What replaced the London Plan density matrix

London Plan Policy D3 sets site capacity through design rather than through a table of densities. The 2016 London Plan did the opposite: Policy 3.4 and its density matrix gave a range of habitable rooms and units per hectare for each combination of setting and public transport accessibility level, and an applicant could open the plan, find the cell, and argue inside the range. That matrix is gone. Since the London Plan 2021 was published, capacity is whatever a design-led process can justify on the particular site, and the argument has moved from arithmetic to evidence about form.

For applicants this is harder, not easier. A number in a table was a defence. A design-led case has to be built, and boroughs are entitled to ask for the design options you tested and rejected.

What the matrix did and why it went

The old matrix was criticised from both directions. Developers used the top of a range as an entitlement, and boroughs used the bottom of it as a cap, and neither reading had much to do with whether the resulting building worked. It also produced odd outcomes at the edges, because two sites with identical accessibility scores could have entirely different capacity depending on their shape, their neighbours and their topography, none of which the matrix noticed.

Policy D3 replaced it with a requirement that all development makes the best use of land by following a design-led approach that optimises the capacity of sites. The word is optimises, not maximises. Optimising means the most appropriate form and land use for the site, which can mean fewer homes than a crude calculation would suggest, and the policy is explicit that it can also mean more.

The three stages the policy expects

The design-led approach runs in sequence, and skipping a stage is the most common weakness in a capacity argument:

1. Understand the site and its context, covering physical constraints, the character and grain of the surrounding area, transport accessibility, existing and planned infrastructure capacity, and any designations.
2. Develop and assess design options that respond to that understanding, examining form, layout, height, massing and land use rather than starting from a unit count.
3. Test the preferred option against the policies of the development plan, including housing quality, daylight, play space, greening and fire safety, and refine it where those tests bite.

The output a borough wants to see is the second stage written down. Two or three options, drawn, with a short statement of why each was carried forward or dropped. Applicants routinely present only the scheme they intend to build, which leaves the officer with no way to test whether capacity was optimised or simply asserted. The published treatment of the policy itself is set out in optimising site capacity under London Plan Policy D3.

Evidence a borough expects on capacity

Accessibility levels change when bus routes change, so quote the level and the date you checked it. On a site in an Opportunity Area the planned infrastructure matters as much as the existing network, and an allocation carries its own expectations about capacity, which is covered in what a London allocation changes.

Where optimisation meets local character policy

This is the live tension in most London applications. The Plan requires optimisation. Borough local plans and neighbourhood plans contain character, height and townscape policies that pull the other way, and both sit in the development plan under section 38(6) of the Planning and Compulsory Purchase Act 2004.

The resolution is not that the London Plan wins. It is that the design-led approach is the mechanism for reconciling them, and a scheme that has properly tested options against local character carries the argument. Where a borough has an adopted character appraisal or a tall buildings policy map, work inside it and say so. Where the proposal exceeds a local height expectation, the case has to be made on design quality and on the transport and infrastructure position, with the options testing showing what a compliant scheme would deliver and why it was rejected.

One honest caveat. The London Plan 2021 remains the adopted spatial development strategy for Greater London, and a review of it has been running, so check the current status of any replacement before relying on the policy numbering in a submission.

Testing capacity against housing quality

A capacity figure that fails the quality policies is not capacity. The tests that most often reduce a unit count are the residential quality requirements: internal space against the Nationally Described Space Standard, floor-to-ceiling heights of at least 2.5 metres across most of the residential floor area, limits on the number of dwellings served from a single core on each floor, and the push away from single aspect homes that face north or sit against a noisy frontage.

Daylight inside the proposal matters as much as daylight to neighbours. Deep plan flats on a tight site fail internal daylight long before they fail the neighbour tests, and the fix is usually fewer, better units. The approach to running those numbers in a dense context is set out in applying BRE 209 daylight guidance in dense London.

Then there is the amenity arithmetic. Child yield drives play space provision, and play space has to be on site before off-site contributions are considered, so a family-sized mix on a constrained plot eats into the footprint available for building. A Play Space Assessment run early will tell you whether the mix and the massing can coexist.

Presenting the capacity argument

Write the capacity case as a short narrative with the drawings doing the work, in this order: what the site is, what the context sets as a reference, what options were tested, what each option delivered, which was chosen and what the quality tests did to it.

Give the officer a single table that reconciles the final unit count with every quality policy. It saves a round of questions and it stops the debate drifting back to units per hectare, which the Plan no longer uses. Where the scheme is tall enough to engage the tall buildings policy or the fire safety policy, flag the supporting documents in the same table rather than leaving them to be discovered.

Avoid two habits. Do not present a density figure as compliance, because there is no range to comply with. And do not describe the design-led approach in the planning statement while the drawings show a single option that was clearly the starting point rather than the conclusion. Officers read both.

Illustrative example: a 0.3 hectare backland site with a moderate accessibility level, surrounded by two-storey terraces and a four-storey mansion block. Three options tested at 3, 4 and 5 storeys produce 18, 24 and 31 homes. The five-storey option fails neighbour daylight on two flank windows and leaves no room for the play space the child yield requires. The four-storey option passes both and provides a dual aspect mix. The capacity argument is then 24 homes, evidenced, rather than 31 asserted. The figures are illustrative only.

Enter your postcode on PlanScope to see which reports a London site is likely to need.

This is general guidance, not legal advice. Check the position with your local planning authority or a qualified professional before relying on it. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.