A Good Building Can Still Make a Bad City

I have been thinking about why so much new development can look competent and still feel wrong.

The buildings are not necessarily ugly. The apartments may be well planned. The materials may be decent. There may be balconies, trees, bicycle parking, a landscaped courtyard and a sustainability certificate attached to the project. Nothing is obviously broken when the development is inspected building by building.

But when I arrive on foot, the place somehow falls apart.

The entrance is behind a parking lot. The ground floor is mostly garage doors, ventilation grilles, storage rooms or blank walls. The shortest route to the bus stop is blocked by a fence, so I have to walk around the entire site. There is a courtyard, but it feels private even when it is technically public. There may be hundreds of new homes, yet nowhere nearby to buy bread, sit outside, repair something, meet someone or do anything except enter an apartment.

Each building may be successful. The urbanism is not.

I think this happens partly because real estate development makes the project or the building the main unit of thought. The site has a clear boundary. The budget has a boundary. The financing, planning permission, construction contract and eventual sale all have boundaries. Inside those boundaries, everything can be measured and assigned.

Someone knows the exact number of apartments, square metres, windows, parking spaces and rentable surfaces. Someone has calculated how much every material costs and how each decision affects the return.

Then I cross the property line and the resolution drops.

The street becomes context. The surrounding buildings become context. The people who already live there become stakeholders, neighbours or objections. The routes through the area, the uses on the ground floor, the distance to ordinary destinations and the relationship between public and private space are considered, but often as external conditions around the main object.

This is the difference between thinking at the level of a building and thinking at the level of an environment.

Building-level thinking asks what can be placed on the site.

Environmental thinking asks what kind of street, block, route, edge and neighbourhood this placement produces.

The difference sounds small, but it changes almost every design decision.

A building can be designed as an object standing inside a site, or it can be designed as one wall of a street. An entrance can be positioned where it is convenient for the internal plan, or where it creates activity and direct access from the pavement. A courtyard can be treated as a private amenity for residents, or as one part of a larger sequence of public and semi-public spaces.

The same floor plan can participate in the city or withdraw from it.

Urban-design research gives some support to this distinction. Reid Ewing and Robert Cervero's meta-analysis of travel and the built environment found that travel behaviour is related not only to density, but also to street-network design, access to destinations, land-use diversity and proximity to transit. Density by itself is not enough. A large number of people can be placed in a development that remains disconnected, monotonous and difficult to navigate without a car.

This seems obvious, but a lot of development is still described through quantities that belong almost entirely to the site. How many homes were added? How much floor area was created? How efficiently was the land used? These are important questions, especially where housing is scarce, but none of them tells me what kind of urban environment was produced.

A dense residential tower beside a highway may contain more people than an older mixed neighbourhood, but it does not necessarily produce more urban life. The tower may be dense as a building while the surrounding environment remains sparse.

Research on walkability often returns to some combination of density, mix and access. There need to be enough people to support activity, enough different uses to create destinations, and a network that makes those destinations reachable. Remove one of these and the others become less useful.

This may be one reason developers and city officials sometimes talk past each other. The developer can point to the number of new homes. The urbanist is asking what kind of place those homes create.

The building count is easy to measure. The relationships are harder.

Even the ground floor reveals the difference.

From the perspective of the building, the ground floor is where many awkward things need to go. Parking entrances, waste rooms, transformers, storage, loading, ventilation, security systems and access cores all compete for space. From the perspective of the street, however, the ground floor is where almost the entire relationship between the building and the city happens.

A study of 588 block faces in New York found that active uses, ground-floor windows and street furniture were associated with greater pedestrian activity. Other research on active frontages has connected them with stronger perceptions of safety, comfort, liveliness and sociability.

This does not mean every street should be covered in shops. Retail cannot survive everywhere, and forcing commercial units into places without enough demand often creates another row of empty windows.

An active edge can also be made from homes, workshops, schools, entrances, visible shared rooms, small offices, gardens or stoops. The important thing is that the building has a relationship with the street. People can see out. People can enter directly. Something happens close to the pavement.

The opposite is a building whose ground floor functions as its technical underside. The architecture may be polished, but the street receives the back of the machine.

Street networks create another problem.

A development may have excellent internal circulation and still make the surrounding city harder to move through. It may be entered through one controlled driveway, with pedestrian routes that end at gates or property boundaries. To the developer, this can appear safe, legible and manageable. To everyone outside the project, it becomes a large obstacle they must walk around.

The project asks how residents enter the development.

Urbanism asks what new connections the development adds to the city.

These are not the same question.

Research on pedestrian movement repeatedly finds that connected networks, route choice, directness and access to destinations matter. A study comparing four Istanbul neighbourhoods found that street-network configuration and directional accessibility helped explain pedestrian flows alongside land uses and streetscape features. Research operationalising Jane Jacobs's ideas in Seoul has similarly tested how mixed uses, block size and building concentration contribute to pedestrian activity, while also showing that the relationships are contextual rather than universal laws.

But connectivity is difficult to produce one isolated development at a time. A route through one site may only become useful when a later project continues it. A street may need to connect across several properties owned by different people. The urban benefit exists between projects, while responsibility is divided inside them.

This may be where low-quality urbanism actually comes from.

It is not always caused by individual incompetence. It may emerge because every participant is doing a reasonable job within a boundary that is too small.

The architect works on the building. The traffic engineer makes the vehicle access function. The landscape architect fills the remaining open space. The developer makes the project financially viable. The city checks whether it follows the rules. The future building manager tries to reduce maintenance costs and control access.

Nobody necessarily decides to produce a disconnected or lifeless environment.

But nobody is fully responsible for the relationships between the parts.

Research on the projectification of urban change describes a similar problem. Projects make complex work manageable by turning it into bounded interventions with budgets, timelines, outputs and completion points. This makes action possible, but it can also favour things that are easy to deliver and measure over changes that depend on long-term coordination, institutional learning and relationships across multiple sites.

A building fits naturally inside project logic.

A neighbourhood does not.

The neighbourhood exists before the project and continues after it. Its streets may last longer than any of the individual buildings. Its social and commercial patterns emerge from many decisions that were never planned together. The consequences of one development may only become visible when the next five are completed.

By then, the original project team may be gone.

This is why I no longer think the main distinction is between good and bad architecture. It is possible to have carefully designed architecture and weak urbanism at the same time.

The more useful distinction is between designing the building as the final object and designing it as material for a larger environment.

A good piece of city may require the individual building to give something up. A little rentable area may be lost to a public route. The most efficient placement may be rejected because it produces a worse street. Parking may need to be less convenient. A courtyard may need to connect to something outside the development. A ground floor may need more doors than the internal circulation strictly requires.

These decisions can look inefficient when the project is viewed alone.

They may only make sense when the unit of evaluation is larger than the property.

There is an important limit to the research behind this argument. Pedestrian counts, walking behaviour and perceptions of public space are not the same thing as total urban quality. A place can be walkable and still be unaffordable, exclusionary, inaccessible to disabled people, ecologically destructive or hostile to the people who were there before. Strong urbanism cannot be reduced to active frontages and connected streets.

But the research does show something more modest and useful. The performance of an urban environment cannot be inferred from the quality or quantity of the individual buildings inside it. It emerges from relationships among buildings, routes, uses, public edges and destinations.

That is probably the central problem. The project boundary is useful for building things, but it is a poor boundary for understanding what has been built.

A developer delivers buildings. People live in environments.

Low-quality urbanism is what happens when every building is treated as a successful individual project, while the relationships between the projects remain nobody's main responsibility.

Connected reading

• The City Is an Interface
• The Best Products Aren't Apps. They're Cities.
• The Smart City Breaks Between the Systems
• The AI Future Is Becoming a Land-Use Problem
• The City Is a Matchmaker

AI authorship note

This essay began with Kris Haamer asking whether people in real estate development mostly think in terms of projects, then correcting the argument toward building-level thinking versus environmental thinking. AI supported the urbanism literature review, source verification, drafting, editing and implementation. Kris narrowed the thesis, rejected the broader systems framing and approved the article for publication.