I allowed the quantitative research to go ahead, I oversaw survey design and device analytics, but I convinced them to let me do some in-home observations as well.
A marquee feature emerged that everyone on the team loved and had put so much energy into. However, when we had people use the device naturalistically in their homes, we discovered the feature did not work at all. Worse, it was so incomprehensible that it made people question the entire device. It undermined the whole opportunity. I would never have gotten that insight had I not been in people’s homes with them, getting their realistic reactions. We managed to get that internally loved feature removed, which was a big deal.
Another example comes from my friends Charley Scull and Jay Hasbrouck, who wrote a paper on the sustainability of fishing. They hung out with fishermen and learned that there’s a concept called the “daily catch” which is more than just a number. It’s a mental model and a cohering metaphor that helps them understand their job. It gets in the way of sustainability because sometimes there’s no daily catch in sustainable practice. If you look at the daily catch objectively, it’s just a bunch of fish, a number. But if you do it ethnographically, you see that it weaves into everything: daily practice, ritual, employment incentives, even recruitment for the next generation of fishers. It changes how people think about sustainability, because sustainability sometimes doesn’t include catches. You’re basically asking people to reinvent everything.
JG: After the fieldwork is complete, once a researcher has their big pile of notes, transcriptions, photos, and videos, how should they begin to make sense of it all? The practice of finding patterns and themes throughout the data aligns with the emic standpoint, but in your book you also mention using existing social theories. Doesn’t that risk slotting people into predefined categories, which was one of our concerns about the etic perspective?
SL: The essential concept that you need to employ is reflexivity. I often think of Dorothy Smith’s Institutional Ethnography. She argued that the goal is not to research the people, but the institutions with which they interact. The people lead you, and then you turn your lens onto the institution to see how it enables or constrains them. If you keep in mind that your job is not to investigate people but the institutions they interact with, it’s so much easier to avoid slotting people into boxes.
Another way is through reflexivity checks. You can show a participant what you wrote about your experience with them and ask for their comment. Very rarely do I find that I have completely missed the mark, but if you give them the opportunity, they will tell you.
You can also use inter-rater reliability tests. You can code all your transcripts and then do a comparison with somebody else’s code. In the tool I use, MAXQDA, there is an AI feature that lets you chat with your coded segments. For example, I can ask if the data represents what Veblen called “conspicuous consumption.” The AI will look at the transcripts and tell me if that is a reasonable inference or point out exceptions. It acts like a research partner, but only after you have done the hard work. You cannot get AI to do everything for you.
JG: Do you see AI tools as a new way to scale up ethnographic research?
SL: I’m bored with the conversation around AI because it’s not revolutionizing my life, though it does offer an opportunity to relieve burden. Coding large transcripts has always been the bane of my existence. Once upon a time I used to write my own transcripts by hand with a foot pedal. AI came for the transcripts. I love it. AI can take that over. AI is now coming for first pass coding, I love it, fantastic, take it over.
But AI cannot come for interpretivist perspectives or emic perspectives. It can help me stay true to my emic perspective. The problem is that the opportunity to scale is not being employed correctly. Scaling should be about minimizing the burden on the human researcher and emphasizing their ability to do interpretive work. Instead, we see AI taking over everything and flattening interpretations into a laundry list of things that happened. Researchers who lack a theoretical background are the most threatened by AI because they do not know how to make that next level of interpretation.
JG: There are now tools for conducting AI-moderated interviews with real people. You wouldn’t call that ethnography, I don’t think anyone would. But at the end of your book, you talk about not being a purist and allowing for a plurality of approaches, because reality is complex and no one method is going to reveal every facet. Is AI-led interviewing a useful new tool to have in the kit?
SL: I rarely condemn any tool wholesale. Some research indicates an AI interviewer may allow participants to feel more comfortable sharing very personal histories, such as medical issues or abuse. However, it is self-defeatingly ironic when stakeholders choose to understand people through a technologically mediated method while claiming to be human-centric.
Deep understanding is not about interviewing thousands of people. AI-driven interviewing often equals zero human exposure. There is something philosophically important about seeing “the other” in a human encounter. You have no soul connection to your users if you use AI for everything. If we can demonstrate that AI interviewing is less triggering for victims of abuse, for example, we should use it. It’s like taking a painless breath sample instead of a blood sample. Of course we should use a less invasive form of interviewing. But do not be fooled into thinking you are changing your human centricity by doing so.
JG: It seems that as people start outsourcing more and more of their work to nonhuman intelligence, the need for real human-centricity becomes only more important. What role do you see ethnography playing as we move into this new technological era?
SL: It’s an opportunity to show unique value. High-touch human experiences are becoming a premium, luxury service, and the same can be said for ethnographic work. It is expensive in terms of time and travel, so ethnographers must demonstrate the “luxury of insight” and the unique value they offer.
The opportunity for ethnographers is about offering the only way to deeply understand other people, which is through visceral connection with another soul. How do you communicate the value of that in your research findings? You’re going to talk about emotion. You’re going to talk about story. One of the things I’ve been noticing a lot lately is that AI-written posts and “articles” are so meandering and boring. They’re not compelling. They’re not surprising. That’s where ethnographers can come in. Telling full, coherent stories with beginnings, middles, and ends. Providing satisfying callbacks. Demonstrating double entendre and unintentional meanings. These are the kinds of values the human ethnographer can bring. But it’s on us. We’ve got to demonstrate it. We’ve got to bring that value.