Unit Compression in Brent: Half of Approved Small Sites Lose Units

The question

When you model a Brent acquisition, what unit count should you underwrite? The number you propose, or something materially lower? Across London, approval is usually treated as a binary outcome. In Brent, it is more accurately described as a negotiation, and the data shows how much ground developers typically give.

The finding

Across 622 decided small site applications in Brent between January 2023 and March 2026, 294 were approved. Of those approvals where proposed and approved unit counts could be compared, 82 out of 160 (51.2%) had units reduced during the determination process. That is the highest compression rate of any London borough in our dataset.

Conversions (CNV) compress at 57.4% (35 of 61 approved schemes lost units). End-of-terrace developments (END) show 80.0% compression, though on a sample of just 5 approvals, too small to generalise from with confidence.

The pattern is clear enough at portfolio level: if you get approved in Brent, there is a coin-flip chance you will build fewer units than you applied for.

What drives this? Brent’s refusal profile is unusual, though not in the way we first reported. Space standards (SPC) are the leading refusal reason at 28.8% of coded reasons, ahead of insufficient information at 17.5%, transport at 11.1% and design at 10.9%. Design runs below the London average of 14.1%, which is the genuinely unusual part.

Correction, August 2026. This article previously reported daylight and sunlight as Brent’s top refusal reason at 32.1%, and described the borough as the only one in London where daylight overtakes design. That was an artefact of the keyword classifier then in use, which measured 46.3% accurate when later tested against an adjudicated sample. Reading the reasons it had labelled daylight shows they concern loss of a family-sized unit, cycle storage, tree surveys, drainage and industrial floorspace; none mentions daylight. Under the August 2026 reclassification, which measures 91.8% accurate, daylight accounts for 0.3% of Brent’s coded reasons. The inversion we described did not exist.

The corrected profile still points somewhere useful. In Brent the binding constraint is the arithmetic of the units themselves: sizes, room dimensions, ceiling heights and private outdoor space. That is consistent with the compression finding above, and it is a constraint an applicant can test before submitting rather than discover on refusal.

Conservation areas tell an unexpected story too. In Brent, being inside a conservation area is associated with a 64.1% approval rate (n=117), compared to 42.8% outside one (n=505). That is a 21.3 percentage point advantage, the reverse of the typical London pattern. The likely explanation: conservation area sites in Brent tend to be larger, lower-density plots where daylight geometry is less constrained.

The overall refusal rate of 53.2% is above the London average but not extreme. The real risk in Brent is not outright refusal. It is approval at a unit count that breaks your numbers.

The implication

Consider a developer proposing 6 units on a mid-terrace site in Wembley. Acquisition price: £550,000. Construction cost budgeted at 6 units. GDV modelled at 6 units.

Given Brent’s 51.2% compression rate, prudent underwriting models viability at 3 to 4 approved units, not 6. A scheme that only works at 6 leaves little margin for error.

At 6 units with an average GDV of £400,000 per unit, gross development value is £2.4 million. At 4 units, it drops to £1.6 million, an £800,000 reduction. Construction costs fall too, but not proportionally. Fixed costs (professional fees, CIL, finance arrangement, project management) remain largely constant. The margin that looked comfortable at 6 units can evaporate at 4.

The practical response is threefold. First, commission a daylight and sunlight assessment before you exchange, not after. In Brent, this is not a submission formality; it is a feasibility tool. A BRE-compliant daylight study on the existing site will tell you whether 6 units is geometrically achievable before you spend £20,000 on a planning application.

Second, design the scheme so that it remains viable if one or two units are removed. This means avoiding layouts where every unit is load-bearing for the viability model. A scheme of 6 units where 4 still work is more robust than one where 5 is the minimum.

Third, recognise that Brent’s £50,000-per-unit affordable housing contribution on schemes of 5 to 9 units creates a double cliff. A 6-unit scheme triggers £300,000 in contributions. If compressed to 4, the contribution falls away entirely (threshold is 5). The compression that damages GDV may simultaneously remove an obligation.

The limit

Compression rates are calculated from applications where both proposed and approved unit counts are recorded. Not all approvals have this data; the 160-application sample represents those where the comparison is possible, not the full 294 approvals. The analysis also cannot distinguish voluntary amendments (developer-initiated redesigns) from officer-required reductions. Some compression may reflect tactical over-bidding rather than genuine negotiation losses.

The full Brent Intelligence Report covers area-by-area compression rates, officer determination patterns, and the daylight refusal profile in detail across all 622 decided applications.

Read the full Brent borough dashboard