In 2012, Marc Stickdorn and Jakob Schneider codified five principles for service design in their book, This is Service Design Thinking: user-centered, co-creative, sequencing, evidencing, and holistic. These have become the shared foundation upon which a generation of practitioners has built its work. These principles are sound — I am not here to dismantle them.
I am here to ask a question that, as far as I can tell, no one else is asking: at what scale do these principles actually hold true?
The Pattern
At Estée Lauder, I conducted Jobs-to-be-Done research around “gifting”. We reframed the entire problem: gifting was not about packaging or personalization (the industry standard), it was about fulfilling social obligations. Customers were trying to find meaningful, affordable products that would ultimately improve their relationships. The reframe was profound and we generated ideas to scaffold the full gifting journey. Stakeholders loved it. The research was rigorous, the insights were deep, and the concepts were actionable.
Then… nothing happened.
When it came time to discuss implementation, suddenly priorities shifted. No one in a position to act was willing to take the risk. Momentum stalled; larger, more “strategic” initiatives consumed the attention and budgets of the project sponsors. And so, the work was packaged and shelved, like a high school trophy. This is the pattern that many corporate service designers should recognize: a promising start, exciting innovative concepts, and limited follow-through.
I used to think this was a courage problem — that if the right leader would simply commit, we could make an impact. After seeing firsthand this pattern repeat across every large organization I have worked in, I now believe it is a scale problem. The distance between decision-makers and the people who actually experience a service is so vast that accountability effectively dissolves. When no single person owns the outcome, every person can reasonably conclude that the failure belongs to someone else. No one is going to take a risk (no matter how much affirmative research is produced) when they will not personally feel the consequences of their inaction or of getting it wrong.
The Counter-Case
Let’s compare that to my experience at a credit union when it had fewer than 500 employees.
There, my team and I had to earn the right to propose something new. We reduced waste and variability in the call center first, optimizing call handling and eliminating unnecessary volume. That work paid for itself, which built trust, and that trust earned my team room to innovate. The dynamic was fundamentally different: my ability to do future work depended on whether my current work actually helped.
The difference was not simply organizational culture. It was scale. We could speak to the actual members who were using what we built. When something did not work, we heard about it directly — not from a quarterly NPS report filtered through three layers of management. Feedback loops were robust and efficient because they were small.
During that time, my team also redesigned the debit card mailer. I was shocked when, recently, I received a replacement debit card in the mail. It was the same one that we had designed seven years earlier! My team also built a streamlined wire transfer process (also still in use). Not because these were brilliant designs, but because they were built with the people who used them, by someone who had a stake in whether or not they worked and who would feel the consequences if they did not.
I do not know what happened to the gifting research at Estée Lauder. I doubt anyone does.
The Unexamined Assumption
Stickdorn and Schneider’s five principles are presented as universal. If you apply them faithfully, you will produce better service experiences — whether at a neighborhood clinic or a federal agency, a family-owned restaurant or a multinational corporation.
But universality implies infinite scalability, and that assumption has largely gone unexamined.
User-centered assumes you can meaningfully understand your users. At what point does “understanding users” become “constructing personas from aggregate data”? There is a fundamental difference between designing for someone you have spoken with and designing for a segment. At Estée Lauder, I had research. At the credit union, I had relationships. The principle worked in one context and not in the other.
Co-creative assumes genuine collaboration with stakeholders. But meaningful co-creation has a participation ceiling. At scale, it becomes “representative” input — workshops with selected participants, surveys with statistical significance, and sticky notes from people you may never see again: the voices are sampled but not heard. The output at scale is functionally diluted beyond recognition.
Sequencing assumes you can map the service journey. At human scale, a journey is a lived experience you can observe and walk alongside. At large scale, a journey map is an abstraction — an averaged path assembled from data fragments that no single person actually walks. “The map is not the territory,” and the further you zoom out, the less it resembles anything real.
Evidencing assumes that making the intangible tangible always helps people. It can. But evidence can also entrap. A simple service agreement between people who know each other is evidence. A forty-page terms-of-service document is also evidence, but it’s also construed as alienating CYA material. One extends human capacity; the other subordinates or disables it.
Holistic assumes you can apprehend the whole system. You can hold a neighborhood business in your head — the owner, the regulars, the staff, the rhythms, the community context. You cannot hold a multinational corporation in your head. At that scale, “holistic” means “we produced a systems diagram” which is a representative abstraction, not a comprehension. There is a hard limit on holistic understanding, and it is smaller than the field wants to admit.
A fair objection here is that the principles are meant to function together, not in isolation. Evidencing without a user-centered lens should produce poor artifacts — that is the point of applying them together. But this interdependence actually deepens the scale problem rather than resolving it. Coordinating five principles simultaneously requires tight feedback, shared context, and ongoing communication among everyone involved. These conditions exist naturally at small scale. At large scale, they must be engineered and enforced and they are usually the first things to break down. For example, at Estée Lauder, the work WAS user-centered and co-creative, but the holistic and sequencing principles collapsed because the organization couldn’t coordinate across all five simultaneously.
None of this means that the principles are wrong. It means they have a threshold — a point beyond which they stop influencing practices for the better. The field has not identified where that threshold is, and in failing to do so, it has quietly accepted that the principles can function at any scale. They cannot.
Five Heuristics for Scale
I want to propose a set of practical tests. These are not derived from theory alone — they come from my own experience of working at scales where the principles held and scales where they did not.
If the service fails, do I experience that failure materially (through lost trust, lost relationships, lost revenue)? Or am I insulated by organizational layers, moving to the next project before the last one is implemented? Skin in the game is not a nice-to-have; it is what makes the difference between principles that function and poster art.
Not personas. Not segments. Not “users.” Can I name the specific individuals who will experience what I am designing? Can I call them and ask them how it is going? If the answer is no, “user-centered” has become an abstraction — useful perhaps, but fundamentally different from the real thing.
If a service breaks on Monday, how long before anyone with the power to change it finds out? Days? Weeks? Quarters? In one context, I could hear about a problem from a member and adjust within the week. In the other, information traveled slowly through research reports, stakeholder presentations, prioritization committees, and roadmap negotiations — a journey measured in months, if it arrived at all. The length of the feedback loop is a direct measure of the system’s capacity to adapt and improve. When that loop stretches beyond a certain point, the principles cannot self-correct, and design quality degrades regardless of the talent involved.
Co-creation can mean “I facilitated a workshop and collected input.” Or it can mean “the people affected by this service are making real decisions about how it works, and they retain the power to change it after I am gone.” The test is whether the community’s capacity to address its own problems is potentially greater after my involvement — not whether they have become more dependent on the solution I built.
As Richard Normann distinguished, every service has two possible trajectories. It can enable people to act — to solve their own problems, serve their own customers, and strengthen their own relationships. Or it can relieve them of those capacities, creating dependency where there was once competence. This is not hypothetical. Consider what large-scale customer service platforms have done to the simple act of one person helping another. It’s a relief these days to talk to a human and not a phone tree. The question is not whether the service works. The question is what it does to the people it claims to be serving.
Five Principles, Refined
The heuristics above are diagnostic — they help you to determine whether you are operating at a scale at which the principles hold. But the principles themselves also need refinement. What follows is a set of updated principles for service design that takes the question of scale seriously, informed by the work of Leopold Kohr, Ivan Illich, E.F. Schumacher, Kirkpatrick Sale and Donella Meadows — economists and social critics, writing between the 1950s and 1990s, who spent their careers arguing that the most important design decision is often how big (or small) something should be.
Service design must center actual people — not abstractions derived from aggregate data. This requires that the designer can know, name, and maintain ongoing relationships with the people affected by the service. When the number of people affected exceeds the designer’s capacity to comprehend them as individuals, the designer must either reduce scope or honestly acknowledge that the work has shifted from person-centered to model-centered. Both can produce value; they should not be confused with one another.
Collaboration is only genuine when every participant has voice, agency, and shared stakes in the outcome, but this has a natural ceiling. When collaboration requires “representative” stakeholders or sampled input, it has exceeded participatory scale. The standard should not be “were stakeholders consulted?” but “do the people affected by this service have an ongoing and meaningful opportunity to shape it?” Not everyone will choose to participate — and that is fine. The principle is not that every service must demand involvement, but that the option for genuine participation must exist, and that the people affected retain the power to exercise it when they choose. Co-creation without that option is extraction with better manners.
A service is experienced as a sequence, but it exists within a system of feedback loops, delays, and non-linear dynamics. Mapping the sequence is necessary; mistaking the map for the system is dangerous. Furthermore, the tighter a sequence is scripted, the less room remains for autonomous human judgment. The sequence should function as flexible scaffolding — supporting the people within it — not as a conveyor belt that reduces them to units.
Making a service tangible through artifacts is valuable, but every artifact carries a question: does this object extend the person’s capacity to act independently, or does it bind them more tightly to the institution that produced it? A membership card that grants access is empowering. A call center scripting document that dictates behavior is entrapping. Service designers must evaluate their artifacts not only for clarity and utility, but for what they do to the autonomy of the people who use them.
Genuine holistic understanding has a limit. Within a bounded system — a neighborhood, a small organization, a specific community — it is possible to apprehend the whole: the relationships, the environment, the rhythms, the interdependencies. Beyond that boundary, holism becomes modeling, and models are always incomplete. The principle should not be “see the whole system,” but “see this system, whole.” The choice of scale is itself a design decision, perhaps the most consequential one.
Co-creation without the option for genuine participation is extraction with better manners.
Where This Leads
I am not writing this as an academic exercise. I am building a service design practice — Convivial, a worker-owned cooperative in Raleigh, North Carolina — on the premise that these refinements matter. I believe that service design practiced at a human scale, with real accountability, for people you actually know, produces outcomes that large-scale design simply cannot.
I should be clear about what I am not saying. I am not saying that large-scale service design is unnecessary, or that governments and healthcare systems should stop trying to design better services for millions of people. Those are real problems that demand serious attention. What I am arguing is that the principles the field relies on behave differently at different scales, and that is a real blind spot. Figuring out how to make these principles function at institutional scale is a separate project. This essay is about recognizing that the problem exists.
I am beginning a listening tour with small business owners in the Triangle not to sell them anything, but to understand their challenges from the inside. These are challenges that operate, almost by definition, at a scale where the five principles can function as genuine practices rather than corporate aspirations.
The service design field has drifted toward the largest possible clients (government agencies, corporations, healthcare systems) because that is where the budgets and prestige are. It is also where the principles quietly break down, where promising service design work is rejected by the corporation’s immune system, and where beautifully crafted journey maps end up archived in SharePoint folders at best, and as wallpaper at worst.
There is another way to practice. It is smaller, messier, closer to people, and far less glamorous. I believe it is also where the work actually holds.