Most service design begins with a persona that doesn't exist. The mythical average user—moderately tech-literate, modestly able-bodied, statistically forgettable—anchors decisions across healthcare systems, government portals, and financial platforms. The result is services that work passably for many and brilliantly for none.

There is a quieter tradition in design that inverts this logic. Instead of optimising for the centre of a distribution, it begins at the edges: the user with low vision navigating a banking app, the parent managing a complex care plan from a hospital corridor, the migrant worker filing taxes in their third language. These are not exceptions to design—they are its most rigorous test conditions.

What this tradition reveals is counterintuitive. Designing for extreme users rarely produces specialised solutions. It produces better baseline services. The curb cut, originally legislated for wheelchair access, now serves cyclists, parents with strollers, and delivery workers. The accommodation became the standard. This pattern repeats across domains, and yet most service organisations continue to treat extreme users as an afterthought, a compliance requirement, or a separate workstream entirely. The question worth examining is why this approach persists, and what changes when designers reverse the assumption.

An extreme user is not a difficult user. They are a high-resolution diagnostic instrument for a service system. Where an average user might tolerate friction, an extreme user encounters it as failure. The system either accommodates them or excludes them, and that binary clarity exposes design assumptions that remain invisible under normal conditions.

Consider hospital discharge processes. For a patient with strong family support, English fluency, and stable housing, even a confusing discharge summary becomes navigable. The slack in their personal system absorbs the slack in the institutional one. For a patient lacking those resources, the same document is an obstacle that produces readmission, missed medication, or worse. The discharge process didn't change—the diagnostic conditions did.

This is why extreme users reveal opportunities that user research with mainstream populations cannot. They surface dependencies, assumptions, and hidden labour that the service implicitly outsources to the user. The well-resourced user pays for poor design with mild inconvenience. The extreme user pays for it with exclusion.

Strategically, this reframes edge cases. They are not outliers to be handled after launch. They are the conditions under which a service must be proven viable. Designing for them first forces decisions about what the service actually does, rather than what its interface suggests it does.

The methodological shift is significant. Instead of asking what the majority of users need, the designer asks under what conditions the service breaks. The answer is rarely about features. It is almost always about the assumptions embedded in language, sequence, timing, and trust.

Takeaway

Extreme users are not exceptions to your service—they are stress tests. What looks like an edge case is usually the clearest view of your design's hidden assumptions.

Inclusive design has a property that often surprises its critics: accommodations made for specific populations tend to generalise upward into improvements for everyone. This is not a happy accident. It reflects something structural about how constraints sharpen design.

When a team designs captions for deaf users, they produce a feature now used by commuters in noisy trains, language learners, and anyone watching video in a quiet office. When a government agency rewrites forms for users with limited literacy, comprehension rates rise across all education levels. The accommodation becomes the affordance.

The mechanism is straightforward. Mainstream users are not maximally capable—they are merely capable enough to compensate for poor design. Cognitive load, fatigue, distraction, and context reduce everyone's effective ability at various moments. A service designed for the user at their worst is a service that works for the user at their average.

This challenges a common organisational defence: that inclusive design serves a small population and therefore deserves proportional investment. The arithmetic of accessibility rarely works that way. Investments at the edge return value through the middle, particularly in services where comprehension, trust, and completion determine outcomes.

What this requires institutionally is a willingness to treat inclusion as design strategy rather than compliance. The compliance frame produces minimum viable accommodation. The strategic frame produces services where the so-called accommodation is indistinguishable from the core experience—because it is the core experience, refined under harder conditions.

Takeaway

Accommodations rarely stay confined to their original users. The features designed for someone's worst day quietly become everyone's best tools.

Centring extreme users does not require abandoning mainstream ones. It requires sequencing the design process differently and using specific methods that keep both populations visible. The risk of inclusive design, when poorly practised, is producing services optimised for edge cases that feel alien to the majority. The discipline lies in avoiding both errors.

One practical method is stratified prototyping: developing service flows that are tested first with users at the constraint extremes, then validated with mainstream populations. The order matters. Testing with the mainstream first produces designs that subsequently require retrofitting. Testing with extremes first produces designs that mainstream users find unremarkable—which is precisely the goal.

Another is the use of capability journey maps that document not only what users do, but what cognitive, physical, linguistic, and temporal resources each step demands. When a step demands resources that a portion of users predictably lack, the step is a candidate for redesign rather than for help text or support staff.

Service blueprints can be extended to include what might be called shadow labour: the unpaid work the service implicitly assumes users will perform, such as translation, advocacy, scheduling, or interpretation of policy. Extreme users perform this labour visibly because they cannot avoid it. Mainstream users perform it invisibly, which is why it often goes unmeasured.

Throughout, the practitioner's task is to resist two temptations. The first is exoticising extreme users as inspirational case studies divorced from operational reality. The second is collapsing them back into the mainstream once initial empathy work is complete. Neither produces durable design. The work is sustained attention to who the service excludes, asked repeatedly, at every stage of development.

Takeaway

Designing from the edges inward is a sequencing decision, not an ideological one. The order in which you test determines the service you ultimately build.

The argument for extreme user design is not primarily moral, though moral arguments can be made. It is methodological. Services designed at the edges of human capability and circumstance are more honest about what they actually require, more robust under variation, and more useful across the populations they serve.

What stands in the way is rarely technical. It is the persistence of the average user as an organising fiction—a convenient abstraction that lets institutions design for conditions that no actual person fully inhabits. Letting go of that fiction is uncomfortable. It means acknowledging that services are political artefacts, distributing ease and difficulty unevenly across the populations they touch.

The strategic designer's contribution is to make that distribution visible and, where possible, to redesign it. Not through grand gestures, but through the slow work of asking who breaks against this service, and what the service might become if it stopped breaking them.