The Most Circular Thing a Builder Can Do Is Keep the House

Seattle issued about 490 residential demolition permits in the 12 months to September 2026. Each one marks a house that someone decided not to keep.

Most talk about the circular economy, the idea that materials should stay in use instead of going to a landfill, starts with what happens next. Salvage crews pull out doors and mantels. Resellers put them on shelves. Online marketplaces try to match old parts to new projects.

Studio Viridian and Viridian Builds develop, design, build and hold homes in Seattle, working alongside Abacus Fine Carpentry, a finish-carpentry company with about 25 carpenters. Over several months we studied those salvage and resale businesses as possible lines of work. Then we widened the question. Of everything a firm like ours could do, what would return the most good for the money, time and risk?

We rated 23 interventions on two scales: what a serious, ten-year investment would change in the world, and what it would take in cost, hassle and risk.

The result is lopsided. Seven interventions rate as strong, and most of them cost little. Six are set aside, including four of the five businesses we first set out to study. Between them sit seven supporting moves, worth doing mainly because they are cheap, and three long shots.

The biggest saving never leaves the lot

A new house emits roughly 40 to 55 metric tons of carbon dioxide equivalent (CO₂e), the standard measure of planet-warming emissions, before anyone moves in. Studies of new houses in Vancouver, across Canada and in Massachusetts put the figure at about 43, 40 and 55.5 tons.

Keeping a 1,500-square-foot house and adding homes to its lot avoids roughly 28 to 45 of those tons. Taking the same house apart by hand and reselling its parts saves about 5 to 8.

Keeping also avoids every problem that makes reclaimed parts hard to use. Removal, storage, refitting and code approval all come from moving a part, and a kept house’s parts never move. A study of more than 65,000 permits in six North American cities found new single-dwelling buildings about 10 times as carbon-intensive as additions to existing ones.

The idea has a Pritzker Prize behind it. At Grand Parc in Bordeaux, the architects Lacaton & Vassal, with Frédéric Druot and Christophe Hutin, transformed 530 occupied apartments for about €50,000 to €65,000 each. Demolishing and rebuilding would have cost about €165,000 each, by the architects’ figures. Their motto is “never demolish, always transform.”

Seattle’s new zoning makes the approach practical here. Rules adopted at the end of 2025 let most lots in the city’s Neighborhood Residential zones hold up to four homes, and some up to six. The city’s preapproved backyard-cottage plans cut permit review from four to eight months to two to six weeks, by its own count. The new rules offer no bonus for keeping the existing house, though the city’s old Residential Small Lot zone had one.

Why salvage stalls

Seattle does not lack salvage crews. Earthwise has removed and resold building parts across the region since 1991, and Second Use since 1994. The trouble is the arithmetic.

Taking a typical house apart by hand costs about $19,000 more than knocking it down. Each ton of material saved takes about 80 crew-hours, and what buyers pay, plus the dump fees avoided, covers about 28 of them.

Reclaimed parts are also slow to use, however beautiful. Each one is its own small project of fitting and approval. Old windows usually fail the energy code. Leaded glass near doors fails safety-glazing rules. Structural lumber needs a grade stamp, and no American grading rule covers reclaimed wood. Faced with all that, clients often choose something cheaper.

What a ton of carbon is worth

Studio Viridian’s draft Impact Investment Policy, part of Rent, Own and the Return on Impact, sets a yardstick. Keeping a ton of carbon out of the air for $100 or less clearly pays. Seattle’s gas-to-heat-pump conversions and measured weatherization cost about $250 to $330 a ton, and serve as the benchmark. Around $600 needs a stated reason.

By that measure, deconstruction with retail resale, at about $2,900 a ton, fails on carbon alone. Best practice brings it to about $650. Reusing salvage on the same lot costs about nothing, and so does keeping the house, as long as keeping costs no more than replacing. Whether it does in Seattle is the first thing to test on real lots.

Who lasts

The history of reuse businesses points the same way. Marketplaces built to match supply with demand have struggled. PlanetReuse launched in 2008 as an “Amazon for reclaimed building materials” and had become a broker by 2013. Chicago’s Rebuilding Exchange closed in October 2025. Reclaimed goods do sell in volume, but on Craigslist, Facebook Marketplace and eBay.

The survivors curate. LASSCo and Retrouvius have lasted decades in London on design and craft. 1stDibs, the polished online marketplace for such pieces, had $89.6 million in revenue in 2025 and lost $13.7 million. In Seattle, the incumbent resellers already fill the curated niche.

The higher the rung, the higher the return

Circular-economy researchers often draw a ladder of strategies, from refusing and rethinking at the top, through reuse and repair, down to recycling and recovering energy. The higher the rung, the more of a product’s value is kept. Where an intervention sits on that ladder predicts most of its rating.

All 23, one by one

The small charts below repeat the grid, one intervention at a time, grouped by the role Viridian would play.

One firm, or one rule

Viridian’s own projects can only go so far. At two or three keep-and-add projects a year, its share comes to roughly 60 to 135 tons of carbon a year. A city rule moves hundreds of projects.

If a floor-area bonus for keeping the house moved even 5 percent of the several hundred houses Seattle demolishes each year, 15 to 35 houses a year would avoid roughly 400 to 1,600 tons. A deconstruction requirement covering about 150 houses, at the 7.6 tons a house that Portland’s program reports, would avoid about 1,100 tons a year.

The timing is good. Seattle Public Utilities is designing a 2027 step for deconstruction, and the city is still adjusting its new zoning. A builder-owner with real cost records is credible in a way an advocate is not, and making the case would take 20 to 40 hours a year.

What AI changes

Better AI makes studies, analyses, software and design libraries cheap, which is a reason to give them away for reach. It does not make land, kept buildings, a crew that can restore a 1912 interior or a name that cities and lenders trust any cheaper. The strong interventions rest on that second list. Only two of the 23 lose value outright as AI improves: the public marketplace and the paperwork service.

A standing test, not a new business

The survey points to a practice rather than a new division: a short set of rules applied to every project.

How far to go is a separate choice, and each level builds on the one before.

None of the strong interventions depends on salvage skill or software. They depend on what the firm already does. It decides what to keep, designs the next change, restores old interiors and pays the maintenance bills for decades. Few firms combine all four.