A California buyer filed a proposed class action against smart ring maker Oura in federal court in San Francisco on August 20, 2026, arguing that the company's advertised sleep staging accuracy figures cannot be produced by a device with no brain or eye sensors. The complaint reads less like a product dispute than an audit of a decade-long paid media operation.

Updated on August 23, 2026 to include a statement from Oura and the studies the company cited in support of it.

The case, Surber v. Oura Inc., was docketed as 3:26-cv-08686 in the United States District Court for the Northern District of California. Clarkson Law Firm filed on behalf of Madison Surber, who purchased an Oura Ring 4 Gold for approximately $513.68 directly from the company's website on or around May 22, 2025, while located in Los Angeles County. Named defendants are Oura Inc. and Oura Health Oy, doing business as Ouraring Inc., both with principal offices at 222 Kearny Street in San Francisco. Venue was pleaded in the San Francisco Division on the grounds that the marketing at issue originated there.

Seven counts follow: fraud by misrepresentation, unjust enrichment, violations of California's Unfair Competition Law, False Advertising Law and Consumers Legal Remedies Act, breach of express warranty under California Commercial Code section 2313, and breach of implied warranty under the Song-Beverly Act. Jurisdiction rests on the Class Action Fairness Act, with the amount in controversy pleaded above $5,000,000 exclusive of costs and interest and the proposed class described as containing 100 or more members. The filing seeks a nationwide class of all United States purchasers who bought for purposes other than resale, plus a California subclass covering the four years preceding the complaint.

The claims the complaint pulled from Oura's own channels

What distinguishes this filing from a routine consumer suit is the inventory of advertising copy it reproduces. The complaint groups these under the label "Challenged Representations" and lists them with source URLs and access dates, most recorded between August 17 and August 19, 2026.

The headline figure is a percentage displayed on the Oura Ring 5 product page: "95% Sleep Staging Accuracy Compared to clinical sleep lab." An earlier company blog post, dated November 16, 2022, is cited for a different number, "79% agreement with gold-standard polysomnography (PSG) for classifying the four stages of sleep (wake, light, deep, and rapid eye movement (REM) sleep)."

Around those numbers sits a layer of unquantified assurance. Product pages carry "Built for accuracy," a phrase the complaint records in two variants, the second extending into a health framing. The Ring 4 page is cited for "Unparalleled Accuracy," the Ring 4 Ceramic page for a Smart Sensing description promising "highly accurate, continuous data." A company page headed Why Oura is cited for "25+ PhDs working across an in-house interdisciplinary science team."

One item is doing heavier work than the rest. A December 6, 2022 Oura blog post is titled, according to the complaint, "Oura Ring: An Alternative to PSG Sleep Studies." A pinned TikTok promotional video for the Ring 4 is described as carrying the overlay "More accurate than ever."

The products covered are the Oura Ring 5, the Oura Ring 4 and the Oura Ring 4 Ceramic, priced from $300 upward.

The sensor argument

The technical core of the complaint is a categorical claim about hardware rather than a dispute over calibration.

Polysomnography, the clinical reference standard, records electroencephalography for brain waves, electrooculography for eye movement, electromyography for muscle tone and electrocardiography for cardiac signals. Under the staging framework the American Academy of Sleep Medicine maintains, REM sleep is identified by three features: rapid eye movements on EOG, muscle atonia on EMG and EEG desynchronisation. Deep sleep, or stage N3, is defined by waves at 0.5 to 2.0 Hz with a minimum amplitude of 75 microvolts, present across at least 20% of a 30-second epoch.

The complaint reproduces the sensor list published on the Oura product pages: red and infrared LEDs for blood oxygen, green and infrared LEDs alternating for heart rate, heart rate variability and respiration rate, a digital temperature sensor, and an accelerometer. No EEG. No EOG. No EMG. No ECG.

From that gap the filing draws a logical rather than empirical conclusion: a device that cannot measure eye movement cannot directly detect the stage defined by eye movement, and a device that cannot measure brain activity cannot directly detect the stage defined by brain activity. What the ring produces instead, according to the complaint, is machine learning inference applied to peripheral physiological signals.

The disclosure that contradicts the sales page

The sharpest passage in the filing is not an outside expert opinion. It is a quotation from Oura's own technical content, in which the company states that using an Oura Ring "is not the same as a PSG (as it does not measure electrical brain activity or eye movements)." A separate February 12, 2024 company blog post is cited for a statement that polysomnography relies on "electroencephalography (EEG) to measure electrical brain waves" in order to identify the stage a sleeper is in.

The complaint places those admissions directly against the product page figure of 79% agreement with polysomnography and against the blog headline describing the ring as an alternative to PSG studies. The alleged defect is one of placement rather than absence: the qualification exists, but it sits in blog and support material rather than at the point of sale.

That structural argument has become the recognisable shape of this litigation category. A federal complaint filed in San Diego on August 12, 2026 made the same argument about front-of-pack ingredient prominence on a Mars rice product, and a November 2025 filing applied it to flag imagery on coffee packaging. In each, the theory is that prominent front-facing signalling creates an impression which inconspicuous secondary text does not correct.

What the studies say

The complaint cites three studies, one of them funded by the defendant.

A study published in Nature on March 19, 2025 gathered data across 45 nights from 45 patients wearing Oura rings. According to the complaint's summary, the rings underestimated light sleep by 13.84 minutes and deep sleep by 5.98 minutes while overestimating REM sleep by 31.56 minutes, producing overall stage classification accuracy of 53.18%. The researchers warned that reasonable average agreement with polysomnography can mask substantial individual-level error.

A second study, from researchers at the University of Massachusetts and published in Sleep Advances in 2026, found significant differences from polysomnography in both younger and older adult groups, with lower accuracy in older adults.

The third is Oura's own. A 2024 validation study of 96 participants and 421,045 epochs, cited in the complaint as company-funded, reported that the products may overestimate light sleep when it runs shorter than 175 minutes and underestimate it beyond 240 minutes, and may overestimate deep sleep under 40 minutes while underestimating it past 90 minutes. The complaint uses that paper twice: once as evidence of inaccuracy, and once as evidence of knowledge.

Oura disputes the allegations

Oura responded to the filing on August 22, 2026 with a statement attributed to a company spokesperson. "We stand behind our science, research, and accuracy claims," the statement opened.

The substance of the response is an argument about inference. According to Oura, the ring estimates sleep stages from multiple physiological signals covering heart rate, heart rate variability, movement, breathing patterns and temperature, and there is well-documented independent peer-reviewed evidence that sleep stages are associated with distinct, measurable and reproducible changes in physiology. That association, the company said, is why those signals are reliably used to classify sleep stages. The framing accepts the complaint's premise that the ring measures peripheral signals rather than neurological ones, and contests the conclusion drawn from it.

The company stated that the ring is not a medical device and not a substitute for a clinical sleep study, while maintaining that its sleep staging has been validated and has compared favourably against polysomnography in several studies. It said the data collection behind its sleep staging algorithm, which it identified by the initials NSSA, covers more than 1,200 nights and is unparalleled in sleep research, that it published a peer-reviewed paper describing how the algorithm was developed and how it uses the signals the ring collects, and that multiple third-party independent studies support its accuracy claims. Oura also said it has reported transparently on the mechanisms and measures behind its sleep staging, and that it remains committed to communicating clearly what the ring measures, what it estimates, and how members are intended to use that information.

On the litigation itself, the position is short. "We dispute the allegations and will defend against them in the appropriate legal forum," the spokesperson said.

Four studies were supplied as supporting sources: a 2018 review of central and autonomic nervous system coupling during sleep by de Zambotti and colleagues in Neuroscience and Biobehavioral Reviews; a 2021 multi-night validation of the ring in adolescents and adults by Altini and Kinnunen in Sensors; a 2022 home polysomnography validation by Ghorbani and colleagues in Nature and Science of Sleep; and a 2024 in-laboratory polysomnography validation by Robbins and colleagues, also in Sensors.

Where the two evidence sets diverge

The two citation lists do not describe the same body of literature, and the points of contact are worth setting out precisely.

They overlap on one paper. The 2021 Altini and Kinnunen article Oura cites appears at Sensors volume 21, article 4302. The complaint cites the same volume and article number, under a variant title, as authority for the proposition that proper sleep measurement requires electroencephalography, electrooculography, electrocardiography and electromyography, and for the statement that Oura's own scientific staff understood the distinction between direct measurement and inference. Both parties are resting part of their case on the same 2021 publication.

Beyond that, the sets do not meet. None of the four studies Oura supplied is the March 2025 Nature paper reporting 53.18% stage classification accuracy across 45 nights, and none is the 2026 Sleep Advances paper from University of Massachusetts researchers reporting lower accuracy in older adults. The most recent validation in Oura's list is dated October 2024. The company also did not address the 79% and 95% figures directly, nor the allegations concerning approximately 91 active Meta advertisements and dynamic creative assembly.

Two wording points sit close to the pleaded claims. The statement applies the word unparalleled to the volume of data behind the algorithm, whereas the complaint challenges "Unparalleled Accuracy" as a phrase appearing on the Ring 4 page. And the assertion that the ring is not a substitute for a clinical sleep study runs alongside the December 2022 company blog headline the complaint reproduces, which describes the ring as an alternative to PSG sleep studies. The complaint's argument on that point concerns placement rather than the existence of the qualification, and the statement does not address where each formulation appears.

There is also a naming gap. Oura identifies its algorithm as NSSA. The self-funded 2024 validation the complaint uses as evidence of knowledge concerns an earlier version the filing labels OSSA 2.0. The statement does not specify which product generations run which algorithm.

Section D of the complaint is unusual for a consumer class action. It is an account of programmatic and social ad delivery mechanics, written to establish that a plaintiff who cannot recall specific advertisements was nonetheless exposed to a uniform message.

According to the filing, Meta Ad Library records reviewed on August 17, 2026 showed approximately 91 advertisements actively running for the brand, and nearly all of them used one of Meta's dynamic services. The complaint describes Dynamic Creative as a system in which images, videos and headlines are automatically mixed to produce personalised ad versions, and describes TikTok Smart Creative as its counterpart. It then does the arithmetic: with four variations each of primary text, description, headline and creative asset, 1,820 possible combinations exist.

That number is the pleading's answer to an evidentiary problem. If every user sees a distinct combination generated in real time, no individual can reasonably be expected to identify which advertisement they saw. The complaint converts creative automation from a performance feature into an argument for class-wide uniformity of message.

Marketers will recognise the underlying products. Meta switched Dynamic Media to on by default for Advantage+ catalog ads in September 2025, aligning API behaviour with an Ads Manager toggle that had defaulted to on since launch. TikTok's Smart+ system, launched in October 2024, automates creative selection alongside targeting and bidding, and gained an ad preview capability in January 2026 precisely because advertisers complained they could not see what combinations were being served. Meta's Advantage+ generates up to 150 creative variations automatically and expands targeting beyond advertiser-defined parameters.

The complaint also devotes several paragraphs to retargeting mechanics, describing pixels as code that tags site visitors so advertisers can reconnect with them later. It cites a 2018 industry figure putting the conversion uplift from retargeting at 70% and branded search exposure gains at a minimum of 1000%, and a 2019 randomised field experiment published in the Journal of Marketing Research finding that retargeting caused 14.6% more users to return to a website within four weeks.

Those statistics are used as an aggravating factor rather than a violation. The theory is that repetition and personalisation compound exposure, so that a claim which might be shrugged off once becomes, in the filing's phrase, a message that follows consumers across platforms.

Scale, endorsement and the audience the complaint describes

The complaint sets out the commercial position it says the marketing produced. Oura Health Oy was founded in Oulu, Finland in 2013 and launched its first product through a 2015 Kickstarter campaign that raised over $650,000, alongside a $2.3 million seed round led by Lifeline Ventures. The company reached a valuation of approximately $11 billion after a Series E round in 2025, with more than 5.5 million rings sold and revenue reported above $1 billion in 2025. Nearly 3 million rings were sold in 2025 alone.

Channel figures recorded as of August 17, 2026 give the paid and organic footprint: roughly 131,000 Facebook followers, about 55,300 followers and 20,200 posts on X, 23,400 YouTube subscribers, approximately 1.1 million Instagramfollowers across 1,254 posts, and around 130,400 TikTok followers with close to 1 million cumulative likes. The earliest promotional post identified is dated March 3, 2015, giving a campaign the complaint describes as running for nearly eleven years.

Celebrity adoption features in the pleading as an amplification mechanism. The filing names Jennifer Aniston, Prince Harry and Kim Kardashian as figures who adopted or endorsed the products publicly and on social media, citing press coverage from 2018 and 2021.

On the demand side, the complaint leans on a November 2023 American Academy of Sleep Medicine finding that 68% of consumers using sleep trackers changed behaviour after learning more about their sleep from those devices. It also cites market sizing placing global sleep trackers at $6.76 billion in 2025 with a projection of $25 billion by 2034.

The competitive framing is narrow and specific. Smart ring makers SleepOn and Circul are named as companies selling comparable sensor sets without stage-accuracy claims, and EEG headband systems are named as the category that measures brain activity directly. The alleged injury therefore runs to competitors as well as buyers.

Notice, remedies and the request that matters to media teams

Counsel sent a CLRA notice letter to Oura Health Oy by certified mail on March 26, 2026, addressed to the Kearny Street headquarters and to the registered agent for service of process. According to the complaint, no adequate correction followed in the roughly five months before filing.

The relief sought includes damages, restitution, disgorgement and punitive damages. It also includes an injunction with a specific media component: a requirement that the defendants run what the complaint calls an affirmative advertising campaign to dispel the public misperception the marketing is alleged to have created, and that prominent qualifications or disclaimers appear on front labelling.

Corrective advertising orders are rare, and their commercial weight is different in kind from a settlement cheque. A company ordered to buy media against its own prior claims carries both the production cost and the reputational effect into the same budget line.

Why this lands on media buyers rather than lawyers

Three features of this complaint make it relevant beyond the wearables category.

The first is evidentiary. Ad transparency repositories were built as accountability tools and have long doubled as competitive intelligence resources. Meta's library added a WhatsApp filter and low impression labels in December 2025, continuing a trajectory toward more granular public disclosure. This filing uses that repository as a plaintiff's exhibit, counting active creatives and characterising the delivery system behind them. Whatever an advertiser runs on those platforms is discoverable without discovery.

The second is about automation and liability. Dynamic creative systems assemble messages the advertiser never individually approved, and the IAB's disclosure framework published this month is explicit that advertisers remain responsible for accuracy and substantiation regardless of authorship method. The Oura complaint applies that principle in the opposite direction, treating combinatorial output as proof that a single core message reached everyone.

The third is category exposure. Health and wellness marketing has been drawing filings at a steady clip. A complaint over patches named for a hormone the products do not contain was filed on August 13, 2026 under the same combination of California statutes. The Federal Trade Commission has a long substantiation record in the sector, including refunds distributed to buyers of weight-loss supplements marketed with unsupported clinical claims, and regulators elsewhere have reached comparable conclusions, with India's consumer authority fining a parenting startup over unverified developmental claims in early 2026.

There is a fourth thread that is easy to miss. Oura data does not stay inside Oura's app. Amazon announced Alexa+ health integrations covering Oura, Withings and Wyze in late 2025, and confirmed early access for Oura ring data inside Alexa+ in January 2026, with the assistant surfacing morning rundowns of sleep and recovery status. A sleep score contested in litigation is also a signal moving into an ambient assistant layer owned by one of the largest advertising businesses in the world.

Nothing has been entered on the docket by the defendants at this stage, and every allegation described here remains an allegation, contested by the company. The specific factual question the case raises, however, is narrower than most advertising disputes and unusually testable: whether a percentage figure describing agreement with a clinical instrument can be advertised as accuracy by a device that does not contain the sensors the instrument uses. Oura's answer is that physiology carries the signal reliably enough for classification. The complaint's answer is that inference and measurement are not interchangeable at the point of sale.

Timeline

Summary

Who: Madison Surber, a California resident represented by Clarkson Law Firm, filed against Oura Inc. and Oura Health Oy, doing business as Ouraring Inc. The proposed nationwide class covers United States purchasers of the Oura Ring 5, Oura Ring 4 and Oura Ring 4 Ceramic, with a California subclass spanning four years. Oura responded through a company spokesperson.

What: A seven-count class action complaint alleging fraud by misrepresentation, unjust enrichment, and violations of California's Unfair Competition Law, False Advertising Law, Consumers Legal Remedies Act, express warranty provisions and the Song-Beverly Act. The core allegation is that advertised sleep staging accuracy figures of 79% and 95% cannot be substantiated by a finger-worn device containing no EEG, EOG, EMG or ECG sensors, and that peer-reviewed testing put overall stage classification at 53.18%. Oura disputes the allegations, stating that the ring is not a medical device or a substitute for a clinical sleep study while maintaining that physiological signals classify sleep stages reliably and that several validation studies support its accuracy claims.

When: The complaint was filed on August 20, 2026, and Oura issued its statement on August 22, 2026. The plaintiff's purchase occurred on or around May 22, 2025. A CLRA notice letter was sent on March 26, 2026. Product page and social media access dates were recorded between August 17 and August 19, 2026.

Where: The United States District Court for the Northern District of California, San Francisco Division, case number 3:26-cv-08686. Both defendants maintain principal offices at 222 Kearny Street, San Francisco.

Why: The filing matters to advertising practitioners because it treats a public ad transparency repository as primary evidence, counting roughly 91 active Meta advertisements and using dynamic creative combinatorics to argue class-wide uniformity of message. It also seeks a corrective advertising injunction, a remedy that converts a legal outcome into a media budget line, and it lands in a health and wellness category where substantiation claims have been generating filings and regulatory action throughout 2026. The dispute over evidence is unusually direct: both sides cite the same 2021 Sensors paper, while the studies each side relies on otherwise do not overlap.