We took 142 health claims off the marketing pages of 30 standing desks and matched each one to the strongest peer-reviewed evidence we could find. Most of them do not hold up.
Every distinct claim was scored as one of four things: supported (direction and magnitude match the evidence), partially supported (right direction, overstated or short-term), unsupported (no evidence base as marketed), or contradicted (the best evidence points the other way). Here is the whole audit in one bar.
Only 18 of 142 claims are fully supported. The single largest bucket is unsupported. Add the contradicted claims and more than half the marketing on these pages has no evidence behind it, or has evidence pointing the wrong way.
The audit is not uniformly bleak. Sit-stand desks genuinely help with a couple of things, and the evidence says so. The rest is where the marketing outruns the science. Each bar is one claim category, split by verdict and sorted from best-supported to worst. Same evidence bar for every category.
Click any category for the receipts: every claim in that category, the product it came from, its verdict, and the study it was graded against.
The same physiology is for sale from every brand, so the difference is in the copywriting. Our transparency score (0 to 100) rewards citing primary literature, hedging contested claims with "may" and "supports", and avoiding superlatives like cure, eliminate, or guaranteed. Scoring was blinded to the verdict counts to avoid circularity. Sorted most transparent first.
Click any brand for its full claim sheet: every product, every claim, and how each one graded. No brand in the audit clears 75 of 100.
This is a health-adjacent audit, so the grading rule matters as much as the result. Claims were matched to the strongest available source in a fixed preference order, graded on Cochrane GRADE conventions, then scored against that evidence with one of four verdicts.
Method. Across 30 sit-to-stand products from 14 brands sold in the United States during Q2 2026, every distinct health, productivity, or ergonomic claim was captured from the brand product page, its category landing page, and the first paid placement for the product name, then deduplicated to a canonical list of 142. Mechanical-spec claims (height range, lift speed, weight capacity) were excluded. Each claim was matched to its strongest source and scored on the four-verdict rubric above. The anchor studies are the Cochrane workplace-interventions review (back pain), Saeidifard et al. 2018 (energy expenditure, roughly 8 to 9 kcal per hour standing versus sitting), Buckley et al. 2014 (post-meal glucose), Stamatakis et al. 2018 (standing and cardiovascular mortality), Tuchsen et al. 2005 (standing and varicose veins), and Chambers et al. 2019 (productivity). Brand transparency is a 0 to 100 composite of (a) citations on the product page, (b) accuracy of any cited studies, (c) absence of unsupported superlatives, and (d) hedged language on contested claims, scored blind to verdict counts.
Data provenance. SOURCED: brand claims collected from each product page and marketing copy (Q2 2026), plus the strongest available peer-reviewed evidence matched to each. DERIVED: the claim-to-evidence grading and the transparency scores. Every number on this page is computed live in your browser from the same open dataset linked below.
Limitations. Five apply. Claim collection is a single point in time in Q2 2026; brands routinely refresh product pages and may have moderated claims since. Evidence grades depend on the strongest study at the time of writing; future Cochrane updates may shift them. The 30-product sample covers the largest direct-to-consumer and contract brands but excludes regional and white-label brands. Transparency scoring is partly subjective despite blinding; the rubric is published so anyone can re-score. And this audit evaluates claim accuracy, not ergonomic performance or build quality, a brand can score low on transparency and still make a well-engineered desk. Read the numbers as a graded snapshot of marketing copy, not a verdict on the hardware.
Every verdict, category split, and brand score on this page is generated from the raw claim-level records. Take the data, read the grading rule, and re-score any claim yourself.
The CSV is one row per claim: product, brand, price, category, claim text, evidence grade, verdict, and the citation it was graded against. The full narrative write-up, with every finding and chart, lives on the dataset page.