Quick answer: Sleeping position and mattress firmness together move the pillow height your neck needs across a range of 1.5 in to 6.25 in — a spread of nearly five inches, or 12 cm. Mattress firmness alone shifts a side sleeper’s target by 1.25 in, and it moves in the direction most people guess wrong: a softer mattress needs a lower pillow, because the shoulder sinks in and closes part of the gap. For someone who sleeps on both their side and their back, the arithmetic is worse: in 7 of the 9 build-and-mattress combinations, no single pillow height sits within half an inch of what either position needs. On the evidence side, side lying is the only position with a measured protective association against waking neck, shoulder-blade and arm pain, and the popular claim that stomach sleeping causes neck pain is not supported by the largest survey on the question.
In brief: this page maps every combination of sleeping position and mattress firmness to the supported pillow loft it requires, to what happens mechanically when the pillow does not change, and to what the published literature does — and does not — actually show about pain in that position. The loft figures are arithmetic from the formula published on our pillow height calculator. Every claim about where pain presents carries a link to the study it comes from. Where no study exists, the cell says so. Free to cite and reuse with attribution.
Finding 1: for a combination sleeper, no single pillow height fits both positions
Side sleeping and back sleeping do not ask for the same pillow. On our published formula an average-framed adult on a medium mattress needs a supported loft of 5.0 in on their side and 3.625 in on their back — a gap of 1.375 in that opens and closes every time they roll over. A fixed pillow cannot be in two places at once, so the calculator splits the difference and returns 4.3125 in.
The problem is that splitting the difference does not land inside the tolerance. The calculator reports every result as a ±0.5 in band, on the reasonable view that being half an inch out is not the same as being two inches out. But the compromise height sits 0.6875 in from both requirements for that average sleeper — outside the band in both directions at once. Run the same arithmetic across all nine combinations of frame width and mattress firmness and 7 of the 9 exceed the tolerance. The two that do not are both on soft mattresses, where shoulder sink compresses the difference between the positions.
The gap also widens predictably: it is smallest for a narrow frame on a soft mattress (0.375 in) and largest for a broad frame on a firm one (0.9375 in). Broad-shouldered combination sleepers on firm mattresses are the group a fixed-loft pillow serves worst, and the arithmetic says so before any product enters the discussion.
Finding 2: your mattress moves the target further than the tolerance allows — downward as it softens
Supported loft is the gap between the ear and the mattress surface once the shoulder has sunk in. That makes mattress firmness a direct term in the equation rather than a side issue. Our formula subtracts 1.5 in of shoulder sink on a soft mattress, 0.75 in on medium and 0.25 in on firm, halved for back sleeping because less shoulder is involved.
The consequence is arithmetic. Moving a side sleeper from a firm mattress to a soft one lowers their required pillow height by 1.25 in. For a back sleeper it is 0.625 in, and for a combination sleeper 0.9375 in. Only the side-sleeping figure is more than twice the ±0.5 in tolerance — which means a side sleeper who buys a new mattress and keeps the pillow that was working has moved fully outside their own correct band without changing anything about the pillow.
The direction is the part that catches people out. A softer mattress needs a lower pillow, not a higher one, because the shoulder drops into the surface and closes part of the gap the pillow was filling. This is not a novel observation — it appears in pillow-fitting guidance elsewhere — but it is rarely quantified, which is what the numbers below are for.
Stomach sleeping is the exception: it is treated as a fixed 1–2 in regardless of mattress, because no loft figure makes the position neutral for the neck. Mattress firmness does not move that target because height is not the operative variable there.
Finding 3: the published evidence lines up with position much less tidily than pillow marketing does
The geometry above is arithmetic and behaves. The clinical picture does not, and it is worth being blunt about where it disagrees with the received wisdom.
The largest survey we found on the question — 812 completed responses, Gordon, Grimmer and Trott (2007) — reported that side lying significantly protected against waking cervical, scapular and arm pain and was associated with higher sleep-quality ratings. It identified no advantage in adopting the supine (back) position. And it found that the prone (stomach) position was not significantly related to waking symptom reports, though the authors note that small subject numbers limited that particular analysis. The near-universal internet claim that stomach sleeping is the position that wrecks your neck does not have that survey behind it. The mechanical case — that the head must rotate to one side for hours, which Cleveland Clinic describes as stretching the neck muscles and extending the neck backwards — is a geometric argument, not an epidemiological finding, and should be quoted as one.
Where video evidence exists, it points somewhere unexpected. Cary, Jacques and Briffa (2021) filmed 53 people over two nights with infrared cameras in their own bedrooms. People waking with cervical symptoms spent a median 185.1 minutes a night in what the authors call a provocative posture, against 83.8 minutes for controls — roughly 2.2 times as long. But the defining feature of a provocative side-lying posture in that study is the top thigh advanced forward of the bottom thigh, twisting the spine. That is a trunk-and-hip variable, not a pillow-height one. The same study found no equivalent difference for the lumbar-symptom group.
Side lying is also where shoulder trouble concentrates. In a polysomnography study of 71 people (Yatağanbaba and colleagues, Acta Medica), patients with subacromial impingement syndrome spent significantly more time in lateral decubitus than controls (p=0.003) and spent 76.9% of that time lying on the painful shoulder rather than the other one (p<0.001). Forty-seven per cent reported shoulder pain on waking. That study is cross-sectional, so it cannot say whether the sleeping side caused the shoulder or the shoulder chose the sleeping side.
On mattresses, the one high-quality randomised trial we could find measured the lower back, not the neck. Kovacs and colleagues (2003, Lancet) randomised 313 adults with chronic non-specific low-back pain to a medium-firm or a firm mattress and reported odds ratios at 90 days of 2.36 for pain in bed, 1.93 for pain on rising and 2.10 for disability, all favouring medium-firm. That is a real result, and it is not a result about pillows, necks or loft. Nothing in it tells you what height pillow to put on a medium-firm mattress; that part is still arithmetic.
The full dataset
| Sleeping position | Mattress | Supported loft (in) | Supported loft (cm) | Shift vs medium mattress | What happens if the pillow does not change | What the published evidence shows for this position | Source |
|---|---|---|---|---|---|---|---|
| Side | Soft | 3.75–4.75 in | 9.5–12.1 cm | −0.75 in | Shoulder sinks 1.5 in, closing part of the gap. An unchanged pillow now sits 0.75 in too high and tips the head up toward the opposite shoulder. | The only position with a measured protective association: side lying significantly protected against waking cervical, scapular and arm pain in a survey of 812 people (Gordon 2007). It is also where shoulder trouble concentrates — patients with subacromial impingement spent 76.9% of lateral-decubitus time on the painful shoulder, n=71 (Yatağanbaba, Acta Medica). Provocative side lying — top thigh advanced forward, twisting the spine — ran 185.1 vs 83.8 min/night in people with cervical symptoms, n=53 (Cary 2021). | Gordon 2007; Acta Medica; Cary 2021 |
| Side | Medium | 4.5–5.5 in | 11.4–14.0 cm | reference | Reference case. Shoulder sinks 0.75 in; the pillow fills the remaining 5.0 in between ear and mattress. | ||
| Side | Firm | 5–6 in | 12.7–15.2 cm | +0.5 in | Shoulder barely sinks, so the pillow must fill almost the whole gap. An unchanged pillow sits 0.5 in too low and the head drops toward the mattress. | ||
| Back | Soft | 2.75–3.75 in | 7.0–9.5 cm | −0.375 in | Only half the shoulder sink applies, so the target moves 0.375 in. Too much pillow here pushes the chin toward the chest. | No advantage identified over side lying for waking cervical, scapular or arm pain in the 812-person survey (Gordon 2007). We found no trial linking back sleeping to a specific waking pain site. | Gordon 2007 |
| Back | Medium | 3.25–4.25 in | 8.3–10.8 cm | reference | Reference case. Base ear-to-mattress gap 4.0 in less 0.375 in of half-applied shoulder sink. | ||
| Back | Firm | 3.5–4.5 in | 8.9–11.4 cm | +0.25 in | Target rises 0.25 in. Under-lofting flattens the neck curve rather than supporting it. | ||
| Side + back (combination) | Soft | 3.25–4.25 in | 8.3–10.8 cm | −0.5625 in | Compromise height sits 0.5 in from each position — the closest the two ever get, and still at the edge of tolerance. | Not published. We found no study that reports waking symptoms for people who change position during the night as a separate group, so we could not determine what pain, if any, is characteristic of combination sleeping. | not published |
| Side + back (combination) | Medium | 3.75–4.75 in | 9.5–12.1 cm | reference | Compromise height sits 0.6875 in from both the side and the back requirement: outside the ±0.5 in band in both directions at once. | ||
| Side + back (combination) | Firm | 4.25–5.25 in | 10.8–13.3 cm | +0.375 in | Worst case for a fixed loft: 0.8125 in from each position. An adjustable or dual-zone pillow is the only geometric answer. | ||
| Stomach (prone) | Any | 1–2 in | 2.5–5.1 cm | no change | Height is not the operative variable. The head must rotate to one side to breathe, and adding loft increases the rotation and the backward extension rather than reducing it. | Contrary to the common claim, the prone position was not significantly related to waking symptom reports in the 812-person survey, though the authors note small subject numbers limited that analysis (Gordon 2007). The mechanical argument — sustained cervical rotation and extension — is described by Cleveland Clinic but is geometry, not an epidemiological finding. | Gordon 2007; Cleveland Clinic |
| Frame width | Side | Back | Combination compromise | Distance from each position |
|---|---|---|---|---|
| Narrow | 4.25 in | 3.125 in | 3.6875 in | 0.5625 in |
| Average | 5 in | 3.625 in | 4.3125 in | 0.6875 in |
| Broad | 5.75 in | 4.125 in | 4.9375 in | 0.8125 in |
Method and limitations
- The loft numbers are arithmetic, not measurement. Every height on this page is computed from the formula published on our pillow height calculator: supported loft = ear-to-mattress gap − shoulder sink. Base gaps are 5.0″ narrow / 5.75″ average / 6.5″ broad for side sleeping and 3.5″ / 4.0″ / 4.5″ for back sleeping; shoulder sink is 1.5″ soft, 0.75″ medium, 0.25″ firm, halved for back sleeping and taken at three-quarters for combination. Stomach sleeping is a fixed 1–2″. Anyone can reproduce every figure here from those inputs.
- It is a model, not a clinical measurement. Those base figures are a practical model built from pillow-loft conventions and cervical-alignment principles. They are a starting range to test, not a prescription, and they carry a ±0.5″ tolerance.
- No first-hand testing. This site does not buy, test, weigh or sleep on pillows or mattresses. Nothing here is our own measurement of a product.
- The evidence column is position-level, not firmness-level. We found no study that reports waking pain broken down by mattress firmness, so the evidence text is identical down all three mattress rows for a position. Read the mattress rows as arithmetic only.
- Everything cited is observational. None of the position studies randomised people to a sleeping position, so none of them shows that a position causes the pain associated with it. The subacromial impingement study is explicitly cross-sectional — the shoulder may be choosing the sleeping side rather than the other way round.
- Combination sleeping has no evidence base we could find. Recorded as “not published” rather than inferred from the side and back rows.
- What this page does not show. It does not show that changing your position or pillow height will relieve pain you already have. It does not model how a pillow compresses under the weight of a head, so a loft printed on a box is not the supported loft in this table. It does not cover pregnancy, post-surgical positioning or sleeping in a recliner. And it makes no claim about mattress firmness and neck symptoms, because the one randomised mattress trial we cite measured the lower back.
- Geometry, not clinical advice. These are calculations about body dimensions. Persistent neck, shoulder or arm pain — and any numbness, tingling or pain radiating into an arm — is a matter for a clinician, not a pillow.
Sources
- Gordon SJ, Grimmer KA, Trott P (2007). Sleep Position, Age, Gender, Sleep Quality and Waking Cervico-Thoracic Symptoms. Internet Journal of Allied Health Sciences and Practice 5(1). 812 completed surveys. nsuworks.nova.edu/ijahsp/vol5/iss1/6/
- Cary D, Jacques A, Briffa K (2021). Examining relationships between sleep posture, waking spinal symptoms and quality of sleep: A cross sectional study. PLOS ONE 16(11): e0260582. 53 participants, infrared video over two nights. journals.plos.org
- Gordon SJ, Grimmer-Somers KA, Trott PH (2010). Pillow use: the behavior of cervical stiffness, headache and scapular/arm pain. Journal of Pain Research 3:137–145. Five pillow types tested against participants’ own pillows in side sleepers. dovepress.com
- Yatağanbaba A et al. The relationship between sleep posture and subacromial impingement syndrome. Acta Medica. 71 participants (34 with subacromial impingement, 37 controls), polysomnography. actamedica.org
- Kovacs FM et al. (2003). Effect of firmness of mattress on chronic non-specific low-back pain: randomised, double-blind, controlled, multicentre trial. Lancet 362:1599–1604. 313 adults. pubmed.ncbi.nlm.nih.gov/14630439/
- Cleveland Clinic. Sleeping on Your Stomach: Is It Bad for You? Mechanical description of prone sleeping, quoting Andrew Bang, DC. health.clevelandclinic.org
- Sleep Align. Pillow Height Calculator — the loft formula and its published base values, from which every arithmetic figure on this page is derived. sleepalign.org/pillow-height-calculator/
Suggested citation: Sleep Align (2026). Sleep Position and Pain Map: pillow loft by position and mattress firmness, and what the evidence supports. Retrieved from https://sleepalign.org/sleep-position-pain-map/
Corrections are welcome at hello@sleepalign.org. If we have mis-stated a study or missed one that stratifies waking symptoms by mattress firmness, send the link and the page gets corrected.
Machine-readable. The loft arithmetic behind Table 1 is also available as a free open API: /wp-json/sleepalign/v1/loft returns the supported height range for any position, build and mattress firmness, with no key and no rate limit. See the Sleep Reference Library for the rest.
What to do with this
Three practical readings come out of the table. First, work out the number for the position you actually spend the night in, not the one you fall asleep in — the map moves by more than an inch between them. Second, if you have changed mattress and your pillow stopped working, the pillow did not change; the shoulder sink did, and the correction is downward on a softer bed. Third, if you genuinely alternate between side and back, the arithmetic says a fixed loft cannot serve you: an adjustable-fill pillow, or a contoured one with a higher side section and a lower centre, is the only geometry that covers both.
And the honest caveat, which is the reason this page exists in the form it does: the geometry is tidy and the evidence is not. Side lying is the position with the best measured record for waking neck symptoms, stomach sleeping’s bad reputation is not well supported by the largest survey on it, and the one thing filmed studies point at most strongly — a twisted trunk in side lying — is not something a pillow fixes at all.
Work out your own number. The pillow height calculator takes your position, frame width and mattress firmness and returns the supported range used throughout this page — then the pillow specification database shows which of 32 catalogued pillows publish a loft inside it. Only 16 of the 32 publish a height at all.
If your result lands in the contoured memory-foam band, the Derila Ergo is the contoured pillow this site links to; it carries a 60-day money-back guarantee on the official store. That is an affiliate link and nothing else on this page is.
Questions this page answers
Does a softer mattress need a higher or a lower pillow?
Lower. On a soft mattress the shoulder sinks about 1.5 in into the surface, closing part of the gap the pillow would otherwise fill. An average-framed side sleeper needs 4.25 in of supported loft on a soft mattress, 5.0 in on medium and 5.5 in on firm. The full soft-to-firm swing is 1.25 in for side sleeping, 0.625 in for back sleeping and 0.9375 in for combination sleeping.
Can one pillow work if I sleep on both my side and my back?
Usually not. The compromise height for an average frame on a medium mattress is 4.3125 in, which is 0.6875 in away from both the side requirement (5.0 in) and the back requirement (3.625 in) — outside the ±0.5 in tolerance in both directions at once. Seven of the nine frame-and-mattress combinations behave the same way.
Is stomach sleeping actually bad for your neck?
The mechanical case is clear; the evidence is weaker than commonly claimed. The prone position was not significantly related to waking symptom reports in the 812-person survey by Gordon, Grimmer and Trott, though small subject numbers limited that analysis. The argument that the head must rotate to one side and the neck extends backwards is geometry, and should be quoted as geometry.
Which sleeping position has the best evidence for fewer waking neck symptoms?
Side lying, which significantly protected against waking cervical, scapular and arm pain in that same survey while the supine position showed no advantage. It is also where shoulder trouble concentrates, so “best evidenced for the neck” is not the same as “best for everything”.
How much does mattress firmness matter for neck pain compared with pillow height?
We could not determine this, and we are not going to guess. No study we found reports waking neck symptoms broken down by mattress firmness. The strongest randomised mattress evidence — Kovacs and colleagues, Lancet 2003 — measured the lower back, not the neck.
Sleep Align is an independent, self-funded pillow research site run by Bogdan Bilt. He is not a clinician and the site carries no medical credentials; it publishes geometry, published specifications and links to other people’s research. This page is part of the Sleep Reference Library. This is geometry, not clinical advice — persistent pain is a matter for a clinician.