Why Does Croydon Refuse 78% of Small Sites on Design Grounds?

The question

If you are acquiring a small site in Croydon, what is actually killing your application? The borough’s 67.2% refusal rate (the second highest in London) suggests something structural. But the cause is not what most developers assume.

The finding

Across 878 decided small site applications between January 2023 and March 2026, Croydon refused 589. Of those refusals where reasons could be extracted and classified (n=470), 367 cited design quality as the primary ground. That is 78.1%.

The London-wide average for design as a primary refusal reason sits at 49.9%. Croydon is 28 percentage points above it.

To put that in proportion: nearly four in five Croydon refusals are fundamentally about how the scheme looks, how it relates to its context, and whether it demonstrates adequate design quality. Not density. Not policy compliance. Not parking or highways. Design.

This shows up consistently across site types. Mid-terrace conversions and extensions (DMR), the bread and butter of small site development, achieve only a 31.1% approval rate from 264 decided applications. Full conversions (CNV) fare marginally better at 36.8% across 372 applications. End-of-terrace schemes (END) drop to 20.4%, though on a smaller sample of 49.

What makes Croydon unusual is what does not matter. Conservation area status, typically one of the strongest predictors of refusal across London, has almost no effect here. Applications within conservation areas were approved at 34.1% (n=82). Applications outside conservation areas: 32.7% (n=796). The difference is within the margin of noise.

In most boroughs, a conservation area location adds 5 to 15 percentage points of refusal risk. In Croydon, it adds nothing measurable. The design bar applies uniformly, regardless of heritage context.

Determination times tell a consistent story. The median decision takes 56 days overall. Approved schemes take slightly longer at 62 days, consistent with negotiation and amendment cycles. Refused applications are dispatched in a median of 56 days. Officers appear to reach negative conclusions quickly.

One further signal: 27.8% of approved schemes lost units between application and consent. Over a quarter of the developers who do get through Croydon’s gates arrive with fewer units than they proposed.

The implication

For a developer acquiring a mid-terrace site in Croydon, the numbers frame the decision sharply. A 31.1% approval rate on DMR schemes means roughly two in three applications fail. If 78% of those failures are design-driven, the single highest-return investment is not in policy consultants, not in daylight assessments, not in heritage statements. It is in the architect.

Consider a worked example. You are looking at a mid-terrace property in Addiscombe, guide price £400,000. Your scheme proposes 4 units at an estimated GDV of £1.6 million. Standard pre-application and design costs might run to £15,000–£20,000. A more thorough design process (a recognised architect, a design review panel submission, detailed contextual analysis, high-quality CGIs) might cost £35,000–£45,000.

The additional £20,000 spent on design quality is not a luxury. Against a 31.1% approval rate, it is the most efficient way to shift your odds. No other single variable in the Croydon dataset correlates as strongly with refusal as design quality.

This also means pre-application engagement carries different weight in Croydon. Pre-app is most valuable when the council’s concerns are negotiable: density, unit mix, affordable housing contributions. When the dominant issue is design quality, pre-app becomes a design review mechanism. The question to ask is not “will the council accept 4 units?” but “does this design meet the standard they are evidently enforcing?” (Worth noting: the raw correlation between pre-app and approval rate in Croydon is sharply negative, for reasons of how the officer reports are written rather than what the advice itself does.)

The limit

This analysis classifies refusal reasons into a standard taxonomy. “Design” as a category covers character, streetscene, massing, materials, and detailed architectural quality; it does not distinguish between them. A scheme refused for poor materials choice and one refused for excessive massing both register as DES. The dataset also cannot capture applications withdrawn before determination, which may represent schemes where design concerns were raised informally at pre-app stage.

The Croydon dashboard covers area-by-area profiles, officer patterns, determination timelines, and viability modelling across all 878 applications.