The site shell every building should start with

Most building records are built upside down. They begin with the structure, because the structure is what gets bought, leased and insured, and the land it sits on is treated as context rather than content. The fire risk assessment, the audits, the compliance documents all accumulate around the building, while the ground it stands on, the boundary, the drainage, the trees and the access roads sit outside the record, managed loosely or not at all. It works until the day a risk in the land needs a home and there is nowhere to put it.

The alternative is to start the other way round. Define the site first, as a bounded thing, and let the building be the first occupant of it rather than the whole of it. A site shell is simply that frame: an outline with a place for everything the site contains, the building included. Start there and the land never becomes an afterthought, because it had a home before the building was even recorded.

What a site shell is

A site shell is not a heavyweight thing. At its simplest it is a defined boundary with the capacity to hold the layers a real site carries: one or more buildings, the structures around them, the ground risk, the drainage, the trees, the access, the legal envelope. The point is not complexity. The point is that the container exists before the contents, so nothing the site contains is left without somewhere to sit.

The difference this makes is structural. In a building-first record, land risks are exceptions you remember to add. In a site-first record, they are slots you fill. The retaining wall, the protected tree, the buried services, the sustainable drainage feature each have an obvious place to live, because the site was defined as the thing that holds them. We have made the specific cases elsewhere, in managing the land a building sits on and draw the boundary once and let everything hang off it; the site shell is the practical mechanism that makes both possible.

Why building-first records leave gaps

Consider what a building-first record quietly omits. It tends to have nowhere natural for:

• The boundary itself, and the disputes a defined one prevents.
• Retaining walls, slopes and made ground.
• Trees, including any carrying preservation orders.
• Drainage and sustainable drainage features that must be inspected and proven.
• Buried services, surveyed once and then forgotten.
• Fire service access routes that are part of the building's own safety case.
• Biodiversity net gain commitments running for decades under the Environment Act 2021.

Each of these is a real obligation. In a building-first system they survive on the goodwill and memory of whoever happens to know about them, which is exactly the knowledge most likely to be lost at a handover. The site shell turns memory into structure, because the slot is there whether or not anyone remembers to fill it, and an empty slot is a prompt rather than a silence.

It matters most for the things you cannot see

The strongest argument for starting with the site is that the riskiest parts of a site are often the least visible: what is underground, what holds back the slope, where the fire service can actually reach. A building-first record naturally documents the visible structure well and the invisible context poorly, because the structure is what is in front of you. A site-first record forces the invisible to be acknowledged, because the slots are there waiting to be filled and an empty slot is a visible question rather than a silent gap.

This is why we have argued that you should know what is buried before you dig and treat drainage and SuDS as a proven inspection regime. Both depend on having somewhere to hold the knowledge against the site. Without a shell, that knowledge has no durable home and tends to be re-bought every few years.

The handover dividend

The clearest payoff comes when the site changes hands. A new manager who inherits a site shell inherits a map: the boundary, the buildings, the land risks, the inspection histories, all attached to one outline. A new manager who inherits a building record inherits the structure and a set of blind spots they will spend months discovering, usually one unwelcome surprise at a time. The site shell is, in effect, the difference between handing over a complete picture and handing over the easy half of one.

It also scales cleanly. A portfolio of sites, each a defined shell holding its buildings and land, is a portfolio you can reason about. A portfolio of buildings with the land managed informally around the edges is a portfolio with a consistent, invisible gap running through it.

Start with the container

The recommendation is simple and a little contrarian: do not start with the building. Start with the site. Draw the boundary, create the shell, and let the building be the first thing you place inside it, with the land risks taking their proper slots from day one. It costs almost nothing at the outset and it closes the gap that building-first records carry for their whole lives. In SAMRISK, every plan includes a free site shell for exactly this reason, so a building always begins inside a site that has room for the ground it stands on, and the building is one well-managed element of a well-managed whole rather than an object in a void.