How to Adjust an Office Chair: The Order That Actually Works

Almost every guide to adjusting an office chair begins by telling you to sit in it. That is the wrong end of the problem, and the reason is that three things in a desk setup are fixed and one is not. Your body is fixed. Your desk, unless you bought a sit-stand frame, is fixed. The monitor can move but only after everything under it has stopped moving. The chair is the one adjustable object in the room, which means the chair's job is not to be comfortable in isolation — it is to close the distance between a body it cannot change and a desk it cannot change.

So the first number to write down is not anything about the chair. It is the height of your desk, measured from the floor to the top of the desktop. Most fixed desks land between 29 and 30 inches. If yours moves, this whole article gets easier and you should read our standing desk height guide instead, because an adjustable desk lets you set the chair to your body and then bring the desk to meet it, which is the correct order and the one almost nobody gets to use.

Here is why the 29-inch convention matters more than it looks. ANSI/BIFMA's G1 ergonomics guideline is built on a straightforward measurement: popliteal height plus elbow height, which is the distance from the floor to your elbow when you are seated with your feet flat and your forearms level. That number is the height a work surface should be. BIFMA's revision committee published it against two anthropometric datasets, the 1988 Natick survey of military personnel and the 2012 CAESAR civilian survey, and the two agree closely: a 5th-percentile woman needs about 22.4 inches, and a 95th-percentile man needs about 29.2.

Now put a standard desk next to those figures. BIFMA's own recommendation for a fixed-height input-device surface is no less than 747 mm, which is 29.4 inches — essentially the 95th-percentile man's number. For an adjustable surface it recommends a minimum range running from 575 mm (22.6 in) up to that same 747 mm, which covers everybody. The industry standardised on the top of the range and called it a desk. That is not a scandal, it is just what happens when a surface has to be one height for everyone and the cheap failure mode is being too short for the tall guy. But it means that for most adults a fixed desk is not slightly high. It is several inches high, and the entire correction is charged to the chair.

One more thing worth noticing in BIFMA's data, because it explains a complaint people have about chairs and cannot articulate. Between Natick 1988 and CAESAR 2012, seated hip breadth at the 95th percentile for women went from 17 inches to 20.4 — a 3.4-inch jump, mostly because the population changed from soldiers to civilians. BIFMA's minimum seat width is 457 mm (18.0 in), and the next line of the same guideline says that at 461 mm (18.1 in) a seat "will accommodate at least the 80th percentile female hip breadth" — so the floor of the standard is written to leave a fifth of women out. Cornell's ergonomics notes ask for 20 to 22 inches to fit clothed adults. A well-regarded chair like the Steelcase Series 1 has a 19.5-inch seat. If a chair has ever felt narrow to you in a way you could not pin on the armrests, that gap is where the feeling comes from, and no adjustment fixes it.

This is the step everybody does first and most people do backwards. The instinct is to lower the chair until your feet are flat on the floor, because that is what the diagrams show and because it feels like the natural starting point. OSHA's computer workstations eTool even says it plainly: "The chair height is appropriate when the entire sole of the foot can rest on the floor with the back of the knee slightly higher than the seat of the chair."

That instruction is correct and it is not the whole instruction, because a few lines later the same eTool adds the caveat that makes it work: "If the seat cannot be lowered (for example, it would make the keyboard or monitor too high), use a footrest to provide stable support for the feet." Read those two sentences together and the real rule falls out. Feet-on-the-floor is the goal. Elbows-at-the-desk is the constraint. When the two disagree, the constraint wins and the floor gets moved up to meet you.

So: sit in the chair, pull it to the desk, and set the height so that with your upper arms hanging relaxed and your elbows bent to about a right angle, your forearms are level with the desktop — or a fraction above it, because your hands actually rest on a keyboard rather than the desk itself. Shoulders down, wrists straight. If your shoulders are creeping up toward your ears, the chair is too low. If you are reaching down and your wrists bend upward to find the keys, it is too high. Ignore your feet entirely for now. They are the next section's problem.

Here is where the arithmetic from the previous section arrives, and it produces a conclusion that is the exact opposite of the usual advice. Conventional wisdom says short people need chairs that go low. On a fixed 29-inch desk, short people need chairs that go high — because the desk is nowhere near coming down to them, so the only way to get their elbows up to it is to raise the seat well above the height their legs would choose. The gap between BIFMA's 5th-percentile figure of 22.4 inches and a 29-inch desk is 6.6 inches, and every one of those inches has to come out of the chair's height range.

Which is a problem, because chairs do not have 6.6 inches of range. We went through all eighteen chairs across our three chair guides and pulled every published seat-height spec we could find — which turned out to be the first finding on its own, because only seven of the eighteen publish a range at all. Among those seven, the total travel runs from 2.9 inches to 5.0. Cornell's ergonomics course notes recommend that a seat "should adjust 9 inches" and that the minimum height should be 15 inches, "which designs to the 5th percentile of women with 1" heels." Not one chair here has half of that range, and the ones that reach highest do it by starting high rather than by travelling further. The seat-height adjustment on a modern task chair is roughly a third to a half of what the ergonomics literature asks for, and the shortfall is invisible because nobody publishes travel — they publish two endpoints, and two endpoints look generous when they are 16.5 and 21.5.

The practical consequence is a shopping rule, not just an adjustment rule. If you are short and your desk is fixed at 29 inches, the spec that decides whether a chair works for you is the top of its seat-height range, and you should read it as pass/fail. Our under-$600 guide notes that the Branch Ergonomic Chair Pro's optional tall cylinder is required if your desk does not drop below 29 inches — which is right, and worth restating in the direction people do not expect. That cylinder is not a tall-person upgrade. It is a fixed-desk upgrade, and the people it rescues are disproportionately the ones the standard cylinder was supposedly sized for.

Once the seat is set to the desk, look down. If your feet are flat on the floor with your thighs roughly level, you are finished with this step and you can skip to the next one — this is the position a taller adult on a standard desk usually lands in naturally, which is precisely why the problem below goes unreported by so many reviewers.

If your feet are dangling, or you find yourself hooking them on the chair base or tucking them under you, the floor is in the wrong place and the fix is to raise the floor. That is all a footrest is. It is not a comfort accessory bolted onto a correct setup; it is the second half of a seat-height adjustment that a fixed desk forced you to make. The pressure it relieves is real and specific: with the seat above your popliteal height, your body weight bears down on the underside of your thighs at the front edge of the seat, which is where the sciatic and peroneal nerves and the vessels behind the knee run. Cornell's notes describe exactly this, a seat that is too high "leads to increased pressure at the popliteal fold, decreasing blood circulation and increasing pressure on the nerve." Dangling feet are not a posture preference. They are a load path.

Now measure the gap you actually need to close, because this is where our own guides have been quietly optimistic. The distance from the floor to the soles of your feet, once the seat is set to the desk, is the platform height a footrest has to deliver. For someone at the 5th-percentile end of BIFMA's data on a 29-inch desk, that number is in the neighbourhood of six and a half inches. We went back through the footrests in our own footrest guide and totalled up what they actually reach — and only one of the five gets there, at the top of three fixed steps, with about a tenth of an inch to spare.

That is an uncomfortable finding to publish in a guide whose job is partly to sell footrests, so let us be plain about what it means rather than softening it. For most people the footrest category is sized correctly, because most people are not at the 5th percentile and the gap they need to close is three to five inches, which every product here covers comfortably. Worth knowing before you order, though: the "adjustable" ones adjust in steps rather than continuously — both the HUANUO and the Fellowes give you three fixed positions — so you are picking the nearest step to your measured gap rather than dialling the exact number, and it is the top step that decides whether a product is in play at all. But if you are at the short end of the distribution and your desk does not move, the standard product category runs out before your gap does, and the honest options are a taller platform, a stack (a footrest on a low box, or a purpose-built adjustable model at its top setting), or — the better answer if you can afford it — a desk that goes down instead of a floor that goes up. A sit-stand frame with a low minimum solves this properly, and it is one of the few cases where we would tell someone to spend $400 rather than $30.

Two smaller notes while you are down there. A footrest with some tilt or rocking movement is worth having over a fixed block, because the thing your legs want over eight hours is not one correct angle but a supply of them — which is the same reasoning behind our preference for chairs with a working tilt. And if you are on carpet, remember that your chair's casters sink into it while a footrest's flat base mostly does not, so the effective gap on carpet is slightly smaller than on a hard floor. A chair mat removes that variable entirely, which is a small ergonomic argument for one on top of the flooring-protection argument that usually sells them.

Seat depth is the adjustment most people never touch, partly because a lot of chairs do not have it and partly because it is not obvious what it does. The test is simple. Sit all the way back so your lower back is against the backrest, then check the gap between the front edge of the seat and the back of your knees. You want roughly two to three fingers' width. OSHA phrases the target as a seat pan that "should provide support for most of the thigh without contact between the back of the user's knee and the front edge of the seat pan," and the failure mode in each direction is distinct: too deep and the front edge presses into the backs of your knees so you slide forward and stop using the backrest at all; too shallow and your thighs are unsupported and all your weight lands on your sit bones.

Cornell's notes put the recommended adjustable range at 14 to 18.5 inches, which is 4.5 inches of travel, and flag why the deep end matters: "if the seat depth is greater than the buttock-popliteal length (fifth percentile woman is at 17 in) then the user won't be able to use the backrest." Real chairs give you less than that. The Steelcase Series 1 — one of the better-adjusting chairs at any price and the top pick in our under-$600 guide — publishes a functional seat depth of 15.5 to 17.75 inches, a 2.25-inch range. That is half of Cornell's recommendation, and it is the most generous figure in our guides rather than a bad one. Hold on to the 2.25; it comes back in the next section and it does not survive.

The backrest is two adjustments that are easy to confuse. Back height or lumbar height positions the support at your lower back — the inward curve at your belt line, not the middle of your back, which is the single most common error people make with an adjustable chair and with every cushion sold to fix one. The Series 1 puts its lumbar 6.5 to 8.75 inches above the seat, which is a 2.25-inch window and about right for most adults; if yours adjusts, run it up and down while sitting and stop where the pressure feels centred on the curve rather than above or below it. Back angle or recline is the other one, and OSHA asks only that a chair let you "recline at least 15 degrees from the vertical," which every chair here clears many times over.

Which brings up the number the entire gaming-chair category is sold on and the one you should mostly ignore. Recline figures — 165 degrees, 160, 152, 135 — are marketing's favourite spec because they are large and trivially comparable, and nobody works at any of them. Our gaming chair guide makes this point and it is worth repeating in a piece about adjustment: what you use every day is a tilt lock that holds a slight backward angle without drifting, and tilt tension set heavy enough that leaning back does not feel like falling. Set the tension so that you can rock with light pressure but the chair stays where you put it, lock it a few degrees back from vertical, and leave it. Sitting bolt upright at 90 degrees is not the goal and never was; a slight recline transfers load off your discs onto the backrest, which is what the backrest is for.

Here is the section that made this guide worth writing, and it comes out of holding our own chair guides side by side rather than out of any product's spec sheet. Between them, the seven guides in this hub recommend a chair, a seat cushion, a lumbar cushion, a footrest and a chair mat. Each one is sound advice on its own. Followed together, three of them silently change the same measurements the chair adjustments were set to — and they are spent from a budget that, as the last two sections established, is far smaller than the ergonomics literature asks for.

A seat cushion is a seat-height adjustment. The cushions in our seat cushion guide run 2 to 4 inches thick, and the large Cushion Lab is 5. Put a 3-inch cushion on a chair and you have raised your sitting surface 3 inches, which on the Steelcase Series 1 spends 60% of the chair's entire 5-inch height range in one move. That part is recoverable — you wind the chair back down. What is not recoverable is the armrests, because they are bolted to the chair and travel with the seat pan, not with the top of your cushion.

Follow that through with Steelcase's own published numbers, because the result is sharper than we expected. The Series 1 specifies arm height from the seat at 6.5 to 11 inches. Cornell's notes say cantilevered elbow rests "should be 8-10 in above the seat surface height," and the chair covers that window comfortably. Now add a 3-inch cushion: measured from where you are actually sitting, the arms now reach 3.5 to 8 inches, so the chair's maximum armrest height is the bottom of the recommended window. Add the 4-inch cushion instead and the top of the range is 7 inches — below the recommendation entirely, with no setting on the chair that fixes it. A cushion that solves a genuine tailbone problem can, in the same motion, take your elbows out of support for the rest of the day.

A lumbar cushion is a seat-depth adjustment. This one is worse arithmetic. The cushions in our lumbar support guide are 2.5 to 4 inches deep — the Purple Back Cushion is 2.5, the LoveHome and the Cushion Lab are both 4. That depth sits between your back and the backrest, so it pushes you forward by its own thickness, and every inch it pushes you forward is an inch subtracted from your effective seat depth. A 4-inch lumbar cushion therefore removes almost twice the Series 1's entire 2.25-inch seat-depth range. There is no combination of chair settings that gets it back. If the chair was correctly set before, it is now several inches too shallow, and you will feel it as your thighs losing support and your weight concentrating on your sit bones — which, unhelpfully, is the exact symptom that sells seat cushions.

None of this is an argument against cushions. A lumbar cushion on a chair with a flat unsupportive back is a genuine upgrade, and a tailbone cutout is the right answer to a specific and painful problem. It is an argument about order, and it is the one thing to take away from this article if you take nothing else: fit the cushions first, then adjust the chair around them. Adjusting a chair carefully and then adding a cushion undoes the adjustment, silently, in a direction you will not think to look. If you are using a lumbar cushion, start from the chair's deepest seat setting rather than its fitted one. If you are using a seat cushion, set the armrests last and expect to run them near the top of their travel.

The chair mat is the exception that proves the pattern, and it is worth stating so the table below is not read as a warning about all four. The mats in our chair mat guide are 0.09 to 0.2 inches thick, which is inside the measurement error of everything else here. What a mat actually changes is consistency — on carpet, casters sink by a variable amount depending on pile and padding, so your seat height is not the number you set, and it drifts as you move around. A mat makes the floor one height. That is a small effect and a real one.

Armrests come near the end of the sequence for a reason: their correct position is defined relative to your elbows, and your elbows do not stop moving until the seat height, the footrest and the seat depth are all settled. Adjust them earlier and you will just do it again.

OSHA's phrasing is the best short test there is, because it is two conditions rather than one. Armrests should be "low enough so your shoulders are relaxed during use" and "high enough to provide support for your lower arms when positioned comfortably at your sides." Set them so that when your forearms rest on them your shoulders neither rise nor drop — if you can see your shoulders lift in a mirror, they are high; if your arms feel like they are reaching down, they are low. Then, if the chair allows it, bring them in. OSHA's third condition is that they be "close enough to provide support for your lower arms while keeping your upper arms close to the body," and width is the axis most people never touch even when they have it.

This is where armrest axes stop being a spec-sheet curiosity. Height-only arms give you one degree of freedom and you take what the chair's width gives you; 3D and 4D arms let you bring the pads under your forearms where they actually rest. Our under-$600 guide singles this out as a bigger real-world difference than any recline figure, and having now done the arithmetic on cushions, we would put it more strongly: armrest range is the adjustment with the least slack in it, so it is the one that a cushion, a thick seat pan or an unusual torso most easily pushes out of reach. Cornell's numbers put a second constraint on the width axis too — armrests "should be at least 17.2 in apart to exceed thigh breadth of 95th percentile females" — and chairs with narrow-set arms can press against you before they support you.

Two failure modes worth naming. First, armrests that stop you pulling the chair under the desk are worse than no armrests, because the compensation is leaning forward, which unloads the backrest you spent the last section setting. If the arms collide with the desk edge, drop them below the desktop and use them only when you push back. Second, armrests are not wrist rests. They support your forearms between keystrokes; if you are resting on them while typing your wrists will deviate sideways to reach the keys. That is what a wrist rest is for, and the two do different jobs.

Finish with tilt, which takes ten seconds and gets skipped almost universally. Find the tension knob, usually a large dial under the seat, and turn it until leaning back takes light effort rather than either no effort or a shove. Then set the tilt lock a few degrees back from vertical and leave it there. If your chair has a forward-tilt or seat-angle function, it is worth a try on a fixed desk that is too tall — tipping the seat pan slightly forward opens the hip angle and can buy back a little of the reach the desk is costing you.

Ten minutes, once, and a tape measure for the first two steps. The order matters more than the precision of any individual step, because each one changes the target for the ones after it.

1. Measure the desk. Floor to desktop. If it is between 29 and 30 inches, it is a standard fixed desk and the rest of this list is about working around that. If the desk moves, set the chair to your body first and then bring the desk to your seated elbow height, and read our standing desk height guide for the full method.

2. Fit any cushions now, before touching a single lever. A seat or lumbar cushion changes the datum every following step measures from. Adding one afterwards undoes the work.

3. Set seat height from your elbows. Sit at the desk, upper arms relaxed, elbows at roughly a right angle, forearms level with the desktop or a fraction above. Shoulders down. Ignore your feet completely at this stage.

4. Fix the floor. Feet flat and thighs level, or a footrest tall enough to make them so. Measure the gap before buying — if it is over about five inches, check the product's top step rather than assuming, because these adjust in fixed positions rather than continuously.

5. Set seat depth. Sit right back against the backrest; you want two to three fingers between the seat's front edge and the back of your knees. If you are using a lumbar cushion, start from the deepest setting, because the cushion has already taken several inches off.

6. Set the lumbar height. At your belt line, in the inward curve of your lower back, not in the middle of your back. Run it up and down while seated and stop where the pressure centres on the curve.

7. Set the armrests. Shoulders neither lifted nor dropped, forearms supported, arms brought in under your elbows if the chair allows width adjustment. Check they still clear the desk edge.

8. Set tilt tension and lock. Leaning back should take light effort. Lock a few degrees back from vertical rather than bolt upright.

9. Only now, the monitor. Everything above changed your seated eye height, so a screen set before this point is set to the wrong body position. Top of the screen at or just below eye level, roughly arm's length away — the method and the desk-depth arithmetic are in what size monitor for your desk.

10. Re-check in a week. Chairs settle, foam compresses and your first guess at seat depth is usually a little deep. Ten seconds of adjustment after a week of real use is worth more than another ten minutes today.