One address.
Every number that decides it.

A single 50-to-90 grade for how well a California lot can support a detached ADU. It combines five things: whether state law and zoning permit a unit, how much buildable area survives setbacks and existing structures, what the local rent supports economically, fire and flood exposure, and how deliverable the site is. Every component and its data source is shown alongside the number.

84 and above is well suited: room to build with clean constraints. 75 to 83 is suitable and effectively ADU-ready. 65 to 74 is workable, meaning real options exist with some compromise on unit size or placement. Below 65 is limited: buildable, but the constraints are doing the deciding. No lot scores above 88, because the model does not hand out its own maximum.

Because both ends would be misleading. Any parcel that reaches a score has already passed the checks that would disqualify it outright: a lot that genuinely cannot host an ADU is routed to a specialist rather than scored, so the scored population starts viable and the floor reflects that. And no lot is perfect: 90 is the top of the printed scale, and the cap sits at 88, precisely so the model never awards itself full marks.

Buildable capacity carries the most weight at 30%, because usable area is the binding constraint on most lots. Legal position and economics are 22% each, hazard exposure 16%, and delivery logistics 10%. They are combined as a weighted geometric mean rather than a simple average, which matters: a near-zero on any single factor drags the whole score down instead of being averaged away by strong results elsewhere.

Because ADU feasibility is non-compensatory in real life. A lot with abundant space but no legal path is not a good lot, and averaging would let the space score paper over the legal problem. A geometric mean punishes a weak component in a way an average cannot, which produces a grade that behaves like an experienced eye rather than a spreadsheet.

The affected factor is excluded and the remaining weights renormalise. Missing data is never scored as if it were bad data, and nothing is guessed to fill the gap. When enough inputs are unresolved, the page shows your score as a range rather than a single number, and says which check a site visit would close.

No. A low score means the lot is harder, not closed. Common causes are a small buildable envelope after setbacks, a fire-hazard designation that triggers stricter construction, or existing structures crowding the yard. Many low-scoring lots still support a smaller unit or a different placement; the score is a screening tool for prioritising, not a verdict.

Neither. It is a feasibility screen computed from public parcel geometry, hazard layers, and rent data. It does not inspect your soil, review your title, or bind any city. Treat it as a fast, well-sourced way to know whether a property deserves a closer look before you spend money finding out.