Knob or pull, what size, and exactly where the screws go — because the handle is the only part of a kitchen you touch every day, and the only decision you cannot take back with a tin of paint.
Everything else in a kitchen can be undone. Repaint the walls, swap the tap, live with a colour for a year and change your mind. A handle hole is not like that — a permanent hole bored through a finished door, and the only undo is a larger backplate. Which is why the part of this job that matters is not choosing the handle. It is the twenty minutes with a scrap of hardboard before the first hole.
A pull gives you a bar — hook it with two fingers, a fist, a knuckle, or the back of a wrist when both hands are full. A knob asks for a pinch grip, the first grip to fail with arthritis or a healing wrist. If anyone in the house has weak grip, the answer is pulls.
The cost structure is simple: a knob is one hole, a pull is two. Twice the chance of drilling one wrong, and an alignment problem a knob does not have — nobody notices a knob a degree off level; everybody notices a pull.
Knobs also win in tight rooms. A knob projects roughly 1″ from the door face, a chunky bar pull nearer 1½″, with an opening that catches pockets and tea towels — a row of snags at hip height in any galley under 42″.
Drawers end the argument. Put a knob in the middle of a 30″ drawer and you drag it from one point — the front flexes, the drawer racks in its runners, and that is how a soft-close runner starts to sound gravelly. Hence the professional compromise: knobs on doors, pulls on drawers.
Two numbers describe every pull and people mix them up. Centre-to-centre — centres, or CTC — is the distance between the two screw holes; overall length is tip to tip. A 128mm pull might be 150mm long, or 190mm if the legs splay. Order by centres: centres are what you drill.
For size, the rule is thirds: overall length about a third of the front’s width. Not a law, but it heads off both failures: too short and the pull looks stranded, too long and it crowds the reveal.
| Front | Centres to look at | Why that range |
|---|---|---|
| Cabinet door, any width | 3–4″, or a 1¼–1½″ knob | A door opens with a light tug; length crowds the reveal. |
| Drawer under 15″ | 3–4″ (76–102mm) | Longer leaves almost no front either side. |
| Drawer 18–24″ | 4–6¼″ (102–160mm) | The comfortable middle; 128mm is usual. |
| Drawer 24–30″ | 5–7½″ (128–192mm) | Enough grip to pull a loaded drawer straight. |
| Drawer 30–36″ | 10–12″, or two pulls | Past 30″ a single centred pull twists the front. |
| Drawer over 36″ | Two pulls, at the quarter points | Two hands, two points of pull, no racking. |
| Panelled fridge or dishwasher | 12–18″ appliance pull | A cabinet pull here is a lever built for the wrong load. |
That last row is mechanical, not stylistic. A panelled fridge door is heavy and opened with a sharp yank, and the pull is all that stands between your hand and that mass. A standard 160mm pull looks undersized there, and it loosens.
| Centres | The other way round | Where you meet it |
|---|---|---|
| 3″ | 76.2mm | The imperial default. Most door pulls and cup handles. |
| 96mm | a shade over 3¾″ | The European equivalent, a full 20mm off 3″. |
| 4″ | 101.6mm | Doors and narrow drawers where 3″ looks mean. |
| 5″ | 127mm | One millimetre from 128mm, which is a trap. |
| 128mm | just over 5″ | The European workhorse, and the most-sold spacing. |
| 160mm | 6¼″ | Standard on a 24–30″ drawer front. |
| 192mm | just over 7½″ | Wide drawers and pan banks; beyond it, 224mm up. |
About that millimetre: some pulls have enough slop in their bolt holes to bridge 5″ to 128mm and some do not, and forcing it leaves a bent screw and a stress point in the door face. Measure the pull in your hand.
The rule with no exceptions Buy the hardware first, then drill. The pull defines the holes. Hoping to find one that fits what you already bored is not a plan.
The hinge is the pivot, and the further from it you apply force the less you need. So the fitting belongs on the corner furthest from the hinge: the bottom outer corner on a wall cabinet, the top outer corner on a base. Get it backwards and the kitchen feels subtly wrong for years.
How far in? On a plain slab door, 2″ to 3″ from both edges. Under 2″ the fitting crowds the corner and approaches its neighbour across the reveal. Past 3″ it drifts toward the middle and looks like a measuring error.
On a Shaker door the frame overrules that number: centre the fitting on the stile — the vertical frame member at the door’s edge — and align it to the rail, either centred on the rail or on its inner edge. With a stile a little over 2″ wide that lands about an inch and a quarter from the edge — closer in than the slab figure, and correct, because on a framed door the eye reads the frame. Pick one convention and keep it: mixing slab drawers under Shaker doors puts the two at different distances from the edge, and it shows.
Frameless cabinets sharpen this. Full-overlay doors leave only about an eighth of an inch between fronts, with no face frame to absorb a fitting set too close to the edge. On a pair of doors meeting in the middle that is fine; at an inside corner it is not.
The planner lets you set where the handle sits on doors and on drawers, then look at it in 3D with the doors open, before anything is drilled. A flat plan draws fronts, not fittings, so a corner collision is not on the page at all. Check tall pantry doors too: a handle centred on a full-height door puts your hand at chest level, so line it up with the base run beside it.
Left to right, a single pull is centred. Always. The eye catches two millimetres of error on a drawer front the way it never does on a door, because the two vertical edges give it a reference.
Vertically it depends on height. Up to about 6″ of front — a cutlery drawer, the top of a three-drawer base — centre it. On a tall front, a fitting at true centre reads as low, because the optical centre of a rectangle sits above the measured one. On a 10–12″ pan drawer, set the pull about a third down from the top.
Both schools are legitimate. One centres every pull on every front, which never looks like a mistake — the three pulls on a base then sit at uneven spacings, which is fine, because the fronts are uneven too. The other raises the deep ones. Choose one per kitchen.
Past roughly 30″ of width, go to two pulls at the quarter points — on a 36″ front, 9″ in from each end. Two hands, two points of pull, nothing twists. One very long bar is the other honest answer. A modest pull marooned in the middle is not.
Vertical pulls are a tool, not a compromise. Use them where a horizontal pull would collide — beside an inside corner, or an appliance handle — and repeat them: one alone looks like a repair, a matched pair flanking the range looks like a decision. Standard drawer widths are here.
What they search at midnight “where to put cabinet handles” — usually with the doors already hanging and the drill on charge. Put the drill down. The template comes first.
The trade truth that makes this easy to accept: an installer who has fitted two hundred kitchens will still put a handle in the wrong place working from a tape and a pencil. Not because they cannot measure — because a kitchen is forty fronts, and the fortieth comes at the end of a long day.
One detail ambushes people on the day: screw length. The screws in the bag suit a door, around three quarters of an inch. A drawer front on a separate sub-front is double that, and they spin uselessly a thread short.
If a hole ends up wrong the rescues are limited. A backplate covers it, but then every front in the run needs one. A longer pull can sometimes swallow the old holes. On a painted door you can fill and touch up, though the sheen rarely matches; on foil or laminate, filling is not realistic. More on repairs here.
You are choosing the surface that gets touched more than any other object in the house — thousands of contacts a year, in the same spot, with skin oil. That is a harder brief than any other finish in the room, and it is why hardware fails in predictable places.
Plated and lacquered finishes wear at the touch points first — the underside of a bar where thumbs ride, the crown of a knob. Once the coating is through the patch goes cloudy and cannot be spot-repaired; the fix is a new handle, so buy a spare or two at the start. Unlacquered brass and oil-rubbed bronze are living finishes: they change unevenly, fastest where hands land. That is the point. Matte black is two things — a real powder coat is tough and has a faint texture you can feel; cheap matte black is thin paint over a zinc casting and wears silver along the edges within a couple of years.
On fingerprints the trade-off is boring. Polished anything shows every print and wipes clean in one pass. Matte black and satin brass hide prints but hold a greasy sheen. Brushed nickel is where most people land: it hides prints and hard-water spotting and has never dated. Near the sink, matte and satin finishes show white mineral spots — limescale, not a defect. Handle a sample with a greasy hand first.
Mixing metals needs contrast and a hierarchy. Black hardware with an unlacquered brass tap works, because nobody could confuse the two. Brushed nickel beside a stainless tap and a chrome pot filler does not, because three near-identical greys read as three failed attempts at the same colour. Two metals in one sightline, obviously different — and stainless appliances are already a metal in the room. The tap need not match the handles, but the tap, strainer and soap dispenser live in one square metre and must match each other. Door colour interacts with this.
Two routes, and they are not the same decision. One is a mechanism that can be added to almost any kitchen. The other is a shape built into the door or carcase, designed in from the first drawing.
Best on a short run of upper doors, or an island back you rarely touch
A sprung catch in the hinge, the runner, or behind the front. Push, it pops; push, it closes. No fitting at all, which on a plain slab door in a strong colour is striking.
The costs are real. You now push on the door face, so the finish collects an oil map where the hand lands — brutal on matte black and high gloss. A mechanical latch needs travel, so the front sits fractionally proud when closed unless the system is powered. And they open when nudged: a hip, a dog, a drawer bouncing off its own soft-close. Beside a dishwasher, a door that keeps drifting open stops being charming within a fortnight.
Best when the whole kitchen is designed around it from the first sketch
A J-pull door has a return machined into its top edge, so you hook your fingers underneath. A continuous channel between drawer banks does the same along a run. Both give a handle-free face with something real to grip — no mechanism, nothing to service, and it works with a wet hand.
The cost is a system-level commitment: the profile has to be in the door programme or the carcase from day one, and you cannot decide you want it in a year. Channels collect crumbs. And a horizontal lip, while easier than a pinch grip, is still harder than a proper D-pull — if accessibility is the driver, a good pull beats every handle-free system. Edge pulls screwed to a drawer front’s top edge are the retrofittable cousin, but they project into the reveal.
Our own cabinets are frameless with full-overlay doors, the construction all of these systems assume: no face frame, a continuous plane of fronts, a small even gap. That is what makes an integrated profile look right — and a badly placed pull look wrong. Frameless versus framed is here.
The first two are catchable weeks earlier, on a screen: set the handle positions in the planner, open the fronts in 3D, and look at every inside corner and every cabinet touching an appliance. How to drive the planner is here.
Decide it by hand, not by looks. A pull can be gripped with a fist, a knuckle or a wet hand, so it wins for anyone with a weak grip, and it wins on drawers, because a heavy drawer dragged from one knob racks in its runners. A knob is one hole instead of two and snags less in a narrow aisle. The usual compromise: knobs on doors, pulls on drawers.
Aim for an overall length around one third of the drawer width: 3 to 4 inch centres on doors and drawers under 15 inches, 4 to 6 inch centres on an 18 to 24 inch drawer, 5 to 8 inch centres on a 24 to 30 inch drawer, and 10 to 12 inch centres or two pulls beyond 30 inches. Panel-ready appliances take their own pulls, typically 12 to 18 inch centres.
Centre-to-centre, often written CTC or just centres, is the distance between the two screw holes. Overall length is the tip-to-tip measurement. You order by centres, because that is what you drill, and judge the look by overall length, because that is what the eye sees.
On the corner furthest from the hinge: the bottom outer corner on wall cabinets, the top outer corner on bases. On a plain slab door, 2 to 3 inches in from both edges. On a Shaker door the frame decides: centre the fitting on the stile and align it to the rail, which lands a little over an inch from the edge.
Centred left to right, always. Vertically, centre the pull on a shallow front up to about 6 inches high. On a tall front such as a 10 to 12 inch pan drawer, set it a third of the way down from the top, because a fitting at true centre looks low. Past 30 inches of width, use two pulls at the quarter points.
The set you will meet is 3 inches, 96mm, 4 inches, 5 inches, 128mm, 160mm and 192mm, with 224mm and larger for wide drawers. Watch the near-misses: 3 inches is 76.2mm against a metric 96mm, and 5 inches is 127mm against a metric 128mm. Some pulls bridge that millimetre and plenty do not.
Yes, and so does everyone else. An installer who has fitted hundreds of kitchens will still misplace a handle working from a tape and a pencil, because a kitchen is forty fronts and the fortieth comes at the end of a long day. Make a jig from quarter-inch hardboard with a square corner and two drilled holes, and register it against the front's corner.
Only partly, so drill carefully the first time. A backplate will cover a hole, but then every front in that run needs one. A longer pull can sometimes swallow the old holes. On a painted door you can fill and touch up, though the sheen rarely matches; on foil or laminate, filling is not realistic.
They give a completely clean door face, and on a short run of upper doors they are lovely. The honest costs: you push on the finish, so fingerprints land on the door, not on metal. A latch needs travel, so fronts sit fractionally proud when closed. And they pop open when nudged by a hip, a pet or a drawer bouncing off its soft-close.
Set the fitting positions on your own doors and drawers, open everything in 3D, and find the corner collisions while they are still just pixels.
Open the 3D planner →