Every developer and consultant can access the same data — applications, refusals, decisions, inspector reports, committee minutes. The problem is not access. It's the work of reading across all of it, finding the pattern, and reaching a judgment that's actually useful.
Development Intent does that work. This page explains how — and where we'll tell you not to trust a figure.
Methodology v2 · Last updated May 2026The same planning data answers different questions at every stage. We organise it by the decision you're making — four stages on top — not by the data itself: seven intelligence layers running continuously beneath. Click any stage or layer to trace the linkages.
Assess a site before instructing anyone. Know whether a scheme has a meaningful chance before committing time or fees.
The intelligence is decision-ready before you engage a consultant, solicitor, or design team. Less abortive spend.
Every judgment is grounded in what has actually happened on comparable schemes — not what you expect the council to say.
One number, 0–100, for any parcel of land — five pillars, an exclusion gate, and every input traceable. Read the published method.
Enter a postcode and see the whole route ahead — delegated, panel or committee, the people who decide, every obstacle, and how long until you can actually build.
Every number comes from public records or verified landowner submissions. Below is exactly what we collect, how we store and score it, and where we flag a figure as too thin to trust. Read it if you want to check our working — skip it if you just want to use the platform.
An intent score is a weighted combination of up to five independent signals from the public record: planning history, listing status, policy environment, political landscape and market timing. One signal is noise. Three is a pattern. Five is a deal.
Refused applications, withdrawn applications, pre-app enquiries with no follow-up application, and sites with repeated attempts by different applicants. These indicate a motivated owner or a difficult site — context that changes what someone will pay.
We monitor Rightmove development-land listings, major auction houses (Savills, Allsop, Auction House London), and price-reduction events on listings previously appearing in our planning data. A listing with a planning history is materially different from a fresh listing.
Borough policy environment (how strict the adopted local plan is, and how often inspectors overturn it on appeal) and political landscape (committee behaviour, ward-level refusal patterns). Two identical sites in two boroughs can have very different approval odds.
Comparable sales momentum in the area, time-on-market for development land, and borough-level decision-volume trends. These contextualise whether the moment is right independently of the site itself.
Landowners come to us directly for an architect's view of what their land could become. We index those conversations, cross-reference them with planning, policy and site constraints, and surface qualified opportunities to subscribers.
When a subscriber's criteria match a site, they see a teaser first (location, scale, planning context). If interested they can request a full introduction; the landowner chooses whether to proceed. We do not broker — we introduce.
We typically take a 1.5% to 2.5% introduction fee at completion of an off-market deal we facilitated. The exact rate is agreed before the introduction and varies with the size and nature of the opportunity. No fee is payable on on-market listings. Commission is only earned on successful introductions, aligning our incentive with the outcome.
Application-level data is drawn from the statutory planning registers of all 33 London local planning authorities. Each register is a public record maintained under the Town and Country Planning (Development Management Procedure) (England) Order 2015.
Appeal outcome data is sourced from the Planning Inspectorate's published decision records. These cover appeals decided under section 78 of the Town and Country Planning Act 1990 — the standard route for refused applications.
Policy data covers the full adopted local plan for each analysed borough, cross-referenced against appeal decision letters to identify which policies inspectors uphold and which they routinely overturn. Westminster (901 policies) was indexed first; other boroughs follow as data is validated.
Listing and auction signals are drawn from public property listings and auction catalogue entries. Land Registry data is used for price paid, ownership changes and title detail where permitted under the Registry's licensing terms.
Ownership-type classification (private individual, family trust, property company, institutional, public sector, housing association, overseas/offshore) is derived from Land Registry proprietor data cross-referenced with Companies House records.
A development is classified as a major residential scheme when it proposes 10 or more dwellings, or where the site area is 0.5 hectares or more and the number of dwellings is not specified. This follows the definition in the Town and Country Planning (Development Management Procedure) (England) Order 2015, Schedule 4.
Householder applications, changes of use, prior approval notifications, listed building consents, non-material amendments and minor residential schemes (under 10 dwellings) are excluded from refusal rate calculations. Mixing application types would make borough comparisons meaningless.
Refusal rate = (refused decisions ÷ total decided applications) × 100, where both sides are filtered to major residential schemes within the relevant date range. Withdrawn applications are excluded from both.
The default view covers 2017 to 2025. This captures a full cycle of market conditions and planning policy changes. Trend views are available year-by-year from 2017.
We use decision date rather than application date. This reflects the actual output of the planning system in a given period and avoids distortion from applications submitted in one year but decided in another.
Every time-to-decision figure on this site is the MEDIAN: line up all the dated cases from fastest to slowest and take the middle one. It is labelled "median" wherever it appears — consultant profiles, borough pages, the planning-queue investigation. One basis, everywhere.
The mean lets a handful of forgotten files distort a borough. Real example from our own records: Harrow has 1,351 dated tree applications; the middle case took 10.0 weeks, but the mean reads 12.3 — pulled up by conservation-area notices lodged in October 2020 that were only formally closed, as "no objection", in November 2024, plus withdrawn cases shut years late. Those are administrative artefacts, not anyone’s real waiting time. The median tells you what will typically happen to you; the mean tells you about the council’s filing habits.
The clock runs from the date the council records receiving the application to the date it issues the decision, converted to weeks. Applications missing either date are excluded from time figures (they still count in volumes and approval rates); boroughs with no dated cases for a given view show as no-data rather than a guessed number.
Workload totals (e.g. "application-weeks" burden) sum the real recorded weeks across all cases — there the long-open files genuinely are part of the burden, so they are counted, not trimmed.
An appeal is counted as overturned when the Planning Inspectorate allows an appeal against a refusal — meaning the inspector disagreed with the council. The overturn rate for a borough is: allowed appeals ÷ total decided appeals (excluding withdrawn appeals).
Borough-level overturn rates — previously a standalone Intelligence page — now live inside the Intelligence and the borough performance track record table. The methodology is unchanged: raw allowed-appeal rates against decided-appeal counts, with the London average shown alongside so a borough can be read relative to the citywide norm.
Inspector-level data tracks the outcome of every appeal assigned to each inspector. Rates are not adjusted for appeal type or scheme size. We present raw rates alongside decision counts so users can assess statistical reliability — 3 decisions is not the same as 80.
Our appeal dataset covers 10,000+ Planning Inspectorate decisions for London from 2017 onwards, refreshed every weekday morning from the Inspectorate’s casework record. We focus on section 78 appeals (refused applications and non-determination). Enforcement, lawful development certificate and other appeal types are excluded.
Single datasets describe one thing at a time. The insight is in the intersections — which refusal reasons survive appeal, which inspector overturns which committee, which officer-consultant pairing actually gets approved, and how much affordable housing is really delivered after viability negotiation. The Intelligence combines two datasets and surfaces the pattern.
Each appeal decision letter is parsed to identify which local plan policies the council cited as refusal reasons, and whether the inspector upheld or dismissed those reasons. A policy overturn rate is: appeals where the policy was overturned ÷ total appeals where the policy was cited. A high overturn rate means the policy has been a weak ground for refusal when challenged — not that it is unlawful.
S106 and viability analysis compares policy-compliant affordable housing targets against what is actually agreed after Financial Viability Assessment negotiation. Sources are committee reports and decision notices; figures are sometimes partially redacted and therefore modelled rather than exact. Modelled figures always carry a confidence indicator and are never presented as precise.
Every score we present carries a confidence indicator: HIGH (full supporting data), MEDIUM (partial data), LOW (thin data, treat as indicative) or UNVERIFIED (surfacing only, not recommended for decisions). Nobody else admits when their data is incomplete; we do.
Sample size (how many decisions support the rate), recency (how recent the supporting data is), cross-validation (whether the signal is corroborated by independent data) and source integrity (whether the underlying record is verifiable). Geographic heat surfaces apply the same rule — cells below a minimum sample are shown faded, never coloured as if certain.
Refusal and overturn rates describe what has happened. Every planning decision depends on the specific proposal, site, officer and political context at the time — none of which is captured in an aggregate rate. Scores are input to judgment, not a substitute for it.
London planning registers are maintained independently by each LPA and vary in completeness and classification accuracy. We validate by spot-checking against published decision notices; minor errors in borough totals are possible.
Planning Inspectorate decision records typically appear 4–8 weeks after the decision is issued. Appeal data is therefore not fully real-time for the most recent period.
Planning and appeal data are updated monthly. Listing and auction data refresh weekly. The date of the last update is shown on all data views.
Scores are input to judgment, not a substitute for it. Every planning decision depends on the specific proposal, site, officer and political context at the time. Development Intent surfaces pattern data; the decision to proceed is always yours.