Two winters ago I helped reset 34 metres of border along a south-facing wall in Herefordshire. The owner had bought flexible steel the summer before. He pushed it into bare soil with a rubber mallet. By February, every second section had lifted above the turf. The metal was fine. The trench was the problem. I dug it out, re-cut the line 150mm wide and 200mm deep, and bedded the strip on compacted grit. It has not moved since.
Everything Rests on the Trench
Almost every edging failure I am called to is a base failure. Material choice gets all the attention. Installation decides the outcome. A steel strip pressed into loose loam will creep and tilt. The same strip set on 50mm of compacted grit, with stakes driven below the frost line, holds a straight line for a decade or more.
Depth and width both count. I cut new borders 150mm wide so a stake and a shoulder of soil fit beside the strip. I take the trench to 200mm deep wherever a mower wheel or a boot will cross it. For a plain spaded edge, the RHS guide to creating a lawn edge works to a drop of about 7.5cm (3in) between lawn and border. That drop keeps grass off the bed and mulch off the grass. If the cut has to contain something as tough as bluegrass, Illinois Extension suggests a trough roughly 4 inches deep and wide.
Turf does not respect a line for long. Illinois Extension reckons grass advances 3 to 4 inches into a bed every year. A spaded edge wants re-cutting once or twice a season. That is why a permanent strip earns its money on long runs.
Steel: Thin, Sharp and Married to Its Coating
Steel gives the crispest line you can buy. It also lasts longest. Illinois Extension names steel and aluminium the most durable edging materials, and recommends steel where a great deal of side-strength is needed, such as beside a driveway. Thickness drives stiffness. The flimsy rolls sold as flexible steel are often 1mm or thinner. I will not fit them. The useful range starts near 1.5mm and stiffens properly at 2mm. Virginia Cooperative Extension reports that galvanised steel at 14-gauge, about 1.9mm, can run 35 to 50 years before the first perforation.
Coating is the weak point. Powder-coated and galvanised steel stay rust-free only while the surface is intact. Cut a strip to fit a curve and you expose bare metal at the end. Rust starts there. A blister then lifts the coating from underneath. So touch every cut end with zinc-rich paint before you bury it. Keep the string trimmer off the top edge too. RHS advice on maintaining lawn edges stresses sharp tools for a clean cut, and a clean cut is kinder to the coating than a blunted blade dragged along it.
Frost is the other enemy. Illinois Extension flags frost heaving as a known problem with steel edging. Deep staking and a coarse, free-draining base reduce it. Wet clay in a shallow trench makes it worse.
Aluminium, and Where It Bends Better
Aluminium never rusts. That one property makes it the sensible pick for a wet site or a coastal one. It is lighter than steel, so a 3m length is easy to carry and shape. It also takes a tighter radius without kinking. The trade is stiffness. Aluminium gives under a mower wheel in a way 2mm steel does not. Profiles run from about 1.5mm up to 3mm. Illinois Extension notes that aluminium is easier to work with than steel, though the bare silver finish can look out of place against soil and planting. Anodised black solves that.
One detail people miss. Aluminium edging often ships with aluminium stakes. Those pull out of soft ground. Swap in galvanised steel stakes on any run with a curve in it.
Timber Borders and the Rot Clock
Timber looks right in a cottage garden. It comes with a clock ticking from the day it goes in. Wood in damp soil rots, and no finish stops that for long. Species sets the pace. Oak and sweet chestnut hold up best. Larch and Douglas fir sit in the middle. Untreated pine and fir go quickest. Georgia’s CAES Field Report describes eastern red cedar as rot-resistant and long-lived in direct soil contact, while pine has almost no resistance and a very short life there.
I sank a larch board 150mm into a wet bed four years ago to watch what happened. The top half is sound. The buried section has gone soft enough to push a screwdriver through by hand. That is the honest picture. Unless the timber sits in air, the base rots first.
Pressure treatment changes the chemistry, not the physics. OSU Extension lists the modern preservatives: copper azole, micronized copper azole and alkaline copper quaternary. All are copper-based. The arsenic-based preservative, chromated copper arsenate, left residential use two decades ago, and the EPA records the manufacturers’ voluntary withdrawal in December 2003. The old arsenic scare now applies mainly to salvaged timber. NPIC is blunt about the rest: CCA, creosote and pentachlorophenol are not registered for consumer use, garden beds included.
Copper is still copper. University of Maryland Extension says micronized copper azole carries no arsenic but does carry copper, then offers cover: paint the timber, stain it, or run a heavy plastic liner between the frame and the soil. For an edible bed I would not lean on treated timber at all. UGA points out that USDA Organic rules bar pressure-treated wood from touching food plants. Cedar, oak, or a raised aluminium bed dodges the question.
Brick and Stone Need a Real Base
Brick laid straight onto soil shifts. I have re-laid that mistake more than once. The worst was a 12m run in Shropshire, set on bare clay, where the bricks had stepped 30mm out of line by the second spring. On a compacted limestone base, brick stays put. Illinois Extension advises well-fired brick so it does not crumble when wet, plus a sand or concrete base for strength. Stone needs bulk to block grass. The same source suggests pieces at least 3 inches thick and 4 inches wide.
Mortar quantities catch people out. A 20kg bag of dry-mixed mortar beds about four to five linear metres of a single brick course at 50mm depth and 100mm width. A 30m border therefore wants six or seven bags for bedding alone, before any pointing. Dry-lay first. Cutting brick to a curve is slow, and a radius under about 1.5m usually needs a diamond blade rather than a bolster.
Plastic and Composite, Without the Sales Talk
Plastic roll edging is the cheapest thing on the shelf and the shortest-lived. It is fine for holding mulch at the foot of a hedge. It is not fine for a lawn edge. UV makes it brittle in three or four seasons. Frost lifts it. Then the mower catches the lifted lip. Penn State Extension describes a slope where the flimsy plastic edging sagged and buckled, replaced in the end with stiff, properly staked sections.
Rigid plastic and composite lumber are better products. Virginia Cooperative Extension notes that HDPE lumber resists degradation for decades. Composite boards, a mix of wood fibre and plastic, carry a caveat. Manufacturers do not recommend them in direct soil contact, and the wood fraction can still decay. Keep composite above grade and support it. One more point from Penn State Extension: skipping plastic hardscaping such as landscape edging cuts fossil-fuel use, because most weatherproof plastics come from petroleum.
Mowing Edge, Retaining Edge, Decorative Edge
Work out the job before you buy a single length. A mowing edge has to sit flush. Its top rides just below blade height, so the deck glides over it and you stop trimming the line by hand. RHS low-maintenance advice suggests a permanent edge of bricks or paving stones, or small gravel set lower than the lawn, for exactly this reason. Stiffness and a stable base matter most here.
A retaining edge is a different animal. It holds soil back, so it feels sideways load and needs depth below grade. Steel wins. Illinois Extension recommends steel where a great deal of side-strength is required, such as beside a driveway. Anything above about 300mm of retained height should be designed, not improvised. It also needs drainage behind it. RHS guidance on permeable paving notes that hard, non-porous surfaces push rainwater run-off up by as much as 50 percent. Trapped water is what topples small walls.
A decorative edge only has to look deliberate. Set stone or brick slightly proud and let plants spill over it. That is a design choice, and it costs nothing extra.
Width matters on a mowing strip. A 100mm band of paving is too narrow for a wide deck to ride on evenly. Go to 150mm and it works. Set the band dead level with the lawn, or a fraction low, then mow straight across it. If the soil beneath is soft, a wheel sinks at the joint and scalps the turf beside it. Compact a 50mm bed of grit first. Bed each paver on that grit and tap it level with a mallet. Brush kiln-dried sand into the joints afterwards.
What It Costs, and What Drives the Number
Treat every per-metre figure as a band. The drivers move more than the material does. Run length, access, soil type, how much of the edge curves, and whether you dig it yourself all shift the total.
As rough material bands per linear metre, plastic roll edging sits at the bottom. Flexible steel runs several times that. Cast or mortared brick and stone land above both. Virginia Cooperative Extension published a materials-and-cost breakdown for raised beds and the pattern is plain: the same growing volume costs very different sums depending on the wall material, and HDPE plastic lumber came out the most expensive of the lot, with metal kits near the bottom.
Installed cost is a different story. Labour usually beats materials. A 30m run of steel on a straight boundary is a day’s work for two people with the right tools. The same 30m on a tight curve against a hedge could take three days. Add the cost of carting away spoil. That line rarely appears in a quote.
Steel prices swing on gauge and finish more than on length. A 2mm powder-coated strip costs more per metre than a 1.2mm galvanised one. Long runs buy a small discount per metre. Curves cost more, because every cut end needs sealing before it is buried.
Five Years On
I keep a short list of edges that have outlived their owners’ patience. A 2mm powder-coated steel border along a gravel drive, set at 200mm depth, is still true after nine years. A run of reclaimed brick on 50mm of limestone, reset once, is still crisp. And two lengths of untreated pine that lasted four seasons before the base turned to mush.
The failures taught me more. Flexible steel pushed into soil with a cut end left bare, rusting at the joint within two winters. A curved brick edge on bare clay, heaving 30mm out of line. Plastic roll edging in Cumbria that popped its stakes after a single hard frost.
None of those were material failures. Every one was a base failure. Get the trench right, compact the base, and set the top edge at the right height. A modest material laid properly will beat an expensive one laid badly. Get it wrong and no budget on earth will save the line.