A consultation is only honest while the drawings can still change

I think most consultation by residential developers in France is a sales activity held too late to be anything else. A scheme is designed, its architect and its financial model are settled, the permit application is nearly ready, and only then are the neighbours invited to a public meeting to see it. They are asked for views on a building whose height, footprint and number of homes are no longer open. My position is that a developer should consult only while the drawings can still change, should say at the outset which choices are open and which are not, and should publish afterwards what it refused as well as what it conceded.

This is not a claim that neighbours are always right, or that every objection deserves a concession. Some opposition to new housing is opposition to new neighbours, and a developer that yields to it fails the households who need the homes. Nor is it an argument about legal obligation. French planning law requires prior consultation for certain public development operations, but a private developer can usually file for a building permit on an ordinary urban site without asking anyone. The neighbours then have two months from the display of the permit on site to challenge it before the administrative courts, and a challenge, even one that fails, can hold a programme up for a year or more. My argument concerns what consultation is for, and what it costs to do it honestly.

What residents are asked when nothing is left to decide

For most of my career I watched consultation follow a familiar sequence. The development team prepared boards with perspective drawings and a site plan, hired a hall near the site and presented the scheme. Residents raised height, overlooking, parking, trees and traffic. The team listened courteously, explained why each concern had already been considered and promised another meeting. Little changed, because little could. The architect's fees had been spent on that design, the land price had been agreed on that number of homes and the permit timetable was tied to a land payment.

Residents understood this perfectly well. In my experience they objected less to the scheme than to being asked for an opinion that could have no effect. Some then used the one instrument that did have an effect, which was an appeal against the permit. I do not think developers are entitled to complain about appeals while running consultations designed to change nothing.

The alternative is to consult earlier, at a point that is genuinely uncomfortable for the developer. In 2024 Valcerne agreed to buy a former printing works in Villeurbanne, on a street of three- and four-storey houses and small apartment blocks, with a row of mature plane trees along one boundary. Our first feasibility study showed 312 homes. Nine months before we planned to file for a permit, we opened a project room in the works' old gatehouse, staffed on two evenings a week and on Saturday mornings.

We published three choices that were open: the height of the buildings facing the street, the fate of the plane trees and the use of the ground floors. We also published what was closed. The city required a share of social housing on a site of that size, and we would not reduce it. We would not build fewer than 250 homes, because below that number the land price we had agreed no longer worked. And we would not provide more parking than the local plan required, because we believed it would bring more traffic into the street than it took off it.

Publish the list of what you refused

Over six months about 430 people came to the project room, and we received 212 written comments. The changes that followed were real. We lowered the street-facing buildings by two storeys and moved the height to the back of the site, against a railway embankment. We kept all eleven plane trees, which meant moving an underground car park and losing one building. We opened a public passage through the site to a school on the next street, which residents asked for repeatedly and our architect had not considered. The scheme fell from 312 homes to 268.

We also refused things, and I believe the refusals mattered as much as the concessions. One group of residents asked us to drop the social housing; we declined, explaining that the city required it and that we supported it. Several asked for more parking, and we declined for the reasons given at the start. Others wanted the scheme cut to 200 homes, below the floor we had published. We set out every refusal, with its reason, in a document placed in the project room, on our website and at the town hall before we filed.

That document was the most useful thing we produced. It told residents that their comments had been read, including those we did not act on, and it gave them a fixed account to hold us to. When the permit was displayed in 2025, nobody appealed. I cannot prove that the consultation caused that. Some sites attract no appeal anyway, and this neighbourhood badly wanted the passage to the school. But several residents told us afterwards that the refusals document was why they believed the concessions.

The cost of listening, in homes and months

Consultation of this kind is not free, and developers who say it pays for itself are usually counting only the appeals it avoids. In Villeurbanne we gave up 44 homes, about €2.9 million of margin at the prices we expect, and nine months of programme time before filing. We paid for staff in the project room, for two further design rounds and for redesigning the car park around the trees. Our chief financial officer accepted those costs on the reasoning that an appeal would probably have cost more. I think she was right about this site. I would not claim the same for every site.

The approach has limits. It works where the developer controls the land and has time before a land payment falls due. It is harder where a landowner insists on a permit by a fixed date, and harder still where a public development company has already fixed the volumes. It asks the development team to live with uncertainty about the number of homes for longer than it would like. And it requires a marketing function willing to write a document listing the requests it turned down, which at first feels like advertising against oneself.

Since 2024 Valcerne has applied this as a written charter on sites above a certain size: consultation opens before the massing is fixed, the open and closed choices are published at the start, and a list of changes and refusals is published before filing. Seven sites have gone through it. None has yet faced an appeal, though two are still within the period in which one could be lodged.

I would put one test to any developer that says it consults. Before the first meeting, write down the choices that could still change because of what residents say. If the list is empty, cancel the meeting and call what remains a presentation. If the list is long, publish it, and be ready to publish what you refused.