Articles
Urban Development
May 19, 2026
.png)
A floor area ratio of 1.5 sounds precise. It fits neatly into a zoning document, a municipal target sheet, or a developer's investment memo. But FAR alone says almost nothing about what a neighborhood will actually feel like. Two projects with identical FAR can produce radically different outcomes in terms of daylight, courtyard quality, building depth, and the experience of walking through the site at street level.
This gap between density as a number and density as a spatial reality is where most early-stage planning mistakes happen. Not because teams lack ambition, but because the tools they use to explore density operate at the wrong resolution.
Municipal planning documents typically express density targets through a handful of metrics: floor area ratio, site coverage ratio, maximum building height, and sometimes a unit count target. These numbers define a regulatory envelope. They do not define a neighborhood.
Consider a 15,000 square meter site with a FAR target of 1.2 and a maximum site coverage of 40%. That translates to roughly 18,000 square meters of gross floor area distributed across 6,000 square meters of ground coverage. The arithmetic is simple. The spatial question is not.
Those 18,000 square meters could be delivered as three-story lamella rows covering the full 40% footprint, producing wide, open spaces between buildings but limited height variation. Or they could be delivered as six-story perimeter blocks covering only 20% of the site, creating enclosed courtyards with strong spatial definition but taller facades. Or as a hybrid of point towers and low-rise base buildings, concentrating density in specific nodes while leaving large portions of the site open.
Each configuration meets the FAR target. Each produces a fundamentally different place. The density number is the same. Everything else is different.
In a traditional early-stage workflow, an architect or planner sketches a single massing concept for the site. That concept reflects their experience, their instinct for what works, and the specific constraints they happen to prioritize. If the concept meets the density target and looks reasonable, it moves forward. If it doesn't, it gets adjusted until it does.
This process is not wrong. It is incomplete. A single concept tested against a density target produces one data point in a solution space that contains hundreds of valid configurations. The team commits to a direction without knowing what they are leaving on the table, or what trade-offs they are making implicitly.
The perimeter block layout might deliver better courtyard quality but worse direct sun hours on ground-floor apartments. The lamella layout might maximize sunlight but create wind corridors between buildings. The hybrid approach might offer the best open space metrics but require more complex structural engineering. These trade-offs exist whether or not the team explores them. The question is whether they are made consciously or by default.
When density exploration moves from manual sketching to constraint-aware generation, the question changes. Instead of asking whether a specific concept can hit the density target, the team asks what the full range of valid solutions looks like and where the interesting trade-offs sit.
Define the site boundary, the height envelope, the coverage limit, the setback requirements, and the noise constraints. Let the generation engine produce dozens or hundreds of configurations that satisfy all of these simultaneously. Then compare them on the metrics that actually matter for livability: courtyard dimensions, facade-to-facade distances, sunlight penetration, open space continuity, parking access, and building depth.
What emerges is not a single answer but a density landscape. A map of what is possible within the regulatory envelope, showing where different typology choices lead and what each one costs in spatial terms. The FAR stays constant. The spatial quality varies enormously.
The most consequential density decision on any site is the relationship between building height and ground coverage. Higher buildings on a smaller footprint produce more open space at ground level but create taller facades, longer shadows, and more complex construction. Lower buildings spread across more of the site reduce height but compress the space between structures.
This trade-off is not abstract. On a 10,000 square meter site targeting 12,000 square meters of GFA, a four-story scheme at 30% coverage produces a very different neighborhood than a six-story scheme at 20% coverage. The four-story version has shorter buildings but narrower courtyards and less continuous open space. The six-story version has more generous ground conditions but requires elevator access and produces more shadow on adjacent parcels.
Neither is inherently better. The right answer depends on context: the surrounding building heights, the orientation of the site, the municipality's priorities for outdoor space, and the developer's target market. What matters is that the team sees both options, with their full spatial consequences, before committing to one.
Tools like Hektar's typology system make this comparison possible by generating both configurations against the same constraint set and presenting them side by side with comparable metrics. The decision becomes explicit rather than inherited from whichever concept the architect happened to sketch first.
Building typology is not a stylistic preference in early-stage planning. It is a density lever. Parcel geometry and typology interact in ways that are difficult to predict intuitively but become clear when tested systematically.
A narrow, elongated parcel strongly favors lamella rows aligned along the long axis. Perimeter blocks on the same parcel produce awkward proportions with courtyards too narrow to be usable. A square parcel, by contrast, accommodates perimeter blocks naturally but may waste space with lamella rows that leave large, undefined areas between building ends.
These relationships between site geometry and viable typologies are precisely where generative tools add the most value. A team that tests three typology families against a specific parcel in the first hour of a project will make better density decisions than a team that refines a single typology choice over three weeks. The exploration is not about generating more options for the sake of it. It is about understanding which spatial strategy the site actually wants.
Density numbers travel well through organizations. They fit in board presentations, municipal reports, and investment memos. Spatial quality does not travel as easily. This creates a recurring problem in early-stage planning: decisions are made on the basis of metrics that are easy to communicate but insufficient to evaluate.
When a developer tells a municipality that a site will deliver 400 units at a FAR of 1.3, the municipality hears a number. They do not hear whether those 400 units come with adequate outdoor space, reasonable building depths, or sufficient daylight on lower floors. The spatial consequences of the density commitment are invisible until the detailed plan stage, by which point the fundamental layout is already locked.
Generative feasibility tools bridge this gap by producing spatial evidence early enough to inform the density conversation. Instead of a number, the developer brings a set of tested configurations showing how 400 units actually sit on the site under different typology strategies. The municipality can see the courtyard dimensions, the height profile, the shadow patterns. The conversation shifts from whether the density is acceptable to which version of that density produces the best neighborhood.
This is where structured feasibility becomes a communication tool, not just an analytical one. The density target stays the same. The shared understanding of what it means in spatial terms becomes dramatically richer.
In most development processes, livability is assessed after the massing is fixed. Daylight studies, sky view factor analysis, wind comfort modeling, and outdoor noise assessment all happen in the detailed design phase, when the fundamental building positions and heights are already committed. If the analysis reveals problems, the corrections are cosmetic rather than structural. Moving a building 10 meters or changing its height by two stories is not a design revision at that point. It is a project restart.
Running livability checks during the density exploration phase inverts this sequence. When sunlight, sky view, and noise are evaluated for every generated configuration, the team discovers livability constraints at the same time they discover density possibilities. A configuration that technically meets the FAR target but produces courtyards with less than two hours of direct sun in December gets flagged immediately, not six months later when the detailed plan is already in municipal review.
This front-loading of quality assessment is perhaps the most underappreciated benefit of generative density exploration. It does not make the analysis more sophisticated. It makes it timely.
Density targets are necessary. They give structure to municipal planning, create accountability for developers, and provide a shared language for discussing what a site should deliver. But density targets expressed only as numbers create a false sense of precision. A FAR of 1.2 is not a design. It is a constraint within which hundreds of designs are possible, each with different consequences for the people who will eventually live there.
The teams that produce the best density outcomes are not the ones that hit the target most efficiently. They are the ones that explore the full range of ways to hit it, understand the spatial trade-offs involved, and choose consciously rather than by default. That exploration is now possible in hours rather than weeks. The question is no longer whether it can be done, but whether teams are willing to look at what they find.
