Crack Monitoring and Level Monitoring: What 12 Months of Readings Tell Your Engineer
Monitoring is the slow, unglamorous stage of a subsidence investigation, and the one that decides whether you spend a few thousand pounds or tens of thousands. Here is how it is done, why tell-tales are not enough, and what the four patterns of readings mean.
In short: a crack tells you a building has moved; only monitoring tells you whether it is still moving, and that is the question every repair decision hangs on. Crack monitoring measures the width of each significant crack from fixed steel studs with a digital calliper, to a tenth of a millimetre, every few weeks. Level monitoring measures the height of the foundations themselves from markers around the building with a precision level, every six to eight weeks. Six to twelve months of readings, spanning a summer and a winter, separate the four patterns that matter: seasonal (clay and trees), progressive down (drains or soft ground), progressive up (heave), and static (historic, repair it). Glass tell-tales cannot do this. Cracks are repaired when the readings go flat, not before.
Why the trend matters more than the crack
Engineers grade cracks by width, using the six categories in BRE Digest 251 that our guide to types of cracks in walls sets out, but the categories come with a caveat that matters more than any of them: a crack that has stopped moving can usually be lived with or simply repaired, whatever its width, while a crack that is still growing needs its cause found, whatever its width. A static 6 mm crack in a Victorian terrace is a repointing job. A 2 mm crack that was 1 mm in the spring is an investigation. The only way to tell them apart is to measure, repeatedly, over enough time for the cause to reveal itself.
The commercial logic follows. Most subsidence is resolved by dealing with the cause, a drain, a tree, and repairing the cracks once the ground has settled; a minority needs underpinning. Monitoring is what puts a case in the right group. It feels slow, but it is far cheaper than underpinning a wall that stopped moving years ago, or than repairing cracks that reopen because the ground is still on the move.
Crack monitoring: how it is actually done
The first visit produces a crack record: a sketch of every elevation and a plan of every room, with each significant crack drawn on, its maximum width measured, and the point of measurement marked so the next reading is taken in the same place. Old, repointed and filled cracks are recorded too, because historic damage that does not match the current pattern is a clue in itself.
Monitoring points are then fixed at the widest part of each crack that matters. The reliable method is a pair of small stainless steel studs glued to the masonry or plaster either side of the crack with a fast-setting epoxy, read with a digital calliper that measures the distance between them to 0.1 mm. Three studs in a triangle allow movement along the crack as well as across it to be resolved. A photograph and a note of the temperature and weather accompany each reading, because brickwork expands and contracts and a warm afternoon can close a fine crack by a fraction of a millimetre on its own.
What about tell-tales? The glass strip glued across a crack that snaps if the crack moves is not a monitoring device: it records that something happened, not when or how much, it breaks from a knock or a cold night, and it cannot show a crack that is closing, which is half the story in a clay case. Calibrated plastic tell-tales with a printed grid can be read to about half a millimetre and are acceptable for a simple watching brief. For a diagnosis that will be relied on by an insurer, a lender or a buyer, studs and a calliper are the standard.
Level monitoring: watching the foundations, not the crack
Cracks are a symptom in the superstructure; subsidence and heave are movements of the foundations. Many engineers therefore regard level monitoring as the more important of the two, because it measures the thing that is actually moving, shows which parts of the building are involved and by how much, and is the only way to prove that tree works have done their job.
Markers are small brass or stainless screws set in plastic plugs in the brickwork near ground level: one at every corner, two or three along each wall, at the corners of bays and porches and at the junction between the original house and any extension, and where possible on internal load-bearing walls. At each visit a precision optical level and a staff that seats accurately on each marker read the height of every point to 0.1 mm, closing the circuit back to the start with an error of less than a millimetre. Readings are taken relative to a datum on a part of the building that is not moving, or, if the whole building may be moving, to a deep datum founded at least 3 m down in ground outside the influence of trees and drains. A manhole surround can serve; a manhole cover, which gets lifted, cannot.
Readings every six to eight weeks are usual, plotted as movement against time for each marker. Precision is everything: the same operator, the same instrument, the same set-up position, because the differences being looked for are a few millimetres over a season. A one-off distortion survey along a brick course near the damp-proof course, and a verticality survey of the walls with a plumb line or total station, is often done at the outset to map how far the building has already moved historically, on the assumption it was built level, which is usually a safe one.
The four patterns, and what each one means
Plotted over a year, foundation and crack readings sort themselves into a small number of shapes, and the shape points at the cause before a spade goes in the ground.
| Pattern in the readings | What it usually means | What happens next |
|---|---|---|
| Down through summer, back up through winter | Shrinkable clay drying around trees or large shrubs and rewetting; the amplitude grows in a dry year | Identify the tree, trial pit for roots and moisture profile, arboricultural advice, crown reduction or removal, then monitor through at least one more summer |
| Steadily down, one way, sometimes faster after rain | A leaking drain softening or washing out the ground, soft or filled ground consolidating, or erosion | CCTV drain survey and water test, trial pit beside the low point; repair the drain and keep monitoring while the ground consolidates, up to a year |
| Steadily up, at a slowing rate | Heave: clay taking water back after a tree was removed or died, or a leak into dry clay | Moisture profile and suction tests to gauge how deep the dry zone is and how long the swelling will run; wait it out or underpin with anti-heave measures |
| Flat, within about a millimetre, no trend | Historic movement that has stopped, or normal thermal and seasonal flexing of the fabric | Repair the cracks and redecorate; keep the markers in place so any future concern can be checked quickly |
Real readings are noisier than the diagram and are read alongside the crack record, the trial pits and the drain survey. A pattern is a lead, not a verdict.
How long it takes
Long enough to see a summer and a winter: six to twelve months is the normal span, and on an insurance claim it is often longer, partly because of the queue in a busy year. The reason is the physics of clay. A tree-and-clay problem announces itself in August and goes quiet in February; a leaking drain does not care what month it is. Readings that cover only one of those seasons cannot tell the two apart, which is why short monitoring is so often repeated. Where a tree has been reduced or removed, the readings continue through at least one further growing season to prove the movement has been brought within what the building can tolerate, and the markers are then left in place with a reading each spring and autumn while the pruning regime continues.
Timing the start matters too. Studs should go on and markers in as soon as subsidence is suspected, not after months of debate, because the first readings are the baseline and every month without them is a month of evidence lost. Boreholes for the moisture profile are best drilled in autumn, when the drying is at its deepest.
Other ways of watching a building
Visual monitoring means repairing and redecorating immediately after the cause has been dealt with, and using the fresh finishes as the monitor: if nothing reappears, the risk paid off; if cracks return, the repair money is lost and the formal exercise starts anyway. It suits cases where the cause is clear and the owner would rather not wait. Remote sensors, fixed to the building and reporting through the mobile network, give continuous readings with temperature alongside, useful on complex or contested cases, at a price. Satellite radar can now track ground movement across a city, but its resolution does not yet resolve the seasonal millimetres of a single house, and it confuses thermal expansion of the building with movement of the ground. For the foreseeable future the calliper and the level remain the tools.
What you get at the end
A report that plots every marker and every crack against time, interprets the pattern, states the cause consistently with everything else the investigation found, gives a prognosis if nothing is done, and sets out the options with their merits and their likely cost. That report is the basis of the repair specification, of any underpinning design, and of the Certificate of Structural Adequacy that closes the job; it is also the document an insurer, a lender or a buyer's surveyor will want to see. On an accepted insurance claim the insurer normally funds the monitoring; outside a claim our fee for installing the studs and markers and taking three seasonal readings is £1,200 to £2,000 + VAT depending on the size of the property, with further visits at £350 + VAT each, and it remains the cheapest part of a subsidence job by a wide margin. Our guides to subsidence repair costs and to subsidence insurance claims put it in context.
Where EMA Structures fits
We set up and run crack and level monitoring as part of a subsidence assessment: the crack record and level survey at the first assessment (£495 to £895 + VAT with a written report, depending on the property), studs and markers installed where movement is suspected and three seasonal readings taken by the same engineer with the same instrument (£1,200 to £2,000 + VAT, further visits £350 + VAT each). Trial pits, drain surveys and root identification are not included in either fee; they are quoted separately, chosen to test what the readings suggest. Because we also design the repairs and any underpinning, and our specialist construction arm can deliver them under engineering oversight, the monitoring is aimed at the decision that matters, not at generating readings. Send us photographs of the cracks and we will tell you whether monitoring is the right next step.
General guidance on UK practice as at September 2026. The timescale and scope of monitoring for any property depend on its cracks, ground, history and any insurance claim, and are set by the engineer. Fees exclude VAT.
Monitoring questions
What is crack monitoring?
Measuring the width of a crack at fixed points over time to find out whether it is opening, closing, cycling with the seasons or staying still. It is done by fixing small stainless steel studs to the wall either side of the crack and measuring between them with a digital calliper to a tenth of a millimetre at each visit, with the positions recorded on a sketch of the elevation. Glass tell-tales glued across a crack are not accurate enough for a diagnosis, although calibrated plastic tell-tales can be used. Readings are normally taken every four to eight weeks for six to twelve months.
How long does crack monitoring take?
Six to twelve months is typical, because the readings need to span a summer and a winter to tell seasonal clay movement from progressive movement. Insurance claims often run longer, and where a tree has been pruned or removed the monitoring continues through at least one more summer to prove the works have worked. Short monitoring, a few readings over a couple of months, rarely settles anything and often has to be repeated.
Are tell-tales accurate enough to monitor a crack?
Glass tell-tales, a strip of glass glued across the crack that breaks if it moves, are not: they only show that something moved, not when or how much, they break from knocks and temperature, and they say nothing about a crack that is closing. Calibrated plastic tell-tales with a grid can be read to about half a millimetre and are acceptable for simple cases. For a diagnosis that will decide between a repair and underpinning, engineers use fixed studs read with a digital calliper to 0.1 mm, usually alongside precise level monitoring.
What is level monitoring?
Measuring whether the foundations themselves are moving, rather than whether a crack is opening. Small metal markers are fixed near ground level at every corner, at two or three points along each wall and at the junctions of any extension, and a precision level reads their heights to 0.1 mm relative to a stable datum at each visit, typically every six to eight weeks. Plotted against time, the readings show which part of the building is moving, by how much and in which direction, and the pattern usually points to the cause.
What do the monitoring results show?
Four patterns cover most cases. Movement down in summer and back up in winter points to shrinkable clay and vegetation. Steady one-way movement down, sometimes faster in wet weather, points to a leaking drain, soft or filled ground, or erosion. Steady movement up at a slowing rate is heave, usually after a tree has been removed. Readings that stay within a millimetre or so with no trend mean the movement is historic and has stopped, and the cracks can be repaired. Everything else in the investigation, trial pits, drain surveys, root identification, is chosen to confirm what the pattern suggests.
Can I monitor cracks myself?
You can and should keep a record: pencil lines across the crack with the date, and photographs with a ruler or tape in shot, taken from the same spot each month. That is enough to show whether a crack is obviously growing and to give an engineer a history. It is not enough to diagnose the cause or to satisfy an insurer, lender or buyer, because it cannot detect the fractions of a millimetre that separate seasonal from progressive movement, and it says nothing about the foundations. Engineer-led monitoring adds the studs, the calliper and the levelling, and the interpretation.
Do I have to wait for monitoring before repairing the cracks?
For anything beyond hairline decoration cracks, yes. Filling and repointing a crack while the movement that caused it continues means the crack reopens, the money is spent twice and the evidence of how the crack behaved is lost. The exception is where an engineer has already established the cause and dealt with it and the risk of further movement is judged low; some owners then repair straight away and use the fresh finishes as the monitor, accepting that if movement recurs the repair is wasted.
Is the crack still moving?
Send us photos of the cracks with anything you have noted about when they appeared, and we will tell you honestly whether monitoring is the right next step.