This is the eighth in our Adaptive Urbanism series. Previous essays have focused on foundational pillars – what we refer to as “imperatives” – for adaptive urbanism. The cognitive imperative focused on data to sense urban change as it is happening, the institutional imperative focused on governance and stewardship, and the economic imperative focused on finance and capitalization for the built environment, in a world of increasing volatility. This is the design imperative, and it is the most visible expression of Adaptive Urbanism.
The Houston Post, a hulking concrete structure at the center of the city, was designed for a single purpose: to process Houston’s mail. Construction was completed in 1962, and the building served its purpose well for half a century before it was decommissioned in 2014. By the mid-2010s, the infrastructure was aging, and the development option – sixteen acres of prime downtown-adjacent land – was more valuable than the building. The Houston Post was scheduled for demolition.
But a developer, Lovett Commercial, saw adaptive potential. The building’s heavy concrete frame is engineered for industrial loads. A regular column grid is sized for sorting operations. Floor-to-floor heights are much greater than contemporary office norms. By virtue of its initial use, The Houston Post had a generous structure that could, decades later, become a remarkable canvas for reinvention.
The developer hired OMA to realize the building’s adaptive potential. The team, led by Jason Long, carved vast atriums out of the sorting floor, built a rooftop lawn, added three monumental staircases – idiosyncratically shaped as an X, an O, and a Z – and brought light flooding into spaces that had been gloomy for fifty years. The building reopened in 2021 as a cultural and commercial hub with food halls, music venues, co-working, retail, and the largest rooftop park in Texas. POST Houston, as it is now called, is worth more (financially and socially) than it ever could have been as anonymous mixed-use development on the site of a demolished postal facility.
POST has an unmistakable life to it. Families gather in a buzzy dining area, office workers cross paths with concert-goers heading to the 713 Music Hall. Above, five acres of rooftop park spread across the building, paths winding between gardens, an organic farm, and an unobstructed view of the downtown skyline.
The building maintains its historic identity, but POST adapts the structure to fit the economic, social, and technological patterns of today. The building learned, playing host to a transformation its original designers could never have imagined – and adaptation was only possible because of its overbuilt structure. Some buildings have this capacity by accident, but every building should have it by design. In a world that changes rapidly, the least risky building is the one that holds options on multiple possible futures.
POST is a remarkable example of wholesale adaptation at a single point in time. Buildings more commonly evolve in fits and starts – clumsily, intermittently, often against their architects’ intentions – in a decades-long sequence of tenant improvements, structural retrofits, and code compliance upgrades.
In How Buildings Learn, Stewart Brand described that process of change: how it happens naturally, and how the physical elements of a building fight against change. To Brand, a building is a set of nested layers, and each wants to cycle at its own pace. He identified six, ordered from slowest to fastest:
Site — the geographical setting and the plot itself. Stable for longer than human timescales.
Structure — foundation and load-bearing elements. Thirty to three hundred years (but usually no more than 60)
Skin — the exterior envelope. Cycles every fifteen to twenty-five years as the building weathers and building codes evolve.
Services — the mechanical, electrical, and plumbing systems. Replaced every seven to fifteen years.
Space Plan — the interior layout. Reconfigured every three to thirty years, and now, in commercial buildings, almost continuously.
Stuff — the contents, like furniture and fixtures. Changes daily.
Brand observed that buildings fail because of layer entanglement. When a building feature that needs to change fast is constrained by one that is slow, the building cannot evolve to meet the demands of the moment. If services are cast into structural slabs, if the skin is inseparable from the frame, or if the space plan is locked to load-bearing interior walls, the building is doomed to obsolescence, locked into the technologies and assumptions of its original design.
Brand’s framework points to layer independence as a design principle. Each layer must be able to change at its own pace without forcing changes to the layers beneath it.
The Houston Post was designed for a specific program, and engineered for permanence, but it was amenable to radical change precisely because nothing fast was locked to anything slow. OMA’s redesign treated the layers of the building separately. The structure stayed, for example, but the services were entirely replaced, threaded through the existing frame using new risers. The Cold War era building will surely continue to evolve over the decades to come – and when it does, the space plan will shift more quickly than the services, the skin more quickly than the structure.
Layer independence is the foundation for an analytical protocol we call conversion friction. The number reflects the cost and complexity of changing a building’s use from a specific “origin” to a specific “destination.” The viability of pivoting any given building from office to residential, for example, or from retail to laboratory, is quantified as a “conversion friction coefficient,” between 0 and 1. A lower conversion friction coefficient means that it will be easier to exercise a particular conversion option; a higher conversion friction coefficient means that specific conversion will be challenging. Conversion friction has four dimensions, each with several contributing factors.
Structural friction encompasses floor-to-floor height, building depth, column grid, and load capacity. Structural friction carries the strongest weight in calculating the conversion coefficient because it tends to be very difficult or expensive to alter after initial construction. Take the example of an office building with 11ft slab-to-slab height. Laboratory spaces require 12 feet or more to accommodate fume hoods and specialized servicing. It may technically be possible to carve through a floor to offer greater height, but it is so challenging that the office to lab conversion option carries a very high conversion coefficient. The option is structurally foreclosed. In that same building, the floor to floor height may work well for residential conversion, but the depth of the floor plates result in a lack of interior daylight – another high-friction variable. The building is designed for a standard structural load capacity of 2.5-3.0 kN/m² – counterintuitively, less than the live load minimum for a parking garage, at 1.92 kN/m². Office to parking conversion is low friction, along with office to advanced manufacturing.
Services friction follows close behind structural. Plumbing constrains office to residential conversion, because drainage requires gravity flow, and routing waste pipes through existing slabs often means cutting through structure. A building with services embedded in the structure has extremely high conversion friction for office to residential, regardless of how generous its other dimensions are. Electrical capacity, vertical riser space, and HVAC compatibility are all part of services.
Envelope friction refers to the building’s skin. Sealed-facade office buildings, common since the 1970s, cannot be converted to housing regardless of what their structure or services would otherwise allow. Operable windows, daylight reach, and thermal performance under tightening building codes all sit in this layer.
Regulatory friction includes use class restrictions, building code differences across uses, historic designation, and accessibility standards that can foreclose options the building’s physical fabric permit. The building can be physically convertible and legally trapped.
Certain conversion origin and destination pairs will tend to have higher conversion coefficients – things like office to housing or warehouse to early education – while others, like office to self storage, will tend to be lower. Right now, those are mostly intuited. The concept of conversion friction gives a quantifiable cost basis.
Stewart Brand observed buildings naturally changing, and identified the time scales typical to each layer. Conversion friction preserves the sequential ordering of Brand’s shearing layers, but shifts the focus from the time scale of natural conversion to the cost and complexity of immediately exercising a conversion option.
The concept of conversion friction shifts the focus from natural conversion to quantifying the cost and complexity of immediately exercising a conversion option.
Considered alongside market signals and capital availability, conversion friction determines the viability of an adaptive reuse project. A specific conversion may be high friction (expensive, complex) but still enjoy returns because the demand for the new use is so great. Most office to residential conversion is high-friction, but the cost of conversion will never be recouped through rental yields in a normal market. Many cities are experimenting with grants to reduce the cost. The concept of conversion friction helps us compare different conversion paths as apples to apples: office to housing, or self storage, or educational, or light industrial.
In that way, it helps surface an important policy question: will a grant program have a lasting, structural impact on the market, or will it temporarily and artificially subsidize a type of conversion that will never be viable in the real estate market without an injection of free capital?
The previous essay in this series presented a thought experiment: Building A, optimized for a single use; Building B, designed for adaptation, costing five to ten percent more at construction. Under stable conditions, A wins. Under conditions of uncertainty, B wins.
The specifications of a building that has many options – a building with low conversion friction to many endpoints – are unglamorous. It needs a floor-to-floor height of at least 12 feet (15 at the ground floor) to preserve residential, office, and laboratory options through the building’s life. It needs a 25-30 foot structural grid that allows subdivision. Load capacity above 5 kN/m² preserves laboratory and light-industrial options at a modest construction premium. Services organized in accessible central risers, with primary and secondary distribution separated, allow tenant-level changes without building-level disruption and allow major upgrades without demolishing finished spaces. It needs floor plates no deeper than 50 feet to glass, and direct street access. Ideally, the structure is separate from a central core.
For generations, architects have known how to design buildings for adaptation. But they have lacked a financial argument for why anyone should pay the premium. If the pacing problem is correct, and the future brings accelerating uncertainty, the financial value of reduced risk far outweighs the increase in upfront cost. The difference is the cheapest insurance an investor can buy.
Stewart Brand’s project was descriptive, showing what adaptation looks like in existing buildings. He did not, except in passing, ask how an architect would intentionally design a building that adapts well. Three decades on, the pacing problem brings that question to the foreground.
Field States worked with LEVER Architecture to answer it. We created the schematic design of collaborative hub building on an innovation campus (the specific client is anonymous while the project is in development). Because the future needs of the emerging industry are highly uncertain, the hub building was designed to maintain maximum optionality and extensibility. The architecture is based on a simple repeating structural unit that can be extended for any number of bays.
The design for the initial launch phase is based on our research and human-centered design process, and includes office, light manufacturing, labs, and an event and demonstration space. But the building is intended to flex and adapt to the observed patterns of use during years zero to five. If tenants spend more time prototyping than on Zoom calls, office areas can convert to light manufacturing bays. If the building is over subscribed, it can extend. If the industry grows more slowly than expected, the structure isn’t over-built, and it can become something else entirely. A light conversion will cost orders of magnitude less than a building that is stranded because it is optimized for a use that never materialized.
The success of the building depends on its capacity to adapt, and the direction of its adaptation depends on its capacity to sense the evolving demands of its users. What do people need, and how does that data inform the evolving design?
This is where the cognitive imperative informs design at the building scale. Quantitative and qualitative data – merging, for example, indoor positioning, tool utilization rates, event calendars, user interviews, and leasing patterns – is gathered in real time and analyzed longitudinally. Together these show what occupants need from the building.
The architects of the Houston Post were not designing for adaptation; they were designing for mail. The building was specified for a program that no longer exists, but its idiosyncrasies, like the floor to floor heights and column grid, are the reason the building still does. It happened to have adaptive capacity and learned in spite of itself.
POST belongs to a wider family of accidentally adaptive buildings: late 19th century warehouses, pre-war commercial blocks, post-war industrial sheds, and certain mid-century institutional structures. All of them have structural slack (born of caution, regulation, or the engineering conventions of their era) that can be taken up by multiple possible uses. These buildings have low conversion friction for many potential uses.
Field States’ project with LEVER embraces that ethos – but it is designed to learn from the start. An upfront investment minimizes the conversion friction for almost every future use, reducing long-term cost of ownership for a high-performing building. The owners – program directors, really – can exercise optionality throughout the building’s life.
The buildings that struggle most under conditions of uncertainty are the ones that have thoroughly internalized optimization, with sealed envelopes that hamper future ventilation needs, floor plates deeper than residential daylight thresholds, services embedded in structure to save inches of clearance, floor-to-floor heights specified for a single use case. These buildings are perfectly tailored and perfectly obsolete, the durable failure of a design culture that treated single ideal solutions as the highest objective.
Under accelerating uncertainty, a building tailored for a single predicted use is the riskiest and most expensive thing an architect can deliver. The cheapest insurance a building can carry is low conversion friction, and an embedded capacity to learn.