Keep or Demolish? A Simple Test for Any Old Building

We keep what can work, not what merely looks old.

The Question Every Project Should Start With

Stand on a site with something already on it, a shed, a wall, an entire derelict block, and there's a decision to make before any design work begins: what stays, and what goes? Get it wrong one way and you bulldoze the very thing that gave the place its worth. Get it wrong the other way and you drag a dead weight through the whole project, spending money to keep something that will never earn its keep.

Most briefs treat this as a yes or no: keep it, or demolish it. In our experience that binary is exactly what leads people astray, because the most valuable buildings on a site are often the ones that don't fit cleanly into either box. So we work with three categories instead of two.

Three Categories, Not Two

We first developed this test on an extreme site, an abandoned island of Poveglia in the Venetian Lagoon, where every structure had to be judged one at a time, and we now run it on any project that comes with existing fabric.

The first category is preserve and transform. The building is sound, or can be made sound, and it carries the identity of the place. It can't be erased or disguised, but its interior is fair game: remove a floor, insert a new volume, change the use entirely. The shell endures while the life inside it gets renewed.

The second is incorporate. The building is too far gone to stand on its own, but too meaningful to throw away. You stabilise whatever is left and build the new architecture into and around it, deliberately keeping the seam between old and new visible so anyone can still read what was there. This is the subtle category, the one most often skipped, and often where the best architecture actually happens.

The third is remove. What's left carries neither a usable structure nor any real meaning. You let it go, without guilt, and without pretending otherwise.

The whole point of three boxes instead of two is that middle one. Incorporate is where a place keeps its memory and still gets a real future, instead of being either stuffed full of the past or flattened clean of it.

How to Tell Which Is Which

Sorting a structure comes down to four honest questions.

• Is it sound? Can it safely carry use for decades more, as it is or with reasonable work? If yes, lean toward preserve. If it's failing but fragments are worth saving, lean toward incorporate.
• Does it carry identity? Would losing it make the place generic, erase something a local person would actually recognise? A high enough identity can pull a building up and out of the remove box even when it's in rough shape. This is roughly where Kevin Lynch's idea of the legible landmark earns its keep.
• Can it adapt? Generous ceiling heights, decent spans and good daylight mean a structure can absorb uses nobody has thought of yet. A building that can only ever be one thing is a weaker candidate than an old shell that could become almost anything.
• Does it help its surroundings? A structure that makes the street or the site better, encloses a square, holds a corner, shelters a walk, earns its place. One that only blocks and divides does not.

A quick way to hold all this as principle: soundness and adaptability decide whether you can keep something. Identity and context decide whether you should.

The Trap: Keeping the Wrong Things

There's a failure mode on the keep side too, and it's worth naming because good intentions are usually what cause it. You can save a building so thoroughly, and so literally, that you kill it: turn a living structure into an embalmed relic that blocks the life around it instead of hosting any. Salvatore Settis's warning about Venice applies just as well to a single building: a place reduced to walls with no working life inside becomes, in his words, a carcass, a funereal scene (Settis, 2016).

Our own rule guards against both the bulldozer and the embalmer:

The existing earns its place in the new by carrying a living function, not by being spared one. We keep what can work, not what merely looks old.

Jane Jacobs gives the reason to keep more than instinct alone would suggest: a lively place needs old buildings, because their long-paid-off costs are what shelter the small, low-overhead uses that keep a district alive (Jacobs, 1961). Done well, reuse aims for minimum intervention and maximum benefit. Not the most preserved building, but the most alive one.

A Worked Example: Poveglia, Sorted

On Poveglia the three-way test produced a clear map. Preserve and transform covered the old residences, the fortification octagon, the boat house and the church: sound enough, and utterly identity-defining, now re-inhabited with new campus uses inside. Incorporate covered the worst-preserved hospital and church fragments, stabilised and folded into new buildings with their edges left legible, so the history still reads. Remove covered only the parts carrying neither structure nor meaning.

That single sorting exercise, done before a single new line was drawn, set the logic for the entire project. The island reads as one continuous story precisely because nothing was kept, or lost, by accident.

Why This Saves Money as Well as Carbon

This isn't only good conservation, it's good economics too. Every structure you can genuinely keep is embodied carbon you don't have to re-spend and construction you don't have to re-pay for, and, increasingly, a box ticked for regulators moving toward whole-life-carbon accounting. But a structure kept badly, the wrong one, saved out of sentiment, is a cost with no return. The three-way test is how you make sure the things you keep are the things that actually pay you back: in carbon, in character, and in use.

Before we design anything on a site with history on it, we run this test first. It's the least glamorous hour of the whole project. It's also the one that decides whether everything that follows makes any sense.

---

References

• Ninković, K. (2023). The Revival: Poveglia’s Identity as Self-Sustaining University Campus.
• Jacobs, J. (1961). The Death and Life of Great American Cities.
• Settis, S. (2016). If Venice Dies.
• Lynch, K. (1960). The Image of the City.
• Embodied-carbon context for the “saves money as well as carbon” claim: Bryden Wood; EU whole-life carbon — One Click LCA (EPBD).
• Concepts: adaptive reuse, palimpsest, the existing as resource, urban diversity, jane jacobs.