Advancing the Value of Ethnography

Leaving the Bed of Procrustean Experience: On the Need for Ethnography


User Experience (UX) Research and Design is a dynamic and diverse domain where designers, social scientists, and hybrids of all sorts are putting theory to work. It has successfully advanced a more holistic framing for human-centered design intervention, ideally keeping our attention on the user as the key unit of analysis at every stage. But we’re also discovering that herein lie potential opportunities for further refinement.

UX has familiar practical limitations, and we debate these continually—the best way to measure, how to communicate, appropriate sampling, sample size, methods, protocols, metrics, and so on. Its fundamental limitations, by contrast, are inherent theoretical assumptions and biases. It is critical to innovate at this level of UX’s underlying principles; to move beyond the generally unspoken assumptions that the user is necessarily an individual and that the user’s perceptions about discrete temporally and spatially bounded experiences are authoritative.

As a case in point, we’ll be looking at the design of a traffic intersection located in Drachten, Netherlands developed by Dutch engineer Hans Monderman. This example will help illustrate some of the ways that a User Experience perspective, as it is currently articulated, can potentially limit the possibilities for addressing the full range of human values. Finally, I will argue that User Experience design is greatly enhanced by establishing classical ethnographic methods as foundational for defining the domain of design intervention.

“User” & “Experience”

Let’s consider two ways to tackle our foundational concepts “User” and “Experience”. We all agree that experience is incontrovertibly user constituted to varying degrees. On one end of the spectrum, we can design experiences when we have relatively well-defined perceptual control, a preselected sample, and use-context control—as in art installations, theme park rides, or other immersive experiences. On the other end of the spectrum, where this type of control is undesirable or impossible, the designer might design for experiences: here the experiences themselves are emergent phenomena structured in situ by the user engagement.

These aren’t mutually exclusive, of course, but it’s a decent heuristic for discussing what model people tend toward when operationalizing experience design practice. This describes a conventional debate within user experience that, while addressing experience at a theoretical level, can be easily contained by mapping it along a gradient describing the degree of uncertainty that exists between user and designer.

Instead, if we question the scale of the system we’re evaluating or designing, we begin to see ambiguity in (1) who constitutes user and (2) what the bounds of experience are, however they are constituted. Are experiences appropriately constrained relative to the subsystem being designed and to individual user perception? Persuasive design and critical design are among relevant challenges to each of these. Typically, the scale of what is to be experienced is addressed (or not) upstream of any given design problem and so tends to disappear from design consideration. And consequently, the unit of analysis for experience is individual perception by default. Changing scale across space and through time, however, starts to complicate these assumptions and make visible their limitations.

An Example

Hans Monderman was a Dutch urban designer who redesigned traffic intersections around the idea of maximizing the necessity for driver engagement by eliminating (symbolic) signage. The result is an empirically safer intersection that feels less safe. You can read about it, among other places, here, here, and here. While Monderman’s justification, to promote individual responsibility, isn’t the best explanation for the socio-cognitive dimensions of why his plan worked, what’s important are the challenges it raises for user experience: Who is the user of the traffic intersection—a person, a community, or something else? What is the relevant method of perception (and whose perception is it)?

"Monderman in calmed traffic circle" by Jerry Michalski CC BY-SA 2.0

“Monderman in calmed traffic circle” by Jerry Michalski CC BY-SA 2.0

A designer tasked with designing better iconography for a road sign, or for a better driver assistance system, would entirely miss Monderman’s solution of eliminating signage. In fact, a review of user-articulated experience qualifiers could lead one in the exact opposite direction, toward increasing signage and formalized way-marking purportedly to increase perceptions of safety. Is Monderman ignoring users? Or, by contrast, is designing a traffic intersection actually an issue of system optimization having nothing to do with human experience at all?

A traffic intersection is very much human, and it exists beyond a narrowly defined problem space. We can easily imagine that user experience research might disclose a preference for safety (of individual drivers and pedestrians) while broader, more traditional, anthropological research might reveal a desire for enhanced ‘personal responsibility’ (among the community at large). Without research we don’t know; the point is, these aren’t necessarily the same. If the goal is to improve navigation about this intersection, delimiting experience inquiry to any one interface contained within it is risky. Of course the goal may be to enable community improvement by the community’s own standard of improvement relative to this traffic intersection. The problem is not merely that positive interface experiences might not add up into a positive holistic experience; but also that they may in fact be parasitic.

Who or What Is a User?

We can dig a little deeper here. We could argue that the user of a traffic intersection is the individual driver who must navigate the traffic circle. From this view, the interfaces to be designed are the signage (symbolic or lexical), or perhaps the formal moderation of traffic flow, or the vehicle’s interface, or something else. Alternatively, if we argue that the user is the community, the design space is different and we are faced with the issue of assessing a community experience, whatever that might be.

A third option is that the user is the individual, but the individual existing at a time scale that extends beyond one instance of interface. Monderman’s work seems to imply that he understands the user to be a combination of the second and third. The user exists across time within a community (is cultivated); and this widens to a great degree the boundaries of potentially considerable experiences—and consequently the methods needed to access, understand, and operationalize them.

Where, When, and for Whom Is an Experience?

How can UX account for subjective, positively regarded experiences that are time and community dependent? Examples of these might include such things as ‘a sense of responsibility’, ‘developing skill and mastery’, ‘sense of purpose’, or ‘sense of community participation’. There may be limitless others. All subjective states are more or less relative to time and culture, but I’d like to focus on those where that structure (diachronic and social) is impossible to ignore, and specifically those that cannot be accounted for by individuals in reference to superficially positive momentary stimuli.

What I mean here is that when Monderman seeks to design an intersection that engenders personal responsibility he is designing for an experience that is the result of some attentiveness training, ongoing commitment, transparent accountability, and skill development—an experiential assemblage that is costly in terms of effort and perhaps unpleasant or anxiety-inducing at the scale of one particular interface. Surely, if asked, no one wants to feel unsafe or exert more effort. Rarely does someone request a steeper learning curve or a series of challenging interactions. And so the individual user cannot be regarded as the authoritative voice in some sensory vacuum, but is instead enmeshed in a community of values within which they develop.

This unreliability of individual respondents is neither a matter of ability nor right to voice. The experience simply does not exist in whole form within the head of any one person to relay regardless of introspection; it is instead enacted through time in and among a community. The bounding of user experience inquiry necessarily precludes whole categories of value that can be loosely classified as ontogenetic or developmental: mastery, cultivation, citizenship, responsibility, stewardship, transcendence, enhancement, etc.


The conventional User Experience perspective, then, recommends two potentially unreliable, or at least ambiguous, foundational principles—user and experience. To get beyond the narrow ambitions of these terms in their most limited sense, we must define them as larger, longer, and more distributed than our current processes comprehend. We are still interested in the user, but must admit that we might not know a priori the relevant boundaries of a domain of use or how to define a user.

We do know that deciding in advance potentially biases our results— precluding positive results at one scale by arbitrarily privileging positive results at another. So when selecting methods for eliciting, describing, operationalizing, and validating user experience, we should be aware of at least the following points:

  • Experience is developed over time, in and across particular contexts.
  • Those contexts play a contributive role.
  • Experiences exist at different scales and are more than distinct affective or cognitive engagements of individuals engaged with discrete stimuli.
  • Users are not static; they also develop over time and with other people, and may themselves be plural.
  • Not all values are compatible with universal experiential accessibility
  • Not all explicit values are compatible with implicit values
  • Not all proximal values are consonant with distal values

Think of an orchestra: “the experience of solidarity and commitment when a group of people achieves mastery at a piece of music they are learning together” is one type of experience that isn’t easily reducible to a single individual interacting with the particular aspects of music production. Indeed, individual engagements with particular instruments may be frustrating, counter-intuitive, or at times baffling. Less empathetically, exclusivity may also play an important role in defining the value and potential of a given activity.

Do we address this “problem” by designing easier-to-play instruments or requiring each part in an ensemble individually to be a fulfilling, pleasing-to-play piece of music? Surely not: these solutions could be corrosive or irrelevant to the larger (actually desired) outcome. Perhaps pedagogy or time management improvements are needed if we’ve indeed identified a problem. We don’t know; what we do know is that what gets defined as “the experience” matters, and that needs to come from the users situated in context. We simply cannot assume what the appropriate level of analysis is before doing the fieldwork. What is the experience of an instrument? of playing? What is the experience of practicing? of music selection? What is the experience of instrument maintenance? These may or may not be relevant, and prioritization framework will be structured by more holistic user goals.

Given these issues, standard UX measures prove to be useful but insufficient. Learnability, performance, retainability, usability, emotional resonance, intelligibility—these and others cannot hope to account for a holistic view of experience that might also include skill, commitment, mastery, conviviality, trust, self-improvement, craftsmanship, pride, etc. Our standard measures may potentially even interfere.

To account for context and interpersonal interaction without losing formal analytic accuracy, we should:

  1. Use classical ethnography to ground your subject and your constructs in an ecological context (to calibrate spatial, temporal, and social scale).
  2. Use formal methods to establish construct validity after you extract the constructs from the field (do not allow them to originate in laboratory contexts).
  3. Consider experiences at the scale they exist in the wild.
  4. Consider that users change over time, depending on situations.
  5. Consider that not all aspiring users are potential or actual users.

What this means, however, is that fully immersive abductive ethnography of the classical tradition is necessary for defining the ontology of the problem space and spatiotemporal domain for appropriate intervention. Properly, hypothesis-agnostic ethnographic methods will need to be used for the initial discovery and framing of design activity; maintaining attendance to longer time scales and ecologically situated social interaction beyond the superficially relevant discrete phenomena trimmed for methodological expediency. The goal is to maintain disciplinary integrity but also to essentially protect our community from economic risk and gradual erosion of shared values.

This isn’t news to practitioners who’ve been doing this for much longer than I have, but I hope it’s a reminder that extends beyond argument into actual practices, as this has implications for how we understand and work with other design research approaches. One such implication is that being user-led in this way might challenge how we measure. We will need to innovate and experiment with new tools to account for, measure, and validate such an extensive conceptualization of experiences.

A second implication is that this may force us to rethink our default assumptions as to what, when, or who users are. Understanding and articulating experience beyond individual engagements is a first step toward developing appropriately scaled measures.

Lastly, while persuasive and critical design (tactically and theoretically) have made significant headway in understanding design from user de-centered perspectives, it is not altogether necessary to accept any dichotomous classification between bottom up and top down approaches. Any system will, in fact, be dynamically interactive and, so long as experience may be defined developmentally and or inter-subjectively, we may find ways to bridge apparent binary distinctions and provide design interventions with lasting, socially intelligible, meaning.

This article was prepared or accomplished by Michael Thomas in his personal capacity. The opinions expressed in this article are the author’s own and do not reflect the view of Ford Motor Company.

michael thomas

Michael Thomas, an anthropologist and designer living in Nanjing, China, is currently working at Ford Motor Company in order to leverage design expertise as well as anthropological sensitivity and rigor toward the engineering of coherent, culturally appropriate, and socially intelligible automotive and mobility solutions. As an active representative of the Society for Anthropological Sciences and doctoral candidate at Wayne State University his research engages cultural anthropology, contemporary archaeology, and STS perspectives to better understand the experience of personhood across the entangled networks of engineering work, artifacts, and people.


Evolution of User Experience Research, Kathy Baxter, Catherine Courage & Kelly Caine

From UI to UX: Building Ethnographic Praxis in a Usability Engineering Culture, Kirsten Bandyopadhyay & Rebecca Buck



Michael Thomas, Ford Motor Company