Could this building become homes?

How this check works

It is a desk check from public records — no site visit, no drawings. It reads what the radar knows about a building (or measures an OpenStreetMap outline in your browser), then runs three layers. Rules version 2026-10-10.1.

1 · Fit

Five parts, each scored 0–10 and weighted: floor plate 30%, daylight 30%, height and storeys 20%, services and structure 10% and location 10%. 70 and up is band A, 50 and up band B. The method follows the published office-conversion screens — Gensler's five categories (whose own scorecards weigh windows and daylight twice their brochure weight) and the TU Delft Transformation Meter's veto and gradual criteria — adapted to what public records hold in Europe.

• Floor plate. Depth façade to façade of 10–16 m is ideal, still workable to 20 m and worth nothing at 30 m (Basel study: up to 16 m good, from 20 m special concepts; Brussels Task Force Bureaux: 17 m takes dual-aspect flats, 20 m and more does not; Banque des Territoires' 2026 guide: under 15 m; BBSR Redevelopment: post-war offices are 12–17 m). A plate of about 750 m² is best.
• Daylight. The share of each floor within 7.5 m of a façade, measured on the outline. On this page the outline is measured again on the base map: walls shared with a neighbour are not façades, and the daylit band shrinks to 2.5 × the clear height where that is less (BDA Bayern, Office to Housing).
• Height. Floor-to-floor height (official height ÷ storeys) on a scale, with 0.4 m of slab and services (BDA): below the country's minimum clear room height it is a warning, below the 2.6 m clear the market expects it is capped (the Transformation Meter), 3.3–4 m earns a bonus and taller floors none. Homes above 4 storeys need a lift, which buildings not built as offices often lack.
• Services and structure. What the building was built for, and its age where a record has it: asbestos is likely from 1930 to 1993, slabs of the 1950s–70s often fail today's fire and sound rules, from 1985 they are sound (BDA; BBSR).
• Location. Steps at 300 / 500 / 800 / 1,200 / 1,700 m to rail or metro (BDA); road noise costs points from 55 dB (the Transformation Meter wants under 50); an industrial estate counts against it.
• Flags. A listed building goes to the challenging cost tier (heritage conversions cost the most); a co-ownership has every owner to win over (Brussels); the year built, listing and land use come from the record or OpenStreetMap.

2 · Programme

It starts from the economic model's site default — shared rooms, ground-floor businesses, furnished mid-term flats — and lets the floor plate move it: a dark core becomes the building co-op's shared rooms (up to 25% of the floor), the ground floor of a building of 3 storeys and up stays commercial, what is still dark is let as storage and workspace or cut out as a light well, and the band 7.5–10 m from a façade suits small furnished flats (up to 50% of the homes) — for awkward plates the Basel study recommends small flats and serviced shared flats run by an operator. The 3D model stacks the layers from the ground up — businesses, the shared rooms both groups use, furnished flats, long-term homes — on the building's own outline; only a dark core breaks the stack. A rough check of floor areas, not a design.

3 · Money

EHC's economic model — its own engine — runs the building on today's site default for Espoo: its floor area and mix, a conversion cost and a purchase price. The conversion cost starts from the country's factor and the building's fit band (the model's good / typical / challenging tiers); a replaced façade adds 20 points of new-build cost (BDA: conversions cost 60–120% of new build, the façade being the biggest item; BBSR: 22–42% of works) and a light well 5. “Works” means the bank lends and the deal balances at the site default's own money, and the homes' fee for the same flat starts at or below the market. The most the building's co-op could pay is the highest price at which it still works. Public pages show it in bands so that a named building's ceiling is not printed for everyone; signed-in members see the figure. A ground lease is modelled as a yearly ground rent on the value, with nothing paid up front. The carbon line compares the materials of a new building (about 505 kg CO₂e per m²) with the conversion's ( ULI 2025, 24 US conversions; materials only), as a range.

Timber storeys on top

Where a building stands lower than its street — measured on satellite heights of it and its neighbours — the check offers to build the missing storeys on top in serial timber, at most 3 and never past a 25 m roof or 8 storeys. It is off until you switch it on: it needs the city's consent and a structural survey, which no desk check knows. The new floor (90% of the footprint a storey, the rest set back) joins the building's: the shared rooms keep their share of all of it, and the homes go up top. It adds no land. It costs 110% of new build a m² (95–130% in the sources: the TU Darmstadt Deutschlandstudie, a Dutch top-up business case and ARGE Kiel: about what new build costs, a little more for small projects), blended with the conversion's cost by floor area; storeys that take the homes above 4 storeys need a lift, a new one about €150,000. The timber's own carbon, 200–400 kg CO₂e a m² of materials, is counted apart from what the conversion saves, the carbon the wood stores not subtracted (timber against concrete, SBUF 2018, on the new-build figure above). The price stays per m² of the building as it stands.

Tested on buildings that were converted

Does the check turn down buildings that real projects did convert? We ran the fit layer on 294 buildings in Paris that received a permit to turn them into homes (Ville de Paris — autorisations d’urbanisme with a change of use to housing (person export, 6 Oct 2026); outlines IGN BD TOPO): 65% scored A, 33% B and 2% C; their median depth is 12 m. Location, services and use are not on the permits, so those parts sat at the middle. It cannot show the reverse — that a low band is right — because nobody publishes the conversions that failed. A C is not a no; it is where cost and compromise pile up.

What it cannot tell you

The façade, the structure, pollutants, heritage listing, whether the owner would sell, and what a planning office will accept. It is a first sort, to decide which buildings deserve a feasibility study — not the study.