The Adaptive City

A building can be inexpensive to buy and expensive to make useful.

That distinction should sit at the center of the conversation about adaptive reuse. A discount to a previous sale price tells us what happened to an asset. It does not tell us whether the next use will work.

For a residential conversion, the more consequential question is what remains after the building has been made safe, serviceable and genuinely desirable as a place to live. How much of its original advantage survives the work required to transform it?

My interest is in the next life of San Francisco’s existing buildings. But that interest begins with a discipline: underwrite from the home backward, not from the discounted building forward.

The purpose is not to prove that every building can be converted. It is to recognize the ones whose next use can justify the cost, risk and intervention—and contribute something worthwhile to the city.

Possibility is not feasibility

SPUR and ULI San Francisco made an important distinction in their 2023 research with Gensler and HR&A Advisors: the physical suitability of an office building for housing is separate from the financial feasibility of converting it. Their findings reflected the buildings, costs, rents and policy scenarios examined then. They are not a permanent verdict on San Francisco or a substitute for evaluating a particular property today.

The distinction is more durable than any individual forecast.

A credible conversion must pass several tests: the building must accommodate the use; the approval path must support it; the economics must justify delivery; and the completed homes must make sense for the people expected to occupy them.

These tests inform one another, but none can stand in for the rest. A zoning opportunity cannot make a dark room desirable. An elegant plan cannot establish the condition of a foundation. An attractive purchase price cannot compensate for every uncertainty still inside the walls.

Read the building before writing the story

An office floor and a residential floor do not make the same demands of a building. HKS identifies floor-plate depth, daylight, plumbing, ventilation and residential life-safety requirements among the practical considerations in conversion work.

The first task is to understand how those conditions interact, rather than assess each in isolation. Where can principal rooms meet the exterior? Can bathrooms and kitchens be arranged around workable service routes? What clear ceiling height remains beneath the structure and new systems? Where will sound travel between homes?

The most revealing plan is not the one with the largest unit count. It is the one that shows a believable day in someone’s life: entering with groceries, finding a place for a desk, opening a bedroom door without colliding with furniture, sitting beside a window that offers privacy as well as light.

That is not decoration added after feasibility. It is a way to test whether the homes can support the rents or prices assumed for them.

Early screening still has limits. Gensler’s San Francisco compatibility assessment expressly called for further property-level investigation; it did not assess hazardous-material abatement, and historic-district considerations were not incorporated into the ranking. A favorable score was an invitation to investigate, not permission to stop.

Nor must the answer be an all-or-nothing conversion. At One Dallas Center, HKS describes residential use on the upper 16 levels of a 30-story building, with commercial tenants below. The lesson is not that every building should be mixed-use. It is that the proposed program should respond to the building, rather than forcing the entire building to serve a predetermined narrative.

The purchase price is only the beginning

Buying below replacement cost is not the same as delivering below the value of the finished property. Replacement cost describes an alternative; it does not establish a conversion’s profit.

The acquisition decision should work backward from a defensible completed value, allowing for construction, professional services, approvals, financing, carrying costs, contingencies and the return required for taking the development risk. The residual is what the project can support for the property—not necessarily what the seller needs to receive.

The area assumptions deserve particular scrutiny. An acquisition price per gross square foot and a sales comparable per residential square foot are not directly comparable. Corridors, stairs, shafts and shared facilities still have to be paid for, whether or not they generate revenue directly.

Consider a simplified, invented example—not a San Francisco cost benchmark. A 100,000-gross-square-foot project has an all-in development budget of $60 million. With 80,000 square feet of net residential area inside the homes, the cost is $750 per net square foot. At 70,000 net square feet, it rises to approximately $857: a 14.3 percent increase without any increase in the total budget.

Holding cost constant isolates the effect of the area assumption. In an actual redesign, both cost and revenue could change.

The conclusion is not to maximize saleable area at any expense. A courtyard might remove floor area while making the remaining homes materially better. Its value must be tested against the lost area, construction cost and achievable pricing—not assumed from a rendering. Equally, preserving an extra room that the market does not value is not genuine efficiency.

A rental strategy and a condominium strategy also need their own complete economics. Net operating income is not gross rent. A condominium sellout is not cash received on the day construction ends. Timing, transaction costs and the recurring costs borne by future occupants belong in the analysis.

“We can always rent it” is only a fallback when a separately tested rental case supports it.

Investigate uncertainty before financing it

Adaptive reuse calls for more than a contingency percentage. It calls for a plan to reduce the uncertainties that percentage is supposed to cover.

I would start with the questions most capable of changing the acquisition decision: the structure’s condition and capacity, the scope of building-system replacement, hazardous materials, and the space needed to deliver a compliant plan. Investigation should be sequenced around consequence, not around which report is easiest to order.

Consider an unforeseen condition requiring structural reinforcement. It could increase cost, take space from a planned residence and extend the schedule. Those are not three unrelated risks. They can be three consequences of one discovery.

The model should test that combination rather than isolate each change in a separate sensitivity table.

Time belongs in this analysis from the beginning. HR&A’s study explicitly included the time needed to empty and convert a building, and identified the approval path and existing occupancy as influences on feasibility. A partially vacant property is not necessarily available for construction. Existing tenancy rights and the timing of possession must be understood, not assumed away.

The construction schedule, loan term and cash reserves should be reconciled. An allowance for additional work is not automatically enough to cover additional months of interest, taxes, insurance and operating shortfalls.

At each major commitment, the useful question is: what uncertainty have we resolved, what remains, and what would make us change course?

Deciding not to proceed after investigation is not a failure of imagination. It is one of the reasons to investigate.

Policy creates options, not certainty

San Francisco Planning’s commercial-to-residential bulletin describes distinct forms of relief, including planning-code modifications, tax and fee provisions, and financing incentives, each with applicable conditions. These are not interchangeable benefits, and eligibility for one does not establish eligibility for all.

For a project team, the task is to translate policy into a documented sequence: which provision applies, what evidence is required, which milestones matter, and when the benefit becomes available. A prospective incentive arriving after completion cannot be treated as cash available for construction without a credible way to bridge the timing.

Additional development capacity deserves the same discipline. Permission to add area is a reason to investigate its value, not a reason to assume the revenue. What will the addition require of the existing structure, circulation and services? Will it improve the project after the full cost and schedule consequences are counted?

The public sector has an equally important distinction to make. Removing an unnecessary obstacle is not the same as weakening a necessary safeguard. Clear requirements and predictable decisions should be goals alongside viable incentives.

And financial feasibility is not synonymous with housing affordability. Where public support is proposed, its intended public benefit should be explicit, proportionate and capable of being measured.

Retain the advantage, not just the structure

A conversion should have an account of what retention contributes and what adaptation demands in return.

That account starts with the physical building but should not end there. A useful structure, a distinctive facade, an established entrance or a particular relationship to the street can be worth preserving. Some elements may require substantial repair; others may frustrate the new use. The case for each should be tested rather than settled by either nostalgia or a preference for everything new.

Carbon is part of that judgment. RMI identifies reuse, retrofit and material efficiency as strategies for reducing embodied emissions—the emissions associated with materials and construction. Retention can avoid future material demand; it does not undo emissions already released when the original building was made.

The responsible question is comparative: what would the credible alternatives require? Account for demolition, reinforcement, new materials and future operating energy over a stated period, with comparable housing output. Tools such as CARE support early comparisons of embodied and operational carbon for renovation and replacement scenarios; a project-specific claim requires project-specific inputs.

“Reused” should describe what happened to the building. It should not substitute for an assessment of environmental performance.

The same care applies to the public story. A neighborhood is not an empty backdrop awaiting an investment thesis. It contains people, businesses and patterns of use that should be understood before they are rearranged.

An entrance that welcomes residents without turning its back on the street; planting that can be maintained; a ground floor with a credible purpose; shared space that invites use without sacrificing privacy: these are modest decisions individually. Together, they determine whether a building participates in the city or merely takes its address.

For me, this is where development discipline and a sense of place belong together. Neither excuses the absence of the other.

A second life, not a convenient story

Not every building should become housing. Some should remain in their existing use, some warrant a different intervention, and some will not support a viable next chapter under present conditions.

Recognizing those limits makes the case for the right conversions stronger.

The promise of adaptive reuse is not that existing buildings are automatically cheaper, faster or better. It is that careful judgment can preserve advantages that replacement would discard—and introduce a new use without making its occupants live with every compromise inherited from the old one.

The work begins with evidence: the structure, the plan, the cost, the approval path. It should end with something less abstract.

A room with good light. A threshold that feels like home. A building that belongs to its street and has become useful to another generation.

A building earns its second life when the people who live there no longer need the conversion story to explain why it is a good home.