Why development sites fail.

Most fail in principle. And the tells are there early.

Reviewed August 2026

The short answer

Development sites rarely fail on design. They fail in principle: the plan does not put growth there, a designation covers the land, nothing can reach it safely, or the numbers will not carry what the site has to pay. Every one of those is readable before you buy.

In practice · Newbold Road

Expensive is not the same as refused.

The distinction is clearest in a site that looked like a failure and was not.

The order of work

01Read the planWhere the development plan directs growth.

02Test the principleWhether the use is acceptable here at all.

03Price the restRemediation and access as numbers.

04Then designA modest scheme, sized to establish the use.

A corner plot of 0.57 hectares at 117 Newbold Road, on a gateway into Rugby town centre. It had been a gas works, then a haulage depot, then a vehicle rental yard. Contaminated ground, a railway to the north, a three storey office hard against the west boundary, and no residential consent of any kind.

Every one is a reason to walk away. Not one is an objection in principle. Brownfield land inside a town centre is where the plan puts housing. Contamination is a remediation cost. The office and railway are design problems. What it lacked was a reason the use could be refused.

The job

Site: A former gas works on a gateway into Rugby town centre
Role: Architect and design author on both consents; the planning submissions were run by others
Consents: Rugby Borough Council refs R17/2113 and R19/1496
Outcome: 98 homes consented, then 122 on the same red line

The first job was to prove the principle, not maximise the plot. The scheme that did it was a block of 98 apartments around a courtyard, and Rugby Borough Council granted it in December 2018. The approved drawings carry our own project reference, so the consented building is ours; the planning submissions were run by others throughout.

Then the arithmetic moved. Decontamination cost more than anyone had priced, and 98 homes stopped paying for the site. Twenty apartments went onto a new third floor repeating the plate below, four turned the eastern corner, and parking rose from 61 to 73 spaces. Same red line, same building line. Committee granted 122 in August 2020.

The order matters more than it looks. Establishing the use first, on a scheme the council could comfortably grant, made the second application a design question rather than a policy one: the principle, the access and the massing had already been accepted, so the only argument left was whether the extra homes disturbed them. The access took the strain of that, gaining dedicated right turn and left turn lanes.

Doing it the other way round is the expensive version. A site with no residential consent, tested first with the largest scheme the arithmetic will bear, invites a council to refuse the principle and the quantum in the same notice. A refusal on the numbers is a redesign. A refusal on the principle is a different site.

The proof

Four sites, four diagnoses.

Constraint · missed

Foleshill Road, Coventry

The site already carried consent for a larger studio block, and a high voltage cable easement nobody had found made that footprint undeliverable. Nothing on the policies map showed it. The answer was a tighter building designed around the cables, and Coventry granted that one instead.

Coventry City Council ref FUL/2022/2402 · granted January 2024

Recommended · refused

Oakdale Nurseries, Brandon

A closed rose nursery in the Green Belt east of Coventry, taken to outline for a care village of 124 retirement homes and a 36 bed care centre. The planning officers recommended approval, subject to a section 106 agreement. The committee then refused it as inappropriate development.

Rugby Borough Council ref R18/0167 · refused 11 November 2019

Policy · refused

Althorpe Street, Leamington

Student accommodation on a town centre employment site. The council accepted that employment-only redevelopment was not viable, then refused on policy: the loss of employment floorspace, and an over-concentration of student housing in that quarter. Winning the argument is not winning the permission.

Warwick District Council ref W/19/0505 · refused March 2020

Capacity · tested first

Frog Lane, Balsall Common

A national housebuilder was weighing a 5.2 hectare site where somebody else's outline application was still undecided. Rather than take the promoter's density on trust, its own house types were plotted across the developable ground: 95 homes, 38 of them affordable, before any offer.

Solihull · a pre-acquisition capacity appraisal, 2024

Every planning application we make in our own name is a matter of public record on a council register, and ours reach more than 100 planning authorities across the UK.

Architects for developers →

The first step

Not sure a site stacks up?

Send us your project and where it is. A Senior Architect will give you an initial appraisal and guide you as to what to do next.

Start a conversation

01 · Where you stand

In principle, or on the detail

A refusal notice does not tell you which of the two you are holding, and the difference decides everything that follows.

A refusal on detail says the council accepts the use and objects to this version of it: too many units for the parking, an overbearing rear elevation, windows looking into next door. Change the drawing and the objection goes. Almost every householder refusal is this kind, and this page is not about them.

A refusal in principle says the council does not accept building there at all, in that form, by anyone. No drawing answers it. On a development site it is the refusal that matters, because it is the one that decides whether the land is worth what was paid for it.

The tell is in the reasons. Count how many cite a policy about where development goes, rather than about how a building looks or behaves. If that is most of them, the argument is about the principle and a redesign is the wrong instrument.

There is a middle case worth naming, because it is the one people misread. A scheme can be refused on detail so comprehensively that it amounts to a refusal in principle: if the only version that answers every detailed objection is a version that does not pay for the land, the site has failed on the numbers, and the notice will never say so.

It runs the other way too. A reason that reads like a matter of principle sometimes turns out to be a matter of evidence. A highway objection with no modelling behind it, or a landscape objection with no assessment, is answered by a specialist's report rather than by a redesign, and answering it is far cheaper than starting again.

02 · The failure modes

Where a site fails in principle

Five modes account for most of it. Each one leaves a tell on the public record, and every one of those tells is free to look up.

Mode 01

The plan says elsewhere

Applications are decided in accordance with the development plan unless material considerations indicate otherwise. That is statute, not guidance. Land the plan has not identified for development starts the argument a long way behind, and the more recent the plan, the further behind it starts.

The tell: no allocation of any kind on the adopted policies map, and a plan adopted in the last few years.

Mode 02

Outside the boundary

Most local plans draw a line around each town and village and treat everything beyond it as open countryside, where new housing is limited to a short list of exceptions. The line itself is local, and the framework never once uses the term. It moves at plan review, and a neighbourhood plan can move it.

The tell: the boundary on the adopted policies map runs along the edge of the site rather than around it.

Mode 03

A designation covers it

Green Belt is the one everybody knows, and it is not a blanket ban, but it is the hardest of the common designations to argue past. A National Landscape, a conservation area, flood zone 3 and a protected habitat each carry a test of their own, and one site can be carrying several of them at once.

The tell: most planning designations are mapped and published, so most surprises are a search nobody ran.

Mode 04

Nothing can reach it

Access is the one a redesign cannot always solve, and the bar for refusing on it is deliberately high: only an unacceptable impact on highway safety, or residual cumulative impacts on the road network that would be severe following mitigation. Sites still fail on it, because the land runs out.

The tell: the visibility splay the highway authority wants runs across ground in somebody else's title.

Mode 05

The numbers stop closing

A site pays before it pays anyone: affordable housing, biodiversity net gain, contributions under a section 106 agreement, remediation, and the cost of building the access. Where the plan is up to date, a scheme that complies with it is expected to be viable, so arguing that it cannot be is uphill.

The tell: the appraisal only balances once one of the plan's own requirements is quietly assumed away.

03 · The odds

The approval rate answers nothing

The commonest question about a site is the least useful one. In the year to March 2026, district planning authorities in England granted 87 per cent of the decisions they made. Just over half of those decisions, 51 per cent of them, were householder applications, so the figure describes extensions and loft conversions more than it describes land.

Residential development is a different slice of the same release, major and minor together, and it ran at 76 per cent. Closer. Still an average across every authority, every plan position and every kind of site in the country, which is to say it is not about your site at all.

The number that matters is narrower again, and nobody publishes it. How many schemes of this kind, on land like this, in this authority, under this plan, have been granted in the last three years. That one you can assemble yourself in an afternoon from the council's own register, and the answer is usually not ambiguous.

Where nothing comparable has been granted, the reading is not that yours will be the first. It is that the principle has not been established here, and that your application is the one that will test it, at your cost and on your programme.

Three years is the window because a local plan, a housing land supply position and a case officer team all turn over on roughly that cycle. Decisions older than that describe a policy landscape that may no longer exist, and they are the ones most often quoted back by a seller.

Search by site, not by keyword. Council registers index the address, and the applications that tell you most about your land are usually the ones on the land beside it: what was asked for, what was granted, what was refused, and which conditions the council thought it needed to impose.

Common questions

Refusals, answered

Why is planning permission refused?

Most refusals turn on a conflict with the development plan. On a development site that conflict is usually about the principle of building there at all.

The reasons on a notice fall into a small number of families: the site is not where the plan directs growth, a designation covers it, the access is unacceptable, the scheme harms a heritage asset or a protected habitat or a neighbour's living conditions, or the evidence on drainage, contamination or noise is missing. On smaller schemes the amenity reasons dominate and a redesign usually answers them. On development land the policy reasons dominate and a redesign usually does not. A notice carrying one reason can be more serious than a notice carrying five, because five detailed objections describe a version of the scheme the council could accept, while one policy objection describes a use it will not.

Should I appeal a refusal or resubmit?

Resubmit where the reasons are about the scheme, and appeal where they are about the principle. A redesign cannot answer a policy objection. Decide early: the appeal deadline runs from the decision.

An Inspector will not redesign a building for you, which is why the split matters. Two practical points sit underneath it, and both cut against resubmitting on reflex. The free repeat application no longer exists: the fee exemption for a further application of the same character on the same site was removed in December 2023, so a second attempt now costs a second full fee. And a refusal is not a clean slate: where an authority has already refused more than one substantially similar application on the same land inside two years, with no appeal against them, it can decline to determine the next one if nothing material has changed. A resubmission that changes nothing is not a free move. The mechanics of choosing between the two are set out in the planning appeal guide.

How do I spot a plot that will never get planning permission?

Look for a small parcel of open field, sold without permission, at a price that only makes sense if permission arrives. Then check the adopted policies map.

The pattern is consistent. A field beyond a settlement boundary is divided into plots and marketed on the prospect of future permission, often with a report attached that describes the potential rather than the policy. The land is real and the sale is a real sale. What is being priced is the chance of the plan changing. Three checks get most of the way in an afternoon and the public versions cost nothing: whether the site sits inside the settlement boundary on the adopted policies map, whether the local plan allocates it for anything, and what the council has already refused on that land or the land beside it. Older planning history is the one that varies, because not every authority's records reach far back online. Where the first two are no and the third turns up refusals, the price is not a discount.

Does a refusal make a site worthless?

No, but it changes what the land can be assumed to be. A refusal on detail costs a round of time and fees. A refusal in principle removes the assumption the purchase price was built on.

A refusal is public information about the land and the next buyer will find it, so the useful question is which of the two kinds it is. A refusal on detail is answered with a better application. A refusal in principle stands until something in the policy position moves, and the things that move it, a local plan review, a shortfall in housing land supply, a change to national policy, run on timescales measured in years rather than months. Reading the notice carefully enough to tell the two apart is the first job on any refused site, and it costs nothing.

Does a refusal count against a site later?

Planning history is a material consideration, so yes, but not in the way people fear. What carries weight is the reason for the refusal, not the fact of it.

The one hard consequence is narrower than people expect. An authority can decline to determine a further application only where the development and the land are substantially the same, nothing material has changed, and either it has already refused more than one such application inside two years without an appeal, or the Secretary of State has refused or dismissed one in that window. Beyond that, an earlier refusal is weighed for what it says about the policy position, and one taken under a plan that has since been replaced carries little. An earlier permission is read the same way and is often overlooked: even a lapsed one is evidence the council found the principle acceptable once. Its weight is for the decision maker and it fades as policy moves on, and unlike a live permission it gives you no fallback.

Can a pre-application enquiry tell me whether a site will fail?

It can tell you what the policy asks and what evidence will be tested. It cannot tell you the answer, because the advice is officer opinion and it does not bind whoever signs the decision.

That is still worth paying for on a development site, because most of what stops a scheme is knowable in advance and a pre-application response puts it in writing. What it will not settle is a decision that is going to a planning committee, or one that turns on the weight given to a designation, because weight is a judgement made at the point of determination. The stronger use of a pre-application is the negative one. It is a cheap way to find out that the answer is no, before the surveys, the application fee and the six months.

Does a shortfall in housing land supply help a site?

It can be decisive. Where an authority cannot show a five year supply of deliverable housing sites, or has delivered far below its requirement, the balance shifts towards granting permission.

What that means in practice is that permission should be granted unless any adverse impacts would significantly and demonstrably outweigh the benefits. The second trigger is the Housing Delivery Test, which bites where delivery has run below three quarters of the requirement over three years. For an unallocated housing site there is no bigger lever, and none that moves as often: the position is recalculated and re-argued at every appeal, so a site can be arguable one year and not the next. It does not sweep designations aside. Where applying a protective policy gives a strong reason for refusing, and Green Belt is the commonest, the shifted balance falls away and the ordinary one returns. What matters there is the outcome of applying the policy, not the presence of the designation on a map.

Written by

Rickett Architects

A RIBA Chartered practice, working since 1970, with more than 2,500 projects behind it. Our senior architects take projects from first survey through planning to completion, for private clients, developers and landowners across the Midlands and well beyond it.