These are the massage stats we can point at a source for, with the source linked and the year of the data attached. Where the honest answer is “nobody has a good number for that,” we say so.
How many people actually get a massage
The best national figure comes from federal survey data rather than from the industry. According to the National Center for Complementary and Integrative Health:
- 10.9% of US adults used massage therapy in 2022.
- That is more than double the 4.8% who used it in 2002.
Two decades, a bit more than double. That is the real growth story, and it is a good deal less dramatic than the compound-annual-growth-rate charts you will find elsewhere.
The workforce, by the numbers
All of these come from the US Bureau of Labor Statistics Occupational Outlook Handbook, which is the closest thing this trade has to an authoritative headcount.
- 168,000 massage therapist jobs in the US in 2024.
- $57,950 median annual wage, May 2024 — about $27.86 an hour. The median across all US occupations that month was $49,500.
- The lowest-paid 10% earned under $33,280; the highest-paid 10% earned over $97,450. That is a wide spread, and it reflects how much of this work is part-time.
- 42% are self-employed — by far the largest single category. Personal care services account for 30%, offices of other health practitioners 10%, chiropractors’ offices 6%, and hotels and similar accommodation 6%.
- Therapists working in chiropractors’ offices had the highest median wage of the main settings, at $66,710.
- Employment is projected to grow 15% between 2024 and 2034, against 3% for all occupations — reaching 193,900 jobs, a rise of 25,900.
- About 24,700 openings a year are projected over that decade, most of them replacing people who leave the occupation.
The part worth pausing on is the 42%. Massage is largely a self-employed trade, which is why so much of the “industry data” floating around is estimated rather than counted.
What the evidence actually shows
Here the useful statistics are not percentages of benefit — they are the sizes of the studies and how confident the reviewers were. NCCIH summarises them like this:
- Low-back pain: a 2015 review of 25 studies covering 3,096 people found short-term improvements in pain, but rated the study quality low and concluded the reviewers had “very little confidence” that massage is an effective treatment. The American College of Physicians’ 2017 guideline lists massage as an option for acute or subacute low-back pain — but not for chronic.
- Neck and shoulder pain: a 2013 review of 12 studies (757 people) found massage beat inactive therapies but not other active ones. A 2016 review of 4 US studies (519 people) found short-term relief of neck pain when sessions were long enough and frequent enough.
- Knee osteoarthritis: of 6 studies (408 people), 5 found short-term pain relief.
- Fibromyalgia: a 2014 evaluation of 9 studies (404 people) found that massage continued for at least 5 weeks improved pain, anxiety and depression — but had no effect on sleep disturbance.
- Cancer care: a 2016 evaluation of 19 studies (more than 1,200 people) found some evidence of help with pain and anxiety, but rated the evidence quality very low.
- Headaches: only a small number of studies exist and the results are inconsistent. In one 2015 trial of 56 people with tension headaches, massage was no better than an inactive treatment.
- Premature infants: a 2017 review covering 20 studies and 1,250 infants found improved weight gain. For healthy full-term infants, no clear benefit has been demonstrated.
- Risk: NCCIH describes the risk of harm as low when a trained practitioner does the work, with serious side effects rare.
Read together, the pattern is consistent: modest, short-term relief of pain and tension, on an evidence base that is thinner than the marketing suggests.
Licensing and training
- 45 states and the District of Columbia regulate massage therapy, according to NCCIH. In those states a licence or certification is required before practising, which typically means graduating from an approved program and passing an exam.
- BLS describes the typical entry route as a postsecondary nondegree award — a diploma or certificate taking several months to a couple of years.
Numbers you will see elsewhere that we do not use
Most massage stats posts are built from three things that cannot be verified: paywalled market-research reports, vendor blogs quoting other vendor blogs, and a handful of figures that appear to have no origin at all.
The one worth naming is the claim that a massage lowers cortisol “by up to 30%.” It is repeated across the whole industry, usually credited to a major medical institution. It does not appear on any of those institutions’ pages. NCCIH does not make that claim, and neither will we.
Market-size and growth-rate figures — global market value, regional share, compound annual growth rates — are absent for the same reason. They may well be accurate. We just cannot check them, so we are not going to publish them.
What this means if you are just booking a massage
Very little of the above should change what you do. Massage stats are useful mainly for keeping expectations honest. Massage is a reasonable, low-risk thing to do for a body that feels tight, and the honest expectation is short-term relief rather than a cure. If you have a medical condition, run it past your doctor first — and tell your therapist about it before the session starts.
If you would like the practical version instead of the statistical one, our FAQ answers the questions people actually ask at the front desk, and you can compare session types on our services page. Sessions start from $69 for the hour across our seven San Diego studios — book whichever is closest when you are ready.
Sources
- NCCIH — Massage Therapy: What You Need To Know (usage, evidence, licensing, risk)
- US Bureau of Labor Statistics — Occupational Outlook Handbook: Massage Therapists (employment, wages, projections; wage data May 2024, projections 2024–34)
Figures on this page were checked against the sources above on 26 August 2026. If a number here cannot be found in one of those two documents, it should not be on the page — tell us and we will fix it.
