For architects
The same evidence base, read at brief stage.
Perfect Scale is the planning dataset that fell out of an unusual exercise: an ARB-registered architect who had authored a London council’s Small Sites SPD reading a lot of decision notices in his own time. Patterns turned up that aren’t in any policy document. Coded, those patterns let an architect see at brief stage what the borough has actually approved on comparable schemes: density, height, plot ratio, fenestration, materials. The same methodology returns the same answer whoever asks it.
The gap between top-decile and bottom-quartile firms is wider than most architects expect
We coded every named agent firm in our dataset by approval outcome. The distribution is wide, and it is descriptive: what the record shows, not a prediction of any single scheme. Figures below are qualified for sample size; we do not publish raw small-sample extremes or name individual firms.
Most of the difference between the top and bottom of that distribution is structural. The top firms file schemes that match the borough’s approved patterns: density, height, plot ratio, fenestration, materials. The bottom firms repeat the same refusal-triggering choices. The gap is around thirty percentage points of approval rate. Read the same evidence at brief stage and a practice can see where its own work sits against that record before it files.
Three ways architect practices use Perfect Scale
Calibrate the brief against the record
The developer arrives with an eight-unit brief. The borough’s approved pattern for that site type may sit lower. Reading the evidence at brief stage puts that conversation onto a shared, independent record rather than a matter of opinion: the same number the borough’s own decision history shows.
A shared reference across the studio
The borough’s full approved-and-refused pattern set in one place (density matrices, refusal taxonomy, decision routes, case studies) gives a studio a common evidence base to brief and review schemes against, rather than each project starting from scratch.
An independent third reference
When a design choice is contested, in-house or with a client, the published decision record is a neutral reference both sides can read the same way. The methodology is open, so the basis of any figure can be checked rather than taken on trust.
Built by an architect, for architects
Perfect Scale is run by Abre Etteh: ARB-registered architect, fourteen years across architecture, council planning and residential development. Author of the London Borough of Merton Small Sites Supplementary Planning Document, the policy framework councils now reference when assessing the kind of schemes Perfect Scale codes.
The methodology is open. The data comes from the GLA Planning Datahub and each council’s own register. Every refusal reason is coded into a ten-category taxonomy with verbatim references. Every finding is graded into one of four evidence tiers based on sample size and statistical significance. Where the data is thin, we say so.
It is not a black-box AI tool. It is the dataset every chartered planning consultant builds in their head over thirty years of practice, made available at the brief stage of a single scheme.
Start with the published evidence
The free borough dashboards carry each borough’s approval rate, refusal taxonomy and quarterly trend. All 33 London boroughs, refreshed quarterly, no sign-up. They are the same evidence base, read at the level of the whole borough.
Around three in five of London’s coded refusal reasons sit in categories addressable at the drawing board: design, amenity, space standards, daylight and transport. For a single named site at brief stage, or to use the evidence base as a studio reference, see how we collaborate or get in touch.