Six myths about AS 2870 site classification, and what's actually true

A site classification looks like a single letter on a report, and that simplicity breeds a lot of confidence that isn't always earned. The letter (A, S, M, H1, H2, E or P under AS 2870) tells a builder how reactive the ground is and how the footings need to respond. But plenty of decisions get made on the back of half-understood beliefs about what that letter means, who it applies to, and when you can lean on someone else's.

Here are six of the most common, sorted out.

Myth 1: The classification is a property of the block

It feels like it should be. The soil is the soil, so surely the class belongs to the land.

What's true is that the ground behaviour is fixed. What's not fixed is the classification, because AS 2870 classifies a site for a particular type of construction under a particular climate zone. The same clay profile can sit differently depending on the depth of the founding soil, the presence of trees, drainage, and whether you're building a slab-on-ground or a suspended floor. Change the design and the reactive movement the footing has to tolerate changes with it.

So a classification is a snapshot of ground plus intended building plus site conditions on the day it was assessed. Treat it as tied to a proposal, not stamped on the dirt forever.

Myth 2: A neighbour's soil test will do

The lots are next to each other. Same street, same estate, same builder pouring slabs down the row. Why pay again?

This one is half-right often enough to be dangerous. In a broad-acre estate on uniform cut-and-fill, adjacent lots frequently do classify the same, and a desktop look at the regional geology will tell you whether that's plausible. Where it falls apart is at the edges: a lot that catches the transition from one geological unit to another, a block with a filled gully running through it, an old dam, a change in slope. Reactive soils and building damage almost always trace back to a spot where the ground wasn't as uniform as the surface suggested.

Use neighbouring data to form an expectation, not to skip the test. If the desktop picture shows a consistent unit under the whole area, one lot's result is a reasonable guide. If there's a boundary or old disturbance nearby, assume nothing.

Myth 3: More reactive always means more expensive, so a lower class is a win

An M costs less to build on than an H2, so an M result is good news. Roughly, yes.

Where the logic breaks is when a class comes back lower than the ground deserves. A borehole placed in the wrong spot, or a test done in a dry season when the clay had already shrunk back, can under-call the reactivity. The cheaper footing gets designed, the slab goes in, and the movement the design never accounted for shows up as cracking a few summers later. The saving evaporates.

The class you actually want is the correct one, not the lowest one. A slightly more conservative footing is far cheaper than a remediation. Read the report for how many test locations were used and where, not just the letter at the top.

Myth 4: Fill just means you dig deeper

Hit fill, dig through it to natural ground, found on that. Problem solved. Sometimes.

Uncontrolled fill is one of the most common reasons a site ends up classified P (problem) rather than getting a clean reactive class. If the fill is deep, poorly compacted, or contains rubbish and organics, you can't reliably found on it, and digging to natural isn't always shallow or cheap. Old fill can also mask what's underneath and change the water movement through the profile. Before anyone books a rig, it's worth knowing whether there's likely fill on the block from historical aerials and records, because that changes the scope of the investigation you commission.

Don't assume fill is a minor dig. Find out early whether it's there, how deep it might run, and whether it's controlled or not.

Myth 5: Class P is a fail

Get a P and the deal is off, the block is unbuildable. Not so.

P simply means the site can't be classified in the ordinary A-to-E scale, so the footing has to be engineered specifically rather than picked from the standard tables. Reasons range from deep uncontrolled fill and soft or unstable soil to steep slopes, mine subsidence, or abnormal moisture. Plenty of P sites get built on every day; they just need a purpose-designed solution such as piers to a competent stratum, a stiffened raft, or a suspended floor.

P means "needs specific engineering," not "walk away." It does mean budget and design certainty come later and cost more, so factor that in before you commit.

Myth 6: A desktop study can give you the classification

The geology, the soil landscape mapping, the bore records, the flood and fill history all sit in databases. Surely that adds up to a class.

A desktop study is genuinely powerful for predicting the likely class, scoping the investigation, and catching the things a driller on the day might not think to check. What it can't do is measure the soil in front of you. AS 2870 classification for a footing design relies on sampled, tested material and characteristic surface movement calculated from it. No amount of mapping substitutes for that sample when a slab depends on it.

So the desktop and the field test aren't rivals. The desktop tells you what to expect and what to look out for; the site investigation confirms it and gives you the classification you can build to. Used together, the desktop stops you buying a test that misses the real problem.

Where a desktop report earns its keep

Every one of these myths gets easier to avoid when you look at the ground before the rig turns up. Regional geology, soil landscape units, historical fill, groundwater and bore records will tell you whether a neighbour's result is trustworthy, whether fill is likely, and whether the site is heading towards a P before you spend a cent on drilling.

To see what the ground under your site is likely doing, pop in your address or see a sample report and read the profile before you commission the test.

LayeredGeo compiles geology, soils, groundwater and site data into an automated geotechnical desktop report for any address in Queensland, New South Wales or Victoria.

About this article. This article was drafted by an AI writing tool using LayeredGeo's data and published automatically. It is general information about how property and ground conditions are commonly assessed in Australia. It is not engineering, planning, legal or financial advice, it is not specific to any property, and it may contain errors or be out of date. Check anything that matters against the source dataset or a suitably qualified professional before you rely on it. If you spot something wrong, tell us at hello@layeredgeo.com.au and we will fix it.

← All articles