The topcoat gets the credit. The client chose the colour, watched it go on, and ran their hand over the finish when you were done. Nobody photographs the primer. Nobody asks which one you used. And yet, the coat that decides whether a job holds for a decade or fails in six weeks is almost always the one underneath — the one nobody sees.
This is written for the working decorator, and it makes an argument you already half-know from the callbacks: the primer governs the things that actually go wrong. Adhesion, stain-blocking, durability — all three live in the primer, not the finish. Get the topcoat wrong and you have a colour you don't love. Get the primer wrong and you have a failure: peeling, flashing, tannin bleed, knots ghosting through a fortnight after handover. The finish is cosmetics. The primer is engineering.
What the primer is actually doing
A topcoat has one job that matters to the client — colour and sheen — and one job that matters to you: hanging on to whatever is beneath it. It is not built to bond to a difficult substrate, block a stain, or seal a porous surface. It assumes that work is already done. The primer is the layer that does it.
Three functions, and the right primer is chosen for whichever of them the substrate demands. Adhesion: the primer grips a surface the topcoat could never hold, and presents a face the topcoat can key into. Sealing: it closes off a porous or absorbent substrate so the topcoat sits on top and cures evenly instead of sinking in and flashing. Stain-blocking: it forms a barrier the migrating contaminant — tannin, resin, nicotine, water-borne salts — cannot pass through. A single substrate often needs two of those at once, which is why primer selection is a reading exercise, not a default.
The mistake that generates most of the trade's callbacks is treating primer as a generic undercoat — one tin, every surface. Bare plaster, glossy old skirting, a melamine cupboard, and a cast-iron radiator are four different chemistries with four different failure modes, and one primer cannot answer all of them. The decorator who reads the substrate and matches the primer to it is the one who doesn't go back.
The chemistries, and what each is for
There are four primer chemistries you reach for in UK domestic work, and knowing what each does — and what it cannot do — is most of the skill.
Water-based acrylic primers are the everyday default. They are low-odour, fast-drying, clean up in water, and they prime previously painted plaster and sound softwood without drama. They seal moderately, key reasonably, and stay white without yellowing. What they do not do is block tannin or resin, and they do not grip a slick non-porous surface. The tannin molecule is small and water-soluble; it diffuses through a porous acrylic film as if the film were not there. Asking a standard acrylic to hold back oak tannin or pine knots is asking it to do a job it has no chemistry for.
Shellac primers are the universal blockers. Shellac — the lac-bug resin dissolved in alcohol — dries hard and fast and forms a barrier that stops just about anything migrating through it: tannin, resin, nicotine, water stains, even smell. In the UK trade, the canonical product is Zinsser B-I-N, a pigmented white shellac primer, and it is the gold standard for problem substrates — knots, MDF faces, nicotine-stained ceilings, water-damaged plaster. It is thin, it dries in around an hour, it sands easily, and it bonds to almost everything. Two caveats that matter on the job: it yellows under UV, so it must always be topcoated rather than left as a finish; and it is a stain block, not a damp barrier — it seals stains but will not hold back moisture pressure, and nothing is gained by relying on it where the real problem is damp.
Oil and alkyd primers — Zinsser Cover Stain is the common UK example — sit between the two. Oil-based, slower to dry, they block most water stains and tannins well and are the economic choice over a large stained area where shellac would be expensive. They flow and key beautifully on timber. The trade-off is the trade-off of all oil paint: solvent odour, longer recoat windows, and yellowing over time, which is why the modern instinct is to reserve them for where their stain-blocking and levelling earn their keep.
Bonding and adhesion primers are the answer to surfaces with no tooth. A bonding primer chemically grips a smooth, sealed, non-porous face — melamine, laminate, glazed tile, UPVC, high-gloss existing paint — and presents a surface the topcoat can hold. Zinsser Bulls Eye 1-2-3 (water-based) is the standard reach for melamine and plastic; Zinsser Allcoat is a water-based universal primer-undercoat used on glossy laminates and veneers where heavy sanding is impractical. On bare and non-ferrous metal you are into the metal-specific end of this family — etch primers and mordant washes that bite into the surface chemically because nothing mechanical will hold. The defining feature of this group is that they solve adhesion, the one thing a topcoat can never solve for itself.
Bare softwood: knots will win if you let them
Softwood — pine, spruce, fir — is the standard UK joinery timber, and its defining problem is the knot. Knots are dense resinous heartwood, and the resin and the tannin around them are mobile. Moisture from humidity, condensation, or central heating cycling reaches the wood, dissolves the tannin, and carries it out through the paint film as a brown or amber stain; heat softens the resin and pushes it up as a greasy halo. A white door with knots in the rails, primed with standard acrylic and topcoated, will show amber ghosting at four to six weeks and look unprimed within two years. The film is sound. The colour is right. The job has failed.
The fix is shellac, applied as a barrier before the ordinary primer goes on. Spot-treat each knot and resin streak with shellac knotting solution, or prime the whole surface with B-I-N and let the shellac chemistry do the blocking. Either route works; what does not work is a porous acrylic primer over bare knots, because the tannin walks straight through it. The sequence is sand, dust off, knot, prime, undercoat, topcoat — and the knot step is not optional on bare softwood. It is thirty minutes of insurance against a half-day callback, and on new pine doors or frames it is non-negotiable regardless of how small the knots look. Pine bleeds through standard primers whatever the knot size.
Hardwood, MDF, and the boards in between
Oak is the hardwood that catches people out, and the mechanism is tannin rather than resin. Oak is tannin-rich, standard primers do not block it, and the result under a pale topcoat is brown or orange discolouration bleeding up through the finish. A tannin-blocking primer — shellac, or a dedicated tannic-acid primer — is essential before painting oak, and because hardwood is dense and does not drink primer the way softwood does, adhesion depends on mechanical keying: sand thoroughly to 120 to 150 grit to give the primer a profile to grip. Skip the sanding on hardwood and you get adhesion failure regardless of which primer you reach for.
MDF is a different animal again, with two distinct problems. The faces come pre-sealed from manufacture and prime up cleanly, but a shellac primer earns its place here because it seals fast, blocks the formaldehyde odour off-gassing from the board's resin binder, and gives an excellent key in about an hour. Water-based primers can raise the fibres on a face that has been sanded or damaged, leaving a texture you then have to flatten. The real trap is the cut edges. Raw MDF edge is exposed fibre core, and it drinks primer like a sponge — one coat looks fine wet and dries to a chalky, hairy mess that sucks every subsequent coat just as hard. Seal the edges first, with a dedicated MDF primer, a water-based shellac, or diluted PVA, then prime the whole piece. The edge is the single most visible failure point on painted built-in furniture, and it is entirely preventable.
Veneered ply and laminate boards follow from the face species and the face material — pine-faced ply behaves like pine and wants knotting, melamine-faced board behaves like melamine and wants a bonding primer. Read the face, not the core.
Melamine, laminate, and the surfaces with no tooth
Melamine-faced board — flat-pack furniture, budget kitchen carcasses, wardrobes — has a hard, non-porous, manufacture-sealed surface with no tooth for mechanical keying. Standard paint will not adhere to it, full stop. Prime it with acrylic and the film peels off in sheets with no more provocation than a fingernail, because the primer wets the surface but never embeds or bonds. The failure is complete and fast, usually within weeks, accelerated by the humidity and temperature swings of a kitchen.
The answer is a bonding primer that chemically grips the smooth face — Bulls Eye 1-2-3 is the standard, with B-I-N as the shellac alternative where its yellowing is acceptable under the topcoat. One coat of bonding primer, two coats of topcoat, and a glossy bright melamine wants two primer coats rather than one. UPVC presents the same adhesion challenge and the same answer: a specialist UPVC or water-based bonding primer. Some all-in-one products claim direct application to these surfaces without a primer — test them on an inconspicuous patch before you commit a whole kitchen to the promise on the tin. Glazed ceramic tile is the same problem in a wetter setting: degrease hard, sand the glaze to a key, and use a tile-specific primer-and-topcoat system, while being honest with the client that painted tile is a cosmetic improvement with a limited life, not a substitute for the original glaze.
Previously glossed surfaces: tooth before everything
Paint will not stick to a shiny surface. A gloss or satin film is slippery at the molecular level; the new coat sits on top rather than embedding, and peeling is the inevitable result. This is the second most common adhesion failure on interior work after damp, and the prevention is simple: break the gloss. Sand it dull — 120 grit minimum on gloss, 180 on eggshell or satin — or use a liquid degloss such as Owatrol ESP, which softens the film chemically and avoids the dust, an advantage that matters on anything pre-1992 that might carry lead. Then wipe off every trace of dust with a damp cloth, let it dry, and prime.
There is a compatibility trap layered on top of the gloss trap. Water-based topcoat over oil-based undercoat that has not fully cured will not key, because the emulsion cannot wet into the hydrophobic oil binder — the two systems are incompatible and the result is peeling or fine crazing as the topcoat shrinks against a substrate it never bonded to. If you are crossing from an old oil system to a water-based one, the oil must be fully cured and mechanically sanded, and the right primer in between removes the ambiguity. Respecting the system — compatible primer and topcoat, or a deliberate isolating primer where you change lanes — is what keeps the film whole.
Metal: ferrous, non-ferrous, and galvanised
Metal is where primer stops being optional and becomes the entire performance of the system. Standard wall paint on a radiator yellows, cracks, and flakes because the film cannot expand and contract with the steel through daily heat cycling; the substrate needs a flexible, heat-tolerant system and a rust-inhibiting primer under it wherever there is surface rust.
Ferrous metals — iron, steel, cast-iron pipe — corrode, and the primer has to do two things: inhibit rust and bond to bare metal. All visible rust comes off first, by wire brush, abrasive, or a phosphoric-acid rust converter that turns rust into a stable paintable coating, and then a rust-inhibiting primer goes on immediately, because fresh rust can bloom again within hours in damp conditions. A primer carrying zinc or epoxy chemistry inhibits the corrosion the topcoat cannot touch.
Non-ferrous metals — aluminium, copper, brass — do not rust but they oxidise, and they are too smooth and inert for an ordinary primer to hold. An etch primer bites into the surface chemically, creating the bond mechanical keying cannot, and then standard paint follows. Galvanised steel is the case that most often goes wrong: bright new galvanising is too smooth and too chemically reactive for ordinary primers, which fail and peel. Two routes work — a mordant wash such as T-Wash, a phosphoric-acid solution containing copper salts that etches the zinc and turns it dark grey to confirm the reaction, overcoated with a metal primer within a couple of days before contamination re-forms; or a dedicated galvanised-metal or etch primer formulated to bond to zinc directly. Weathered, dulled galvanising is easier, but any white zinc-salt bloom must come off first. Across all three, the lesson is identical: the primer is doing the adhesion and the corrosion protection, and the topcoat is along for the colour.
New plaster: the mist coat is a primer in disguise
New plaster is actively porous — a network of capillaries built to absorb moisture — and until it has lost its water of crystallisation it cannot take a topcoat without trapping moisture behind the film and blistering within weeks. The visual test is the only reliable one: when the plaster has gone from dark pink to a uniform pale pink or off-white with no dark patches, it has dried enough to paint, which in ambient conditions is roughly four to six weeks, longer in a cold or unheated property.
The mist coat is the priming step here, and it is the one most often dropped under time pressure. A mist coat is standard matt emulsion thinned roughly 70 per cent paint to 30 per cent clean water, and its purpose is absorption, not coverage — the diluted coat soaks into the plaster and forms the key for the full-strength coats that follow. Applied full strength, emulsion sits on the surface with no mechanical grip and can later peel away in sheets. Contract matt, with minimal vinyl and maximum raw pigment, is the right choice for it; vinyl emulsions create a surface film and make poor mist coats even when vinyl is what you intend as the topcoat. Bare plaster patches left by filling are more absorbent than the painted plaster around them and will flash as dull patches in the finish unless you seal them the same way. Skip this and you get the two classic new-plaster failures at once — peeling from no grip, and flashing from uneven absorption.
Tannin, knots, and the woods that bleed
The bleed problem deserves naming as a failure mode in its own right, because it is the one most likely to come back through a finished, sound film and make a completed job look incompetent. Yellow, brown, or orange stains radiating from knots or scattered across boards are not mould and not a topcoat defect — they are tannin or resin being extracted from the timber by moisture or heat and carried up through the paint. Pine and Douglas fir are notorious; oak is moderate but tannin-rich enough to matter; cedar and redwood are rich in extractives.
The reason it defeats the wrong primer is chemical, not a question of film thickness: standard acrylic and even many oil primers do not block tannin, because the molecule is small and water-soluble and diffuses through a porous film. Only a barrier coating — shellac, bituminous, epoxy — stops it. So the prevention is a shellac knotting coat before the main primer, two coats over dark or obvious knots and one over minor ones, and once it has appeared the only honest remedy is to seal it at source: shellac over the stain, then topcoat, and a strip back to bare timber where it is severe. Over-painting a bleed stain without a barrier coat hides it for a few weeks and then it comes straight back through, which is the difference between fixing the failure and merely postponing the callback.
The failure modes, read backwards to the primer
Peeling and flaking is the headline failure, and it traces almost entirely to primer and prep decisions: a glossy substrate not keyed, a non-porous surface given a primer that cannot grip it, a damp wall painted before it dried, or incompatible oil-and-water systems stacked on each other. The film comes away in clean sheets because it was never bonded. The right primer — bonding for the slick surface, the compatible chemistry for the system, and crucially no primer at all over damp until the moisture is fixed — is what prevents it. Primer is not a substitute for drying; nothing is.
Flashing — random patches of different sheen and colour depth across a wall — is an absorption failure, and it is a primer-and-mist-coat failure specifically. Where bare plaster or filler drinks more paint than the primed area around it, the same topcoat reads duller over the thirsty patches. Priming every surface and every filled patch uniformly, and getting a proper mist coat onto new plaster, equalises the absorption so the finish cures to one sheen. The fault shows worst on dark colours, where unequal film thickness changes the optical density most.
Tannin and resin bleed is the barrier-coat failure already described — preventable only by shellac or another true barrier under the topcoat, never by the topcoat itself. And poor adhesion in its broader sense — the melamine that sheds its paint, the metal that flakes, the gloss that lets the new coat slide off — is the adhesion-primer failure: a surface with no tooth given nothing to bond to. In every one of these, the topcoat is the symptom and the primer is the cause. You cannot finish your way out of a priming mistake.
How the primer talks to the finish
The two coats are not independent, and the better you understand the conversation between them the fewer surprises you hand the client. The first thing the primer owes the finish is colour neutrality where it matters: a shellac primer yellows under UV, and under a thin bright-white topcoat that yellowing can ghost through, so on premium finishes where film thickness is light by design you either lay enough topcoat to cover it or spot-treat with shellac and use a non-yellowing primer over the rest. Allcoat and similar universal primers run slightly yellow-toned for the same reason and are not the first choice under very pale colours unless tinted.
The second thing is sheen continuity. A primer is matt or near-matt, and where the topcoat is eggshell or satin, any thin or missed primer coverage shows during application as dull spots and can telegraph into the cured finish if the coverage was patchy. Complete, even primer coverage is the precondition for an even topcoat, and a second primer coat over a thin first one is cheaper than a flashed finish. The third thing is that the finish reveals the substrate to the degree its sheen allows, and the primer-and-prep underneath has to be specified to the finish above it. Gloss forgives nothing — a filling line invisible under matt becomes a shadow line, and grain telegraphs as variable sheen — which is why a high-gloss job demands micro-sanding and meticulous priming, and why specifying eggshell instead is often the honest call when the substrate cannot justify the gloss preparation.
That interaction shows on real colours, and the database makes it concrete. Deep colours are the most sheen-sensitive, and the finish you choose changes the colour the client sees as much as the primer underneath determines whether it lasts. Farrow & Ball's Railings (#45494C), a very deep blue-black, reads as a soft off-black in matt and becomes a deep reflective blue-black with far more visible chroma in gloss — the audit flags it as very-high finish sensitivity, with flat finishes prone to looking dusty and eggshell sharper. Pitch Black (#3D3C3C) behaves the same way, its warm brown undertone reading velvety in matt and vanishing under an intense reflective near-black in gloss. Benjamin Moore's Hale Navy (#434B56, LRV 8 on the manufacturer's published figure) is described as velvety in matt, lifted in the blue by eggshell, and revealing intense chroma in gloss — the same pigment reading as two entirely different colours depending on the finish. None of that character survives if the primer beneath has let the film flash, peel, or bleed. The finish decides what the colour looks like; the primer decides whether the client ever gets to see it.
The prep sequence that holds
The order of operations is where good intentions either pay off or get undone. On any substrate the spine is the same. Read the surface first and identify what it needs — adhesion, sealing, stain-blocking, or a combination — because that determines the primer before you open a tin. Make it sound: scrape failed paint back to a firm edge, drive out damp before you paint rather than priming over it, and fix the moisture source ahead of any coat, because no primer holds against active damp. Clean it — degrease the kitchen wall, wipe the plaster dust, take off furniture polish and silicone that would otherwise crater the film into fish eyes. Key it — sand the gloss dull, sand the hardwood to a profile, scuff the sound old paint — so the primer has tooth. Then treat the specifics: knot the softwood, seal the MDF edges, mordant the galvanising, mist-coat the new plaster. Only then prime, with the chemistry the substrate demanded, in full even coverage. Sand lightly between coats where the finish warrants it, dust off, and topcoat — two thin coats beating one thick one every time, because thick films skin over and craze.
The discipline that separates the trade from the weekend is that none of those steps is negotiable on the substrate that needs it. The knot gets knotted whether or not it looks small. The melamine gets a bonding primer whether or not the client is in a hurry. The plaster dries before it is painted whether or not the schedule wants it dry. Every one of those steps is cheaper than the callback that skipping it produces, and the primer — chosen right, applied right, in the right order — is where the durability of the whole job is decided.
FAQ
Can I skip primer and use a paint-and-primer-in-one? On a sound, previously painted surface in good condition, a self-priming topcoat is often fine — that is exactly the easy case primers are least needed for. The trouble is that the surfaces where primer matters most are the ones these products handle worst: bare knotty softwood, oak, melamine, glazed tile, fresh galvanising, new plaster. None of those is reliably solved by a combined product, because each needs a specific chemistry — a shellac barrier, a bonding grip, a mist coat — that a general-purpose topcoat does not carry. Match the primer to the substrate when the substrate is difficult; reach for the all-in-one only on the routine repaint.
Why did the knots come through a finished, perfectly good paint job? Because the knots were not sealed with a barrier coat before priming. Resin and tannin in the knot migrate outward through the film when moisture or heat reaches the wood, and standard primers do not block them — the molecule diffuses straight through a porous film. The paint is sound; the stain has come up from underneath. The only proper fix is to lift the paint over each knot, seal with shellac, re-prime and repaint. Over-painting without a barrier hides it for a few weeks and then it returns.
Is Zinsser B-I-N a damp sealer? No, and this catches people out. B-I-N is a shellac-based stain block — it seals stains, nicotine, tannin, and odour exceptionally well, and it is the universal blocker for those. It is not a moisture barrier and will not hold back damp pressure. Where the real problem is water — rising damp, a leak, condensation — the source has to be found and fixed, and a true two-pack epoxy or bituminous moisture barrier used where one is needed. Priming over damp with anything, B-I-N included, just delays the failure.
Do I really need a special primer for melamine and laminate? Yes. Melamine and laminate are non-porous and manufacture-sealed, with no tooth for a standard primer to grip, and paint applied over an ordinary acrylic primer peels off in sheets within weeks. A bonding primer chemically keys to the smooth face and gives the topcoat something to hold. Sand lightly to help, but the bonding primer is doing the real work; without it the failure is not a maybe, it is a when.
What is the single most common reason paint peels off an interior surface? Damp behind the film is the first — water pressure literally pushing the paint off, which no primer can resist and which means the wall must be dry before it is painted. The second is a glossy surface that was not keyed, so the new coat never bonded and slides off. Both are prep-and-primer failures, not topcoat failures, which is the whole point: the layer that determines whether the job holds is the one underneath.
How long should new plaster dry before I prime it? Roughly four to six weeks in normal ambient conditions, longer in a cold or unheated property, and the only reliable check is visual — uniform pale pink or off-white across the whole wall with no dark patches means it has lost enough water to paint. Then the priming step is the mist coat: emulsion thinned to roughly 70 per cent paint, 30 per cent water, to soak in and form the key. Full-strength emulsion straight onto new plaster sits on the surface with no grip and can peel away in sheets.