Abstract: User Experience Design as a profession did neglect its defined and methodologically well-supported core tasks in favor of higher prestige tasks. Such a process is called “Professional regression”. The term User Experience Design might already be the result of an attempt to elevate the profession by extending its definition, however, without providing the methods to actually control the field of work by showing rationality and impact. In the end, the article calls for more focus on professional problem solving and less on one-upping each other with abstract definitions or promises of scaling, based on the argument that the former tends to strengthen the discipline and the latter only strengthens some privileged people and weakens the discipline
Are we there yet?
It is hard to say if the profession of User Experience design achieved its goal of being recognized as a profession; a profession that solves problems that are useful and has a unique expertise being able to solve these problems.
We seem to have a place at the table where important decisions are made. Design is often emphasized as an important factor in developing companies and innovative products and services. However, at the same time, many of us still need to explain the basics of our discipline as if it is 1995: Yes, it is important to do tests, no, it does not suffice to just “make it look nice”. This is exhausting and, if part of a job, can drag on for years. This pattern is widespread enough that many people in our UX-Design think that evangelizing the mere existence of one's role is somehow part of the job.
Both, the hopes for a “seat at the table” and the difficulties to explain why user experience design needs expertise at all might be two sides of the same coin, partly caused by neglecting core aspects of our discipline in favor of a few prestigious management-like jobs that allow distance from the messy work of testing, arranging UI elements and finding out what other people want and need—an instance of “professional regression”.
Prestigious roles and professional tasks in UX Design
The concept of professional regression
Professional regression is a concept from Andrew Abbott that I learned about reading his book “The System of Professions” 1. It describes the process in which professions stratify into lower- and higher prestige areas. Stratification in professions happens along the abstraction of knowledge and the use of that knowledge. This protects professionals from being seen as craftspeople, having less social prestige. The abstracted knowledge also makes them somewhat independent from the specific tasks and technologies they work on—otherwise, if the task would disappear, the profession would disappear, too.
However, abstraction is a double-edged sword, since abstraction means professional prestige and the highest levels of abstraction are not to be found in the most tasks of professionals. Abbott writes: “Since the professions are founded on knowledge, admiration peaks when knowledge is most pure, that is, when it is least deformed by actual application.” (Abbott 1988, Ch 8) This gives us a core problem of stratification, that professions internally stratify along levels of purity: “Professionals admire academics and consultants who work with knowledge alone; the public admires practitioners who work with clients.” (Abbott 1988, Ch.5). The work of a profession that is easily seen as useful and even admirable by people outside of that profession is usually low in prestige from the perspective inside of that profession because it is “deformed by actual application”(Abbott 1988, Ch.8).
To gain prestige, professionals try to get away from the deformed practice, towards more prestigious, more professionally pure activities—which are, however, seen as less useful by people outside of the profession.
I want to describe three developments in User Experience Design that can be seen as professional regression in the profession: Business-consultant roles, roles concerned with scaling, and attempts to increase professional power by claiming a larger field by the use of increasingly abstract definitions.
Consultants, scaling and large-scale definitions
The hybrid of User Experience Design and business consulting is best expressed in the idea of Design Thinking. Influential design agency IDEO fused human-centered design and business consulting 4. Like other business consultants, they promised large scale transformation, innovation and a profitable future for those who hired them. People with design thinking roles often worked in “innovation agencies” doing strategic research (often for big companies in IT or fast moving consumer goods) or were internal consultants to bring “design thinking” into other teams.
Roles concerned with scaling claim to make fellow professionals more efficient by providing the correct infrastructure for their work. This can be tools but also standards, workflows, career development etc., which is often happening under the concept of “Design Ops” (Ops=Operations). These roles make most sense in large companies, where there are other designers to manage and explicit infrastructures to be established.
These roles and areas of concern loomed large at conferences and in discussions about the development of the profession of User Experience Design. They solve some problems that large companies have like the lack of fast innovation and difficulties in scaling. However, the jobs they offer are limited. Design Ops needs other designers that benefit from the operational infrastructure, Design Thinking needs companies that believe in strategic consulting and employ people in need of that consulting. Most user experience professionals work in other roles. Both design thinking consulting and design ops are professionally more pure than a lot of other design work: They do work with and for other professionals in larger organizations and the work happens on a meta level: Designing design, designing innovation and thus is somewhat protected against the status-lowering impurities of other design work that interacts with non-designers or people who do not value design very high.
Last but not least, we have attempted to define UX Design more broadly than it is usually understood, trying to position designers as experts for all things related to design (in the sense of planning or purposeful action) or value creation for customers. These are large-scale definitions, similar to claims that medicine is related to everything concerned with human well-being. They might work for some people; usually ones that have a lot of prestige already. The definitions can also be very aspirational, and motivating for beginners (belonging to a very powerful profession!) but most people without a lot of prestige can't act upon these definitions as it is hard to enforce for most members of the profession: Planning and value creation are often seen as the turf of various managers who don't like intrusion in their field as much as designers like managers who design. It is also not a new problem —even the field and title of “user experience design” can be seen as a difficult claim on “all things related with user experience of a product or service” and thus being already in a still unresolved competition with marketing, other designers and managers.
Having a defendable expertise
To see what expertise UX Design can claim and defend well, we need to look at the mechanisms used for claiming and defending an area of work or, what Abbott calls “jurisdiction”: “Diagnosis, treatment, inference, and academic work provide the cultural machinery of jurisdiction. They construct tasks into known ‘professional problems’” 5 (Abbott 1988, Ch.3). Diagnosis means determining the problem, treatment is an action that resolves or mitigates the problem and inference is what connects the two: What treatment is best suited for this problem?
Diagnosis, inference and treatment in UX Design tasks
Improvements to usability is area of UX Design where diagnosis/inference/treatment can work well together. There are several methods to determine problems (Usability testing, heuristic analysis, log-file analysis) and several standard methods to suggest improvements like avoiding memory load, using metaphors and language known to the user.
It is helpful that most of these methods can be supported by cognitive psychology and its research into working memory and language use, though there are also influences from lesser known academic areas like ecological psychology (which brought us concepts like affordances) and there is a history of direct research in human computer interaction that often builds on ideas of cognitive psychology but also provides some connection to other academic fields, including ethnomethodology, sociology, as well as engineering-based visions of potential uses of technology. This provides a somewhat abstract foundation, not bound to a particular technology, as well as the prestige of academia that allows to make claims seen as true and rational by others 6.
Another interesting property is the that the process of making interfaces more usable is somewhat hard to commodify and automate: While there are principles for justification of action (e.g. “familiarity leads to intuitive understanding”, as in “We should rather use the word_____, since our users are more familiar with the expression”) what exactly shall be done still needs professional expertise. Interface guidelines and design systems systematize and commodify this somewhat, but it seems that for an easy-to-use interface, some knowledge is needed beyond having read the guidelines (aside of the fact that guideline interpretation also needs skill)
I focus on the usability here, that is, designing interfaces that are easy to learn and to remember. The market for this is big, first due the sale of shrink-wrapped, non-custom software (See Grudin 2017 2, p.41) and then, far more extreme, via websites—both situations where offering an explicit or implicit training for use of the software or website is not desirable or simply not feasible. However, similar methods as for usability can be applied to efficiency of use, with studies on performance and a catalog of common treatments e.g. keyboard accelerators, arrangement of UI elements, prioritization of features. Even joy of use can be studied (for example with surveys), though standard methods for treatment are less defined.
Focusing on “fewer errors, faster performance, quicker learning, greater memorability, and enjoyment” (Grudin 20172, p.92) seems to work well from the perspective of Abbott's work on professions: We have somewhat standardized and academically-backed methods for diagnosis and treatment. Professional inference connects diagnosis and treatment, a process that is at least somewhat protected from commodification (that is, packaging it in mechanically applicable “if problem is this, then treatment is that”-way). These are also the tasks in UX design where the positive impact of the work can be shown most easily: Before-and-after studies can be done, but very often the solution is often obvious or at least plausible7.
Impure tasks and devaluation
These areas of our discipline that have diagnosis/treatment, have supporting academic knowledge and show their positive impact most distinctly are also the ones also are somewhat impure: Compromises need to be made constantly and we work with roles that are not UX designers8. The prestige and the conference talks are often focusing on more pure aspects. However, the pure work is not what justifies UX design as being useful in the eyes of non-designers 9. Consequently, I argue, we have neglected the development of methods for diagnosis and treatment for usability, efficiency and enjoyment.
For every workshop and talk on these questions, there seem to be ten on how to work better with managers or how to be one as well as how to scale design. Roles that are concerned with such work are more prestigious, but do not help the development and strengthening of “Diagnosis, treatment, inference, and academic work”.
We also actively participated in our own tasks’ devaluation: With “democratization”10 some made the bet to be the facilitator or consultant for people being professionals in other roles. Interface design, the sub-discipline dealing most obviously with visible results was devalued to a point where no one wants the title UX/UI. Usability testing, the most common way to do an empirically based diagnosis is seen as boring.
Consequently, we might develop into a non-discipline that is all pure periphery and no core; the derelict discipline’s tasks being picked up by developers, product managers and data analysts, none focusing on users but all being able to solve problems for clients and corporations.
A plea to continue working on mundane methods
My plea is thus to focus on the more mundane aspects of our discipline, aspects that also non-designers see as solving problems that need solving and to develop and adapt methods that contribute to these core aspects of the discipline: Diagnosing problems and suggesting treatments. This will need some connection to academia while focusing on the practicality of methods (which does in turn create impurity for academics!).
Some areas where I can imagine development of such methods to be helpful could be:
- Adapting and advocating statistical methods that work on small sample sizes (since it would work well with the common iterative approaches in design
- Adapting and advocating for statistical methods that consider effect size and equivalence testing11.
- Designing for a variety of display methods on different media while pushing back on the idea that designers should just 1:1 adapt the way developers define flexible screens (UI design apps figma and penpot provide visual interfaces for defining these, but they are not read to read)
- Adapting ideas from domain-driven design, coming from software architecture as a means to ensure cross-role communication (could also work well with user story mapping)
- Developing qualitative research away from often decontextualized interviews to learning about practices or learning these yourself (which is what contextual design tried to do, but its focus on large teams made it unwieldy for many UX designers and researchers)
There are just some ideas for ways the discipline could choose as worthwhile challenges that directly could contribute to core tasks.
Developing the groundwork that solves problems for people who are not UX designers would develop the discipline, continue to show its usefulness and allow it to accommodate changes in technology and workplaces in a more concrete way than to bet on a more abstract definition that is impossible to defend. It also keeps the field open to beginners while maintaining professional standards. This does not mean that the profession can be fully learned in the abstract or in a short bootcamp. On the contrary, it would help to show the value of both practical experience and their connection to abstract knowledge without being reducible to it.
I also see some ethical dangers in aiming to recast the professional role as entrepreneurial and business-value-creation. It might describe the idea of the neoliberal cooperation, but that does not mean that I find it aspirational (quite the opposite). Work in UX Design also should bring in a user-informed perspective and pushing back against practices that earn money but hurt users. Without a well-defined core jurisdiction and valuing one’s own profession (even the impure parts), it is hard to push back in the first place, and it is even harder when the redefinition of the discipline means being more aligned with management and its value systems.
Having criticized UX Design as a discipline for focusing on consulting and scaling does not mean that I think these professional tasks are bad or useless: I think these tasks are important in some contexts and that they do important work. However, they are unlikely to develop UX design as a profession and thus we should be cautious with the prestige we assign to these roles.
- Design Methods and (professional) power – On methods as means of coordination and performance of rationality
- Software on itch.io and the costs of scaling – Best practices as being bound to the prestigious roles and contexts of a profession
- On violating the standards of one’s occupation – Why it is important to “act professionally”
- Frames of industrial production and positivism in the design process – Design processes are described as universal and rational; design practice, however, is not.
- Thanks to Stefanie Kegel for providing many helpful comments on a draft of this article
- Donald Schön (The Reflective Practitioner) gives a theory and description for the messy parts of professional work. He is, however, not focusing on the struggles for jurisdiction and definition that happen between different professions.
Abbott, Andrew. 1988. The System of Professions: An Essay on the Division of Expert Labor. Chicago: University of Chicago Press. ↩
As an example from the medical profession, based on how I understand Abbotts writing: A GP has less prestige than the principal consultant in a hospital, because the GP deals with messy cases, many of which are not medically interesting—like convincing people that a problem will just go away. However, in contrast to a GP, the principal consultant works in an institution (the hospital) that helps to enforce structures and they can work with other medical professionals who know how to act correctly according to the ideas of medical professionals. ↩
Irani, Lilly. 2018. “‘Design Thinking’: Defending Silicon Valley at the Apex of Global Labor Hierarchies.” Catalyst: Feminism, Theory, Technoscience 4 (May): 1–19. https://doi.org/10.28968/cftt.v4i1.29638. ↩
Abbott, in this quote, does not order them temporarily, which would be Diagnosis→ inference→treatment. Here we might have another challenge for user experience designers: We emphasize processes a lot. However, to the outside, diagnosis and treatment are most important, since these are visible. This does not mean that process, reflection, iteration are not important, but that it is hard to sell. Dealing mostly with the process itself, again, mostly possible for prestigious roles. An interesting discipline to engage with might be psychotherapy—they deal with slow processes and vagueness of outcomes a lot! ↩
Rationality as value might seem obvious today, but other values were once more important—Gentleman-like behavior, for example (Abbott 1988, Ch. 6) ↩
Checking for correctness seems obvious (Metrics! Rationality! Improvement!) but is actually tricky and very costly; To use a familiar example outside of design, most people trust their GP or maybe get a second GPs opinion but would not think of doing a time-series study on their betterment. ↩
A quote that a Stefanie Kegel pointed out to me: “Listening to user frustrations is the product designer's equivalent of asking a patient where it hurts.” (Kim Goodwin in Designing for the digital age). Such listening and asking questions is inherently messy: People are not professionals in telling where it hurts and it might need a lot of probing and trying until both sides have a common understanding and both might end up not having found the ideal outcome for their concerns. ↩
Rather than solving their problems, UX designers often make the bet that they might be able to tell people that the non-professional’s idea of the problem is wrong ("Yeah, but what is the real problem?!") rather than keeping their problem definition intact and to show how the UX designers own redefinition enables useful work. ↩
I am unsure what that term means. In practice, it seems to be translatable into ‘making it more likely that more people do this activity or have it as an active concern’. This makes it unlike modern state democracy with separation of powers, complex bureaucracies and deliberation and also unlike other forms of collaboration that are concerned with the distribution of power; it seems most similar to launching a cheaper product that is thus more accessible-to-more-people. ↩
Often it is useful to know if two variants are equivalent rather than different:Your new design follows best practices but the old one might have had years of tiny adjustments build in. A simple statistical method to do an equivalence tests is TOST. ↩