Owning the Problem Space: Lessons for Practice
Across this series I have developed an argument in three parts: that planning and design are fundamentally different activities, distinguished by whether the problem space within which work proceeds is given or must be constructed; that service design lacks the formal apparatus - the grammars, the compositional rules, the validity criteria - that would make its designs precise and, increasingly, executable through algorithmic technology; and that public sector transformation programmes frequently fail because they attempt planning, procurement, and governance without having first done the design work that would make those activities meaningful. This post attempts to draw practical lessons from the framework, and in doing so, to be honest about where those lessons run out.
The Core Insight
The central claim of this series, developed with reference to computational planning theory and refined through the military design literature (US Army and Marine Corps, 2006), is that planning and design address different kinds of problems and produce different kinds of outputs. Planning, as the earlier post on state spaces argued, is fundamentally an activity of navigation; given a sufficiently well-defined problem - one in which the possible states, the transitions between them, and the criteria for goal attainment are known - the planner's task is to find a path through that space. Design, by contrast, is an activity of construction; given an ill-defined situation in which the relevant concepts, boundaries, and criteria for success are not yet established, the designer's task is to build the problem space within which planning can subsequently operate (Rittel and Webber, 1973; Zweibelson, 2023).
Most service design situations require design before they permit planning. The institutional pressures, however, run in the opposite direction: timelines, milestones, deliverables, and the governance rituals that accompany them all presuppose a defined problem space. The result, as the preceding posts have argued, is premature planning - plans that assume problem spaces which have not been constructed, and which therefore navigate confidently through territory that may not exist.
Owning the problem space, in the sense intended here, means resisting that pressure long enough to do the design work: to construct the conceptual framework, to surface the constraints, to test whether the apparent consensus about what success looks like is genuine or merely convenient, before committing to a plan that the framework has not yet been built to support.
What Owning the Problem Space Requires
The SCÖ project, discussed in several earlier posts, remains the clearest case in this series of what happens when the problem space is not owned. The Pathway Generator assumed a state space - patient pathways through rehabilitation services - without constructing one. The technology was, by any reasonable measure, sophisticated; the intentions of the project team were sincere; but the design work that would have grounded the technology in operational reality had not been done. What states could patients actually occupy within the Swedish rehabilitation system? What transitions between those states were available, given the organisational boundaries, the data fragmentation across four separate systems with no shared identifier, and the professional resistance that any cross-organisational pathway would encounter? What constraints - legal, technical, cultural - would delimit what was achievable? The algorithm optimised within a problem space that did not correspond to the situation it was intended to address; the algorithm, considered in isolation, worked, but the design within which it was embedded did not.
Owning the problem space would have meant starting differently. Before specifying the algorithm, the work would have been to surface the implicit assumptions embedded in the brief - to ask what "rehabilitation pathway" means, operationally, when patients move between organisations that do not share data, terminology, or professional norms. It would have meant documenting constraints alongside aspirations: not "we want to optimise pathways" but "we want to optimise pathways given these specific conditions of data fragmentation, consent requirements, and capacity limitations". It would have meant testing whether stakeholders actually agreed on what success looked like, or whether the apparent consensus around "better outcomes" masked fundamentally different priorities - as Bailey and Lloyd (2015) observe of policymaking environments, sound ideas on their own are never enough where complex power dynamics make it difficult for any party to be fully transparent about what they want. A design specification that proceeds without reference to the constraints within which it must be enacted is, in a meaningful sense, not a specification at all; it is an aspiration.
This is not, however, merely a matter of individual skill or professional diligence. Individual practitioners can, and sometimes do, ask the right questions, surface assumptions, contest the brief. But doing so means slowing the timeline, introducing uncertainty into a process that rewards the appearance of certainty, and, as Boztepe (2023) documents, encountering the "tissue incompatibility" that arises when design's ways of working meet the institutional logics of the public sector. Every project carries implicit assumptions about its state space - "improve user experience", "integrate services", "deliver digital transformation" - and surfacing those assumptions frequently reveals that the apparent consensus masks fundamental disagreement about what the concepts mean. The value of formal thinking - enumerating states, specifying transitions, articulating promises in the sense developed by Burgess (2015) - is that it forces this confrontation; the exercise exposes gaps and ambiguities that less formal methods leave comfortably vague.
The frugal algorithms work at FOI, discussed in the preceding post, demonstrates the alternative starting point. In military edge computing contexts, constraints come first: what data actually exists, what compute is available, what connectivity can be relied upon; given these realities, what capability is achievable? This constraint-first orientation is, in my experience, the inverse of how most public sector design projects begin.
Why It Rarely Happens
If owning the problem space is as consequential as this series has argued, the question of why it so rarely happens deserves a structural answer, not merely a lament about individual failings. Hay and Vink (in press) frame the issue with precision: the problem is not that individual designers lack reflexivity about power dynamics - it is that contextual factors impede them from acting on whatever reflexivity they possess. The governance apparatus does not merely fail to support design; it actively selects against it.
NPM governance structures are the primary mechanism through which this selection operates. A business case requires quantified benefits before the design work that would identify real benefits has been done, so benefits are invented to satisfy the template. Stage gates demand evidence of progress at fixed intervals, but design produces understanding rather than deliverables, and understanding does not map onto a milestone table. Clarke and Craft (2018) make the related point that design thinking, for all its justified criticism of traditional policy design for insufficiently attending to citizen experience, itself falls short as a comprehensive approach by failing to account for the institutional and governance structures within which policy is made and enacted. The planning/design distinction, if it is to have practical force, needs to be institutionalised, not merely understood by individuals; without institutional support, design work gets governed as planning work, with planning's expectations, timelines, and success criteria applied to an activity they do not fit.
The deeper difficulty is that design, done honestly, threatens. Owning the problem space means being transparent about what is not known - and organisations, as the earlier post on boundary objects argued, are accomplished at absorbing information without responding to it. Hill (2012) is characteristically direct: designers, he argues, too often appear naive in the face of genuinely understanding cultures of decision-making, of how an organisation or sector actually works. The institutional structures that resist design - what Hill terms "dark matter", encompassing organisational culture, policy environments, market mechanisms, legislation, and finance models - are the forces that shape what services actually do, but they sit beneath the surface of what design processes typically engage with. Bailey (2021) offers a more unsettling analysis: design in government contexts has been discursively and practically remodelled to align with existing institutional preoccupations, such that what is called "design" in these settings may already have been shaped to avoid the confrontations that genuine problem-space ownership would require. A journey map will not change a procurement framework; a service blueprint will not shift a governance model; and if the design practice has already been co-opted to fit the governance logic it ought to challenge, the question of what design can actually reach becomes considerably more difficult.
What the Framework Cannot Reach
This is where honest reflection is necessary. The framework developed across this series - state spaces, grammars, planning versus design, constraints - has, I believe, genuine analytical utility. It provides a vocabulary for explaining why the SCÖ project failed, why technomagic persists as an institutional logic, and why NPM governance structures select against the design work that transformation programmes require. But it also has limits, and it would be inconsistent with the series' own argument about the importance of naming constraints to hedge around them.
The first limit is that formalism can become its own kind of technomagic. Proposing a grammar of services - precise vocabularies, compositional rules, validity criteria - carries the same risk that the series identifies in technology procurement: the belief that the right representational tool will solve a problem that is not primarily about representation. Embley and Thalheim (2012) note that specification is frequently restricted to syntactical elements, and that this restriction works adequately only for simple applications; Stolterman and Nelson (2014) argue more sharply that approaches which aim to be comprehensive often lead to problematic oversimplifications. A GP appointment booking can, as the grammar post demonstrated, be specified grammatically without too much strain. A multi-agency child protection service - where the state space is genuinely contested, where professional judgements conflict, where the stakes of specification errors are catastrophic - probably cannot, at least not in any way that would clarify rather than obscure.
The second limit is that the framework treats power as context rather than as content. Blomkvist and Clatworthy (2023) acknowledge that the possibility of foreseeing use is always limited, and likewise the possibility of grasping the dynamics of power relationships within which services operate. A grammar can specify states, transitions, and promises, but it cannot specify who gets to define the states, whose transitions count, whose promises are honoured. These are political questions, not formal ones. Kimbell (2021) identifies a structural feature of how design operates in social contexts: acknowledgement of the political and economic conditions that sustain institutional forms tends to be muted, downplayed in favour of the innovation logic that design more comfortably inhabits. Von Busch and Palmås (2023) put it more forcefully: designers need to engage with the negative valence of power politics, with the reality that co-design and participatory processes unfold in an imperfect world characterised by betrayal, co-option, and drift. Penin and Prendiville (2025) call for a critical re-evaluation of service design as the design of work itself, questioning whether the field has the capacity to address the converging political, economic, and technological forces it claims to design for. The grammar sketched in this series assumes that the primary bottleneck is representational - that if services could be specified more precisely, they would be designed more effectively. But the bottleneck may not be representational at all; it may be political, and no amount of formal precision will address a fundamentally political problem.
The third limit is that the planning/design distinction, powerful as it is as a diagnostic lens, can become its own orthodoxy if applied without judgement. Not every failure is a planning/design confusion; not every problem requires the construction of a new state space. Sometimes the problem space is well-enough understood and the failure is one of execution, resourcing, or political will. The temptation to see planning/design confusions everywhere is itself a failure of the situated judgement that the framework is supposed to develop - and, as Kimbell (2009) argues, descriptions of design thinking that focus on individual cognition fail to account for the situated, collective, and material nature of knowledge production in design practice.
Service design's tools remain what Meroni and Sangiorgi (2012) describe as 'mainly informal and tacit'; Blomkvist and Clatworthy (2023) argue for the development of a 'service design language'; Matthews (2021) demonstrates that even the blueprint - the field's most structured representational tool - models delivery structure without conveying the experiential or socio-cultural dimensions the service claims to address. The gap is real, and developing formal apparatus would help to close it. But it would help with the representational dimension of the problem, not the political one, and it would be dangerous to mistake one for the other.
For the Field
These limits do not invalidate the framework so much as locate it. The planning/design distinction remains a powerful diagnostic for transformation failures; the grammar proposal remains worth pursuing as a means of making service design's tools more rigorous and its specifications more amenable to computational reasoning; constraint-first thinking, as the frugal algorithms work suggests, remains undervalued in a field that tends to foreground aspiration. But the field needs to pursue these alongside - not instead of - the harder questions about power, governance, and the institutional structures within which design is permitted to operate.
Service design can learn from adjacent fields: from computer science, the principles of formal specification and state machines; from systems engineering, requirements analysis and architecture frameworks; from organisational studies, institutional theory and sensemaking; from military design, problem framing and adaptive planning. These fields have thought deeply about problems that service design faces but has not always engaged with on their terms. Equally, however, service design needs to develop its own critical tradition - one that studies failure as rigorously as it showcases success, that asks where its methods break down rather than only where they succeed, and that is honest, in the manner that Bailey (2021), Kimbell (2021), and von Busch and Palmås (2023) variously demand, about the gap between what it claims to design for and what its tools and institutional positioning can actually reach.
At the scale of programme management, organisational strategy, and policy formation, the planning/design confusion becomes simultaneously more consequential and harder to address. Governance structures at this scale are more entrenched; institutional pressures heavier; individual reflexivity, however well-developed, runs into Hill's dark matter. The field needs approaches that work at this level - not only project methods but ways to influence the governance apparatus itself. That, as Clarke and Craft (2018) imply, is a political project as much as a design one.
Reflections
Writing this series was itself, in some sense, a design exercise. I began with a set of questions - why did the SCÖ project fail? What do AI planning and military design have to do with service design? What would a formal service specification look like? - and through the process of writing, constructed a problem space: a vocabulary and framework for thinking about those questions that did not exist, for me at least, before the series began.
The framework is not complete, and the preceding section has tried to be candid about where it breaks down. The planning/design distinction, for all its diagnostic power, risks becoming the kind of totalising lens that flattens the situated complexity it is supposed to illuminate; Kimbell's (2009) caution about design thinking's tendency to privilege individual cognition over collective, material practice applies as much to this framework as to any other. The grammar proposal, for all its appeal as a means of making service design's tools compose and generate, confronts the difficulty that the most consequential service design problems are precisely those where the state space is most contested, where power most determines what gets specified and what gets left out. Formalism would help - but sustained theoretical refinement, tooling, and the cultivation of communities of practice (Wenger, 1998) willing to test and extend these ideas would be needed before it could help in any practical sense.
What the series has reinforced, perhaps above all, is the significance of constraints as design material. Design discourse tends to foreground aspiration, vision, and creative possibility; but the work at FOI, and the retrospective analysis of SCÖ, both suggest that capability emerges from an honest engagement with constraints at least as much as from the articulation of ambitions. That observation may sound modest, but it runs against the grain of a field that has historically marketed itself on the promise of creative transformation - and, as Bailey (2021) documents, has been discursively reshaped in government contexts to amplify precisely that promise.
References
Bailey, J. (2021). Governmentality and Power in 'Design for Government' in the UK, 2008–2017. PhD thesis, University of the Arts London.
Bailey, J. and Lloyd, P. (2015). A View From The Other Side: Perspectives On An Emergent Design Culture In Whitehall. In Proceedings of the 11th European Academy of Design Conference.
Blomkvist, J. and Clatworthy, S. (2023). The Materials of Service Design. The Design Journal, 26(5), 671–690.
Boztepe, S. (2023). Design Making Its Way To The City Hall: Tensions In Design Capacity Building In The Public Sector. Design Studies, 88.
Burgess, M. (2015). Thinking in Promises. O'Reilly Media.
Clarke, A. and Craft, J. (2018). The Twin Faces of Public Sector Design. Governance, 32(1), 5–21.
Embley, D. and Thalheim, B. (2012). Handbook of Conceptual Modeling. Springer.
Gärdenfors, P. (2000). Conceptual Spaces. MIT Press.
Harel, D. (1987). Statecharts: A Visual Formalism for Complex Systems. Science of Computer Programming.
Hay, A.F. and Vink, J. (2024). Why Do Service Designers Struggle to Address Power Dynamics? Strategic Design Research Journal, 17(1).
Hill, D. (2012). Dark Matter and Trojan Horses: A Strategic Design Vocabulary. Strelka Press.
Iqbal, M. (2018). Thinking in Services. BIS Publishers.
Jackson, A.P. (2024). Military Design Thinking: An Historical and Paradigmatic Overview. Routledge.
Kimbell, L. (2009). Beyond Design Thinking: Design-as-Practice and Designs-in-Practice. In CRESC Conference.
Kimbell, L. (2021). Logics of Social Design. Design Issues, 37(1), 76–91.
Matthews, T. (2021). Exploring Sacred Service Design. PhD thesis, Lancaster University.
Meroni, A. and Sangiorgi, D. (2012). Design for Services. Gower.
Penin, L. and Prendiville, A. (eds.) (2025). The Bloomsbury Handbook of Service Design. Bloomsbury.
Rittel, H. and Webber, M. (1973). Dilemmas in a General Theory of Planning. Policy Sciences, 4(2), 155–169.
Stolterman, E. and Nelson, H. (2014). The Design Way: Intentional Change in an Unpredictable World (2nd ed.). MIT Press.
US Army and Marine Corps (2006). FM 3-24: Counterinsurgency. Department of the Army.
Von Busch, O. and Palmås, K. (2023). The Corruption of Co-Design. Routledge.
Wenger, E. (1998). Communities of Practice. Cambridge University Press.
Wilkinson, L. (2005). The Grammar of Graphics. Springer.
Zweibelson, B. (2023). Beyond the Pale: Designing Military Decision-Making Anew. Routledge.