I first read The Corruption of Co-Design (Von Busch and Palmås, 2023) when it came out late last year, around the same time I was writing about organisational metaphors and performance and substance. I found it very funny, bracing and a welcome critique of design's idealism and naivety that I could also very much relate to from my own experiences over the past few years. I already knew Hill's (2012) argument about "dark matter" and designers' naivety about organisational power, and had been writing about the gap between mission-oriented ambition and implementation reality in my own context. Von Busch and Palmås were working similar territory but pushing further - where Hill describes the organisational forces that designers fail to see, they were asking why design's own intellectual traditions leave it structurally unable to see them.
I'm returning to the book now, a few weeks after writing What I Learned at SCÖ and the social defences submission for NORDES, and want to consider the six Realdesign propositions to feel less like theory and more like a post-mortem, and use them to help continue the reflections of what went wrong at SCÖ and what, as part of my PhD Design Research I can draw from these experiences. Tonkinwise's lecture "All Care, No Responsibility" from earlier this year sharpens this further; his argument that service design's tools carry political inheritance - that the service blueprint is fundamentally a management tool, that design's focus on user experience systematically misses what he calls "contexts of aggregation" - connects to the emotional labour I was trying to name in the NORDES submission. The idealisation, splitting, and blame I described there are not incidental to design practice; they are what happens when a discipline markets itself as transformative while providing practitioners with no vocabulary for the politics in which they are working, and gestures at the personal cost, and emotional toll of that work.
But I want to make a distinction that I think both von Busch and Palmås blur. The idealism they diagnose is real, but it lives more in design's self-presentation - the way design is marketed, the programme logics in which it operates, the sacred quality that attaches to its methods and rituals - than in what a design practice grounded in material specificity actually does. Design engineering, the kind of work I do, and which I was doing at SCÖ, is a relentlessly materialist practice: it forces abstraction into specificity, tests technoimaginaries against technical and institutional conditions, produces artefacts that make visible, and proves, or at least tries to empirically test what would actually need to exist, or at least makes real, or tries to, the first functional prototypes of those ideas. That practice did challenge the idealism at SCÖ. It brought things to life, as design should, by insisting on what was real. But it also ran naively into the politics and power structures that von Busch and Palmås describe - because being right about the material situation is not the same as understanding who benefits from it remaining abstract.
I want to work through the Realdesign propositions against the SCÖ experience while it's still close enough to be honest about, holding that tension between design as materialist practice and design as politically naive actor.
Design as materialising practice
Before I get to the propositions, I want to push back on one aspect of how von Busch and Palmås frame the argument, because I think it matters for what follows.
They characterise design as operating through "idealist-tainted glasses" - treating the social world as raw material to be shaped by empathetic designers, assuming the best argument wins on its merits. That's a fair description of something. But it doesn't describe what I was actually doing at SCÖ, and I think the distinction is important.
What this looked like in practice was the algorithm archaeology - running the Pathway Generator's Python scripts on synthetic data, writing typed interface definitions that specified its data structures variable by variable, producing architecture diagrams of what a full implementation would require. Every artefact I produced was an act of materialisation: taking something that existed as a technoimaginary (a collectively performed vision of a technological future) and testing it against material conditions. What does this algorithm actually need? What data structures does it require? What governance frameworks, assessment protocols, and institutional capacities would have to exist before any of this could work?
The idealism at SCÖ was not mine. It was the programme's. It lived in the ESF funding application that promised "data-driven decision support" without verifying whether the data existed. It lived in the consortium agreements that invoked "federated learning" as if naming the technology was equivalent to building it - and, as I explored in an earlier post on sacred service design, that naming served a ritual function, with "federated learning" operating less as a technical specification than as what Matthews (2021) would call a sacred representation, constituting community among the consortium partners and investing the project with a transcendent significance that resisted profane technical scrutiny. It lived in the milestone reports that showed GREEN while the foundational premise was quietly abandoned. The programme treated the social and technical world as something that could be willed into shape through sufficiently ambitious proposals - which is, ironically, exactly the idealism von Busch and Palmås diagnose in designers.
Design's contribution was the opposite: materialising what the programme preferred to keep abstract. Concept maps that forced questions about which data existed and which didn't. Typed interface definitions where every variable demanded: "what protocol gathers this? Does anyone in Sweden actually measure this?" Architecture diagrams that made visible the full infrastructure that would need to exist before a single algorithm could run.
So when von Busch and Palmås say designers need to "get real", I'd add a complication: sometimes design is the real. The practice of materialisation - of making concepts specific enough to test - is what design does that programme management, funding applications, and academic abstracts do not. The problem isn't that design is idealist. The problem is that materialising practice is politically dangerous when the surrounding institution has invested in keeping things abstract. And I wasn't savvy enough about that danger, nor sophisticated enough in thinking that just by showing that the emperor had no clothes everyone would also see it, and be prepared to work together to design him some new ones, so we could at least try to deliver on the programmes aims, or at the very least document our learning from doing so.
Proposition 1: Design must get real - and who whom?
Von Busch and Palmås argue that designers should replace the idealist "what if?" with the realist "who whom?" - who does what to whom, for whose benefit? Design proposals get activated not through the quality of their argument but through tensions between networks of power, incentives, and influences.
I was already asking a version of the materialist question: what would this actually require? That's different from "what if?" - it's closer to "what is?", or "what would it take, materially, functionally, conceptually?" But I wasn't asking the power question: who benefits from things staying abstract? Who has invested in the technoimaginary? Whose position depends on the fiction being sustained?
The consortium was a network of aligned incentives - aligned around sustaining the fiction. The Icelandic academics got a new market for their Pathway Generator. The UK academics got a research collaboration, the Swedish academics got a PhD student. The Swedish organisations got the legitimacy, or kudos of an international data science partnership. The funders got an innovative-sounding project in their portfolio.
I perceived these dynamics at the time but underestimated the power of those relations. I assumed that design would serve as what Brandeis called "the best disinfectant" (Hood and Heald, 2006) - that making things visible and specific would, by itself, be corrective. I understood the political dynamics in the abstract - I'd written about conflicting interests and the appearance of neutrality. But I hadn't done the specific power mapping that von Busch's "who whom?" demands. I hadn't asked: whose career depends on this narrative? Who co-signed the promises? Who controls the relationship with the funders? If I had, I would have understood that my algorithm archaeology wasn't landing in a neutral epistemic space where truth, or even materialism is welcomed. It was landing in a political space where specificity threatened people whose positions depended on, or could more safely hide behind abstraction.
My concept and context maps showed that the Pathway Generator required data infrastructure, governance frameworks, and assessment protocols that didn't exist in Sweden. This was accurate. But accuracy is not the same as leverage. Von Busch would ask: who had the power to act on what this map showed? The answer was nobody who wanted to - because acting on it meant acknowledging that the premise was wrong, which threatened the academics, the funding narrative, and the consortium's collective investment in the fictional expectation.
Proposition 2: The process of change is a process of conflict
Von Busch and Palmås argue that change alters power dynamics - someone's position will be advanced, usually at the disadvantage of another. They ask: "who cleans up the mess and heals the social wounds after the 'design disruptors' have left?"
I wasn't trying to be a disruptor. I was doing what designers do: making things specific and visible. But specificity was disruptive to a system whose coherence depended on abstraction and ambiguity. My artefacts threatened the academics' claim that their tool could transfer, the funders' narrative of a project on schedule. Every design artefact that said "unclear what protocol or data we could use for this" was, in political terms, an act of destabilisation.
The social bonds within the consortium had been built around a shared fiction, or perhaps more generously an extreme act of speculation, but speculation conducted within a NPM governance frame - that ultimately was where the power lay. My materialising work profaned the fiction. What followed was not open conflict but the quiet mechanisms of decoupling - milestones silently redefined, status reports with RAG scores staying green, an "insight report" perpetually promised but never materialising.
The decoupling maps onto what Fotaki and Hyde (2014) call organisational splitting - the defensive separation of incompatible realities so they never have to be confronted together. As I explored in the social defences submission, splitting in design contexts is not merely organisational dishonesty; it is a social defence against the anxiety that would follow from holding both truths in the same frame. At SCÖ, the split was between the project's official narrative (innovative, data-driven, on track) and the material reality my design work was surfacing (technically impossible, data-absent, structurally misaligned). The programme sustained both realities simultaneously by keeping them in separate registers - milestone reports in one, my concept maps in another - so that neither had to confront the other. The anxiety being defended against was not trivial; it was the recognition that the consortium's founding premise was wrong, that careers and funding had been committed to a fiction, and that acknowledging this would threaten every partner's investment in the project.
What I missed
I was treating materialisation as epistemically neutral - as something the project needed in order to proceed. Von Busch is saying that materialisation is always a political act, because making things specific redistributes power. It takes power away from those who benefit from ambiguity and gives it - potentially - to those who need specificity to do their work. At SCÖ, the caseworkers needed specificity (which actual tools, using which data, through which protocols). The consortium needed ambiguity (federated learning as possibility, not as engineering requirement). My design work sided with specificity. I just didn't recognise that as taking a side, or, I underestimated the political implications for myself in what I otherwise considered to be an earnest and good natured attempt to do the job of a designer and design researcher in trying to help the programme deliver on its stated aims.
Proposition 3: There is something rotten in design, and the user is the victim
Von Busch and Palmås's formulation: "Users are hostages. They are held by designers as human shields". If the vision works, credit goes to the designers and the consortium. When it fails, users are left where they started.
The vocational rehabilitation caseworkers and their clients were invoked in every funding document, every milestone report, every consortium meeting. They justified the entire project. But they were nowhere near the actual work. The project's energy was consumed by whether we might find data to test the Pathway Generator, while the service context - the relational infrastructure that actually shapes rehabilitation outcomes - went unaddressed, and, less engaged and interacted with than it should.
I don't think I was naive about this. I'd already described the dynamic in Images of Organisation: "The municipalities want tools that reduce their caseloads. The employment service wants to meet its placement targets... The university researchers want publications and demonstrable innovation". But what von Busch adds is the more critical edge: the invocation of users doesn't just happen to neglect them - it actively serves the interests of those doing the invoking. "Under the delusion of user-power, both use-situations and users themselves are designed to align with the designated distribution of power set by clients, institutions, or funding agencies". Users in the funding application were designed artefacts - they lent legitimacy to a technology project that had, ultimately, and for no lack of trying on my part nothing to offer them.
The invocation of users also served the defensive function I described in the social defences submission: idealisation projected not only onto the designer but onto the users themselves. The caseworkers and their clients were idealised as the beneficiaries of a technological transformation that the consortium had neither the data nor the institutional capacity to deliver. This idealisation sustained the project's legitimacy, because questioning whether the technology could actually serve the users felt like questioning the commitment to serving them; the goodness of the aspiration - who could be against helping vulnerable people find work? - made the technical fiction unassailable.
And - the harder admission - my own design work participated in this. I empathised with the caseworkers. I mapped their context. I genuinely wanted to understand their needs. But my employment was contingent on the project, and the project's purpose was not to serve caseworkers. It was to pilot federated learning. My user-centred framing gave the project a humanistic veneer while the actual work was technology-facing. Von Busch would say that's structural, not personal - it's what happens when design is positioned within projects whose power dynamics it can't challenge.
Tonkinwise (2023) names the personal cost more directly: the expectation to empathise with users while serving a project that cannot deliver for them is a form of what Hochschild (2012) identifies as emotional labour - managing the gap between what you feel and what the role requires. As Hirsch (2020) observes, design research can become a kind of therapeutic encounter conducted without the training or institutional support that therapists receive. I was performing user-centredness within a project structure that had no mechanism for acting on what user-centredness revealed; that performance, at least on my part, was not cynical, but sincerity does not alter the structural position.
Proposition 4: Good is dead
Not nihilism - the argument that Good, as a free-standing vision guided by high-minded intentions, doesn't win on its own merits. Visions not backed by political agency are edgeless.
The SCÖ project was full of good intentions. Everyone genuinely wanted to improve rehabilitation outcomes. But good intentions and fictional expectations are not mutually exclusive. As Beckert (2016) showed, imagined futures mobilise resources precisely because they don't require anyone to verify feasibility. The goodness of the intention sustains the fiction - questioning feasibility feels like questioning the commitment.
There is a connection here to the sacred quality I have been exploring through Matthews' (2021) work on ritual in service design. The consortium's commitment to the federated learning vision had the character not of a falsifiable hypothesis but of a shared article of faith - a sacred narrative whose function was to constitute community among the partners and invest their collective work with transcendent significance. The project meetings and milestone reviews functioned as rituals that reinforced this commitment; the funding application operated as founding myth; the prototypes and diagrams served as sacred symbols. Questioning the technical feasibility of federated learning was not, within this frame, a neutral epistemic act - it was profanation, and profanation provokes not rational reassessment but defensive retrenchment, because what is threatened is not a technical claim but a communal identity built around the claim.
Von Busch and Palmås mount a specific critique of empathy, drawing on Tonkinwise (2010): empathy misses contexts of aggregation - systemic conditions, historic struggles, political settlements. Tonkinwise's (2023) SVA lecture extends this critique into the tools themselves. The service blueprint, he argues, is fundamentally a management tool; the line of visibility does not merely represent a division between frontstage and backstage but conceals the labour that makes the service work. Design's focus on user experience systematically misses what he calls "contexts of aggregation" - and his honest limit captures what I was groping toward at SCÖ: "Design is not the answer. Design can contribute to the creation of sociality, the restoring of sociality".
The tools I used - concept maps, architecture diagrams, mapping of system interfaces - were not service design's standard toolkit, but they shared the same underlying assumption: that making something visible and specific is inherently productive, that representation is a form of agency. Tonkinwise's critique suggests that the problem is not which tools you use but whether you understand whose interests the act of representation serves.
I was operating in the register of truth: if I make the impossibility specific enough, surely people will respond. My materialising practice was grounded - it tested abstractions against reality, it produced accurate accounts of what existed and what didn't. But accuracy without political backing is a spectacle, not a lever. I wrote in The Limits of Making Visible that "the assumption beneath design's visibility paradigm is that organisations want to see clearly". Von Busch sharpens this: the assumption isn't just wrong, it's idealist in a specific way - it mistakes description for force. My artefacts described the situation with precision. They moved nothing.
The idealism I'm guilty of isn't about the situation - I was right about the situation. It's about the theory of change. I assumed that making the truth visible would generate the conditions for acting on it. That's the designer's version of the conviction that the best idea, or at least the most visible, or most tangible idea wins - or is integral to building consensus and designing or infrastructuring the social.
Proposition 5: Realdesign is a creative form of realism
Power constraints are design parameters, not barriers to be dreamed away.
This is the proposition I find most generative. If I'd treated the SCÖ constraints as design material - the funding logic, the academic self-interest, the consortium's collective investment in abstraction, the accountability structures that rewarded performance over substance - what would I have designed differently?
Not better maps of impossibility. The creative realist question would have been: given that federated learning is impossible in this context, what is possible with the resources, relationships, and political energy this project has assembled? The consortium had funding, academic expertise, practitioner relationships, and a mandate for improvement. The technoimaginary was fiction, but the organisational infrastructure was real. Swedish vocational rehabilitation genuinely needs better data practices, better coordination between agencies, better ways of sharing knowledge across institutional boundaries. None of that requires federated learning. The consortium was, in Rip's (2012) terms, an innovation journey - and its institutional context was not a backdrop to the design work but the material within which it had to operate.
Tonkinwise's concept of "contexts of aggregation" points to what I was missing: the systemic conditions, the historical struggles between Swedish agencies, the political settlements that produced the current fragmented landscape of rehabilitation services. Empathy with individual caseworkers - which I genuinely had - could not see these aggregated conditions, because empathy, as Tonkinwise argues, sees the individual interaction but not the system. A creative realist approach would have required mapping not just what individual caseworkers needed but what institutional conditions, inter-agency agreements, and political compromises would need to shift - and which of those shifts the consortium's assembled resources could actually support. That is the work that I was trying to do, and had I not lost my job, might have been able to continue doing.
Von Busch says the primary question must be "constructive under what conditions and constraints - and constructive for whom?" A designer who understood the power dynamics better might have led a pivot rather than documented one, although in fairness to myself I rather ran out of time before my contract was terminated to do so.
It's not that I didn't try, but it was naive because I was still operating under the assumption that truth was sufficient, and perhaps was trying too hard to seek acceptance and acknowledgement of the impossibility or limits of the original project proposal, and a more formally agreed pivot to something more rational. I think I was believing that if I could just show clearly enough what was real and what wasn't, the organisation would correct course. Von Busch is telling me that organisations don't correct course in response to information. They correct course in response to power.
It's an interesting question about why I didn't have more power to do that, and I think it comes back to the fact that I was an employee of the project, and my position was contingent on the project narrative. I had no independent power base within the consortium, and the project narrative was something that every partner had invested in, and co-authored, months before my arrival or involvement - so challenging it felt like challenging the consortium itself, and the consortium, in large part, closed ranks and protected itself - discarded those not willing to sustain the technoimaginary.
Proposition 6: Ask realist questions
The final proposition offers a practical toolkit - power-literate questions that sit beneath the standard design questions of "how" and "what".
Applied to SCÖ, and being specific about things I should have been specific about at the time, the first questions concern who the project actually served. The nominal users - caseworkers and rehabilitation clients - were invoked in every document but structurally absent from the work; their interests were genuine but disconnected from the project's activities. The actual clients were multiple and diverging: the coordination association wanted innovation legitimacy, the Icelandic academics wanted to extend their tool into a new market, the universities wanted publications and a completed PhD, and ESF wanted evidence of data-driven public sector innovation. These interests aligned around the fiction, not around the users. Tonkinwise's point about design's user focus missing "contexts of aggregation" applies here too - the question was never really about individual caseworkers but about the institutional arrangements that shaped their work, and no one in the consortium was asking that question.
The power relationships were equally opaque to me at the time. My formal reporting lines ran to academic supervisors at two universities and to the project leader at SCÖ. The structural frameworks within which all of this operated - ESF funding logic that rewards ambitious proposals over feasible ones, academic career structures that incentivise publications and international collaborations over implementation, consortium governance that distributes accountability so thinly that nobody is responsible for feasibility - all rewarded the maintenance of the fiction and penalised its disruption.
The outcome speaks for itself. My design work, which threatened the narrative, resulted in my position being the only one of the project team terminated. I got burned, and the caseworkers and their clients got nothing. The full cycle of social defences I described in the NORDES submission played out in miniature: the idealisation of the designer as the one who would make things work, the splitting between the project's official narrative and its material reality, and finally the blame - the designer who surfaced uncomfortable truths becoming the target for the discomfort those truths provoked. These were not incidental to the SCÖ experience; they were, as Fotaki and Hyde (2014) describe, organised ways of managing the anxiety that honest assessment would provoke, and the designer who insisted on materialisation became the container for the programme's anxiety about its own impossibility.
What I'm taking forward
I ended What I Learned at SCÖ with questions about design's role in impossible contexts, or contexts where it quickly becomes clear the limits of what is achievable, about how organisations maintain commitment to impossible or idealistic projects, about what happens when design artefacts expose rather than enable. Von Busch and Palmås don't answer these - but they reframe them as questions about power rather than questions about epistemology.
What I got right at SCÖ: the materialising practice. Forcing abstraction into specificity is what design does that programme management, funding applications often do not. The concept maps, the typed interface definitions, the Python testing - these were legitimate design contributions that produced accurate accounts of reality.
What I got wrong: the theory of change. I assumed that materialisation - making the real visible - was sufficient. That truth, well-presented, would generate its own political conditions. Von Busch is saying something uncomfortable but important: design's materialising power is only as effective as the political backing behind it. Without a power analysis - without asking who benefits from abstraction, whose position depends on the fiction, who controls the funding narrative - the most accurate artefact in the world is organisationally inert.
The social defences I described in the NORDES submission - idealisation, splitting, blame - are not separate from this political analysis; they are its psychodynamic expression, the mechanisms through which organisations manage the anxiety that power-literate materialisation would provoke. And the sacred quality that attaches to design's rituals and to the technoimaginaries within which it operates helps explain why fictional expectations resist the kind of profane technical analysis my artefacts offered: what is threatened by specificity is not merely an inaccurate claim but a communal identity built around the claim.
Hill (2012) told me that designers need to engage with "dark matter" - the governance, policy, and finance models that shape what services actually do. I thought I understood that. Von Busch and Palmås are telling me something sharper: it's not enough to see the dark matter. You need to understand who benefits from it being dark.
Ana and I are developing a paper about the metaphors through which the service designer role is understood in public sector contexts - drawing partly on the SCÖ experience, partly on her work at Experio Lab. I'm increasingly convinced that the metaphor problem is a power problem. When different stakeholders hold different metaphors for what the designer is supposed to do, each metaphor serves different institutional interests. The miracle worker metaphor I described last year - where the organisation expects the designer to make the impossible possible, then blames them for not delivering miracles - isn't just a misunderstanding. It's a power arrangement. But that's for the next piece.
References
Beckert, J. (2016). Imagined Futures: Fictional Expectations and Capitalist Dynamics. Harvard University Press.
Fotaki, M. and Hyde, P. (2014). Organisational blind spots: Splitting, blame and idealization in the National Health Service. Human Relations, 68(3), pp. 441-462.
Hill, D. (2012). Dark Matter and Trojan Horses: A Strategic Design Vocabulary. Strelka Press.
Hirsch, T. (2020). Practicing Without a License: Design Research as Psychotherapy. Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems, pp. 1-13.
Hochschild, A.R. (2012). The Managed Heart: Commercialization of Human Feeling. Updated edition. University of California Press.
Hood, C. and Heald, D. (2006). Transparency: The Key to Better Governance? Oxford University Press.
Matthews, T. (2021). Exploring Sacred Service Design. PhD thesis, The Oslo School of Architecture and Design.
Rip, A. (2012). The context of innovation journeys. Creativity and Innovation Management, 21(2), pp. 158-170.
Tonkinwise, C. (2010). "Politics Please, We're Social Designers". Core77, 1 September 2010.
Tonkinwise, C. (2023). "All Care, No Responsibility: Everything Service Designers Need to Know About Politics But Were Afraid to Ask". Guest lecture, MFA Products of Design, SVA.
Von Busch, O. and Palmås, K. (2023). The Corruption of Co-Design. Taylor and Francis.