Most supplement brands tell consumers what an ingredient is supposed to do.
Far fewer tell them how confident they should be that it actually does it.
A vitamin page might say:
- Supports energy
- Promotes calm
- Maintains immunity
- Helps manage stress
- Supports hormonal health
- Encourages restful sleep
The claims may sit beside scientific-looking graphics, medical terminology and pictures of plants or laboratories.
What is usually missing is an admission that the evidence behind those statements varies dramatically.
Before it closed in 2024, personalized supplement company Care/of tried something unusual.
Its product pages displayed an evidence label beside the claimed benefits of individual vitamins, minerals and botanicals.
Depending on the ingredient, consumers could encounter ratings such as:
- Very Strong Research
- Strong Research
- Good Research
- Emerging Research
- Limited Research
- Traditionally Used
The site also provided summaries of the evidence and references that consumers could follow into the research literature. Independent reviews documented that vitamin D received a Very Strong Research grade, while shatavari was labelled Limited Research and Traditionally Used.
Care/of's biggest innovation may therefore not have been personalization, daily sachets or a new vitamin formula.
It was admitting—on the page where the product was being sold—that some supplement claims were much better supported than others.
This is one signal from the Consensys Innovation Signals Engine, which continuously scans a library of more than one million products worldwide for emerging shifts in formulation, positioning and consumer demand.
Signal: Evidence Uncertainty as Consumer Information
Care/of is now a historical case study
Care/of was founded in 2016 as a direct-to-consumer personalized supplement service.
Consumers completed a health and lifestyle questionnaire, received a recommended selection of supplements and could subscribe to daily packets printed with their names.
In November 2020, Bayer acquired a majority stake in the business. Bayer described Care/of as a personalized nutrition company that combined consumer data, technology and direct-to-consumer delivery.
The expansion did not last.
In 2024, Care/of announced that it would stop accepting orders, cancel subscriptions and wind down operations. Reporting linked the closure to a loss of funding within the Bayer-owned business, with 143 employees affected by the shutdown.
Brand: Care/of
Majority owner from 2020: Bayer
Operations ended: June 2024
Current status: Closed
Evidence Correction: Former Bayer-Owned Brand, Not Current Operating Service
The evidence-rating system should therefore be examined as an innovation that disappeared with the business—not as a feature consumers can currently use.
The rating system evolved over time
At Care/of's launch in 2016, reporting described a relatively simple three-level system:
- Very Strong Research
- Mixed Research
- Emerging Research
Co-founder Craig Elbert said the objective was to communicate honestly where evidence was strong and where it was not. The company's scientific advisory board reportedly helped review and classify the research.
Later versions became more granular.
Independent product reviews recorded the following labels:
- Care/of label
- Public explanation reported by reviewers
- Very Strong Research
- Many reputable studies support the use
- Strong Research
- Some high-quality studies demonstrate efficacy
- Good Research
- Fewer high-quality studies or a smaller evidence base
- Emerging Research
- Limited recent evidence
- Limited Research
- Very little evidence for efficacy
- Traditionally Used
- Historical use in herbal or traditional medicine
- Innovation Type: Consumer-Facing Evidence Hierarchy
The system transformed scientific uncertainty into language a shopper could understand quickly.
It was not a conventional medical evidence-grading framework
Evidence-based medicine uses formal systems to evaluate questions including:
- Study design
- Risk of bias
- Consistency
- Precision
- Directness
- Publication bias
- Size and clinical importance of the effect
Frameworks such as GRADE distinguish the certainty of a body of evidence from the strength of a recommendation. A randomized trial may begin as higher-certainty evidence, but concerns involving bias, inconsistency or indirectness can lower confidence.
Care/of's labels were much simpler.
No complete public methodology was located explaining:
- How many studies qualified as "many"
- What counted as a reputable study
- Whether randomized trials received greater weight
- How study quality was assessed
- Whether company-funded trials were downgraded
- How conflicting findings were handled
- Whether evidence had to match the exact dose and extract sold
- Whether health outcomes were separated from biomarker changes
- How frequently ratings were updated
- Evidence Gap: Public Labels Without a Reproducible Scoring Protocol
The ratings offered orientation.
They did not provide enough information for an independent researcher to recreate every grade.
A category label can hide important differences inside the evidence
Consider the phrase Very Strong Research.
It could mean that research strongly supports:
- The nutrient's essential biological function
- Treatment of a diagnosed deficiency
- Prevention of a deficiency disease
- A benefit among a specific high-risk population
- Routine supplementation for the general population
Those are not equivalent claims.
Vitamin D provides a useful example.
There is very strong evidence that vitamin D is essential for bone and mineral metabolism and that supplementation can prevent or treat nutritional rickets and correct severe deficiency. Evidence for broad supplementation producing multiple additional benefits in already sufficient adults is much less consistent.
A label saying Very Strong Research is therefore only meaningful when attached to a clearly defined outcome.
Evidence Principle: Strong Evidence for Which Claim?
Spot check one: Care/of's vitamin D rating was directionally reasonable
Independent reviews reported that Care/of graded vitamin D as Very Strong Research.
That rating is defensible for core claims involving:
- Vitamin D's essential physiological role
- Prevention and treatment of deficiency
- Bone health in relevant populations
- Prevention of nutritional rickets
Clinical reviews conclude that severe vitamin D deficiency should be corrected and that supplementation prevents and treats nutritional rickets.
But the same rating would be too broad if used to imply that routine supplementation reliably:
- Prevents cardiovascular disease
- Prevents cancer
- Improves mood
- Prevents respiratory illness
- Increases energy
- Benefits every consumer regardless of baseline level
- Spot-Check Result: Grade Plausible, Claim Scope Essential
Care/of's label helped consumers compare the maturity of the research, but the exact wording of the associated benefit remained crucial.
Dose and baseline status still matter
A supplement can contain an ingredient supported by strong evidence while delivering a dose inappropriate for the consumer's actual condition.
One review noted that Care/of's vitamin D product contained 1,000 IU and warned that this amount might not be sufficient to treat an established deficiency. The same review also cautioned that an online questionnaire cannot replace blood testing and medical assessment.
That reveals a limitation of evidence ratings.
The research may support the ingredient.
It does not prove that:
- The consumer needs it
- The selected dose is correct
- The consumer is deficient
- The quiz identified the deficiency accurately
- The product will produce the desired result
- Trust Distinction: Evidence for Ingredient Versus Evidence for Recommendation
- Spot check two: shatavari deserved a low-confidence label
Care/of reportedly labelled shatavari as both:
- Traditionally Used
- Limited Research
The explanation acknowledged its long history within Ayurveda while admitting that modern clinical evidence was sparse.
That was a more responsible presentation than simply describing the botanical as proven support for women's health.
Human research on standardized shatavari extracts is now developing, including small randomized trials in women. But the evidence base remains much smaller than that for established nutrient-deficiency treatment, and results depend on the specific extract, dose and population studied.
Spot-Check Result: Appropriate Separation of Tradition From Clinical Proof
The label communicated two different facts:
The plant has historical use.
Historical use is not equivalent to extensive clinical evidence.
That distinction remains rare in supplement marketing.
"Traditionally used" should not sit on a scientific ladder without explanation
Traditional use is not simply a weaker randomized trial.
It is a different type of evidence.
It may document:
- Longstanding cultural use
- Recorded preparation practices
- Practitioner experience
- Historical indications
- Patterns of tolerability
It does not by itself establish:
- Efficacy over placebo
- Optimal dose
- Treatment effect size
- Safety in pregnancy
- Interaction risk
- Effectiveness of a modern extract
Health-research guidance recognises that documented traditional use may support certain claims, but modern evaluation still requires attention to safety, preparation and the nature of the claimed benefit.
Evidence Correction: Traditional Evidence Is a Separate Category, Not Merely Low-Quality Clinical Evidence
Care/of's decision to display both labels was useful, but the relationship between them needed clearer explanation.
Spot check three: magnesium demonstrates why benefit-specific grading matters
Magnesium is unquestionably an essential mineral involved in numerous physiological processes.
Evidence is strong for correcting magnesium deficiency.
Evidence that magnesium supplementation improves sleep in the general population is much weaker.
A systematic review of randomized studies in older adults found only low- to very-low-certainty evidence of improvement in insomnia symptoms.
A later review similarly found a mixed body of sleep-related research rather than one consistent effect across all participants and formulations.
Therefore:
- Magnesium claim
- Approximate evidence position
- Magnesium is an essential nutrient
- Strong
- Supplementation corrects deficiency
- Strong
- Magnesium improves sleep in every adult
- Weak or uncertain
- Magnesium glycinate is universally the best sleep form
- Not established
- Spot-Check Result: One Ingredient Can Require Several Different Evidence Grades
A single badge attached to the entire ingredient could oversimplify these differences.
The system graded ingredients more clearly than personalized recommendations
Care/of's quiz collected information such as:
- Age
- Sex
- Location
- Health concerns
- Diet
- Lifestyle
- Areas of interest such as sleep, digestion, energy or immunity
It then recommended a personalized packet.
The company's transparency system explained the evidence behind the ingredients.
It was less clear about the evidence behind the recommendation algorithm itself.
No robust public validation study was located showing that Care/of's questionnaire:
- Identified nutrient deficiencies accurately
- Improved clinical outcomes
- Outperformed a standard multivitamin
- Reduced unnecessary supplementation
- Matched recommendations based on blood tests
- Detected medication interactions comprehensively
Independent reviewers warned that the questionnaire did not obtain a complete medical history and could recommend products that were unnecessary or potentially unsuitable.
Evidence Gap: Transparent Ingredients, Less Transparent Personalization Logic
The recommendation could feel more medical than it was
Care/of's personalization language risked creating the impression of a clinical prescription.
But an online wellness questionnaire is not equivalent to:
- Laboratory testing
- Dietary assessment by a registered dietitian
- Medication review
- Medical diagnosis
- Evaluation of symptoms by a clinician
A person reporting low energy might be experiencing:
- Iron deficiency
- Inadequate sleep
- Depression
- Thyroid disease
- Low calorie intake
- Medication effects
- Another medical condition
A vitamin quiz cannot reliably separate all of those possibilities.
Risk Signal: Personalization Interface Creating Diagnostic Authority
Evidence transparency around the product did not validate every recommendation generated by the system.
Citations created friction against exaggerated marketing
Ordinary supplement marketing is optimized for speed.
The consumer sees a claim, a benefit icon and a purchase button.
Care/of inserted another step:
Read the evidence label.
Open the research summary.
Inspect the citations.
Notice that some ingredients have weaker support.
That design created productive friction.
Innovation Type: Research Disclosure at the Point of Sale
The company effectively told customers:
You may still choose this ingredient, but you should know that the evidence is emerging.
That is a stronger trust mechanism than hiding uncertainty in small print.
A citation is not automatically substantiation
Publishing research links can still create an illusion of scientific strength.
A bibliography may include:
- Animal studies
- Laboratory studies
- Small uncontrolled trials
- Observational associations
- Studies using a different dose
- Studies using another extract
- Reviews that conclude evidence is insufficient
- Research funded by the ingredient supplier
Consumers may count citations without evaluating them.
Risk Signal: Citation Volume Mistaken for Evidence Quality
The correct question is not:
How many papers are listed?
It is:
- Do the strongest relevant studies support the exact commercial claim?
- The exact product form matters for botanicals
For vitamins and minerals, chemical form and dose still matter.
For botanicals, the problem becomes larger.
Evidence may apply only to:
- A specific species
- A specific plant part
- A standardized extract
- A defined concentration
- A defined daily dose
- A defined duration
- A study of one ashwagandha root extract does not validate every product labelled "ashwagandha."
A study of a shatavari extract does not validate an unspecified powder at a lower dose.
Evidence Principle: Ingredient Name Is Not Research Equivalence
An ideal grading system would assess the exact commercial material rather than the general ingredient category.
Care/of did not publish enough information about grade governance
The company said its scientific advisory board helped evaluate products and research.
Early reporting identified advisers connected with institutions including the USDA Human Nutrition Research Center on Aging at Tufts and the Harvard T.H. Chan School of Public Health.
However, the available public record does not clearly establish:
- Who assigned each final grade
- Whether reviewers voted
- How conflicts were resolved
- Whether advisers were paid
- How often grades were reconsidered
- Whether negative studies triggered downgrades
- Whether the commercial team could challenge a rating
- Whether an independent group audited the process
- Governance Gap: Scientific Advisory Board Without Fully Public Decision Rules
The system was transparent at the output level.
Its internal decision process remained partly opaque.
Evidence grades can become marketing assets
A Very Strong Research badge can improve conversion.
It creates:
- Authority
- Simplification
- Reassurance
- Differentiation
- A reason to choose the brand
This means the company assigning the score also had a financial interest in the result.
Conflict Signal: Seller Also Acts as Evidence Grader
That does not make the grade false.
It means the strongest model would include independent oversight.
A better version would separate three questions
Future supplement evidence systems should display separate ratings for:
1. Evidence that the ingredient changes the proposed outcome
For example, whether magnesium improves measured sleep quality.
2. Evidence that the exact product delivers the studied material
For example, whether the commercial extract matches the tested extract and dose.
3. Evidence that this consumer is likely to benefit
For example, whether the person is deficient or belongs to the population studied.
Innovation Territory: Three-Layer Supplement Evidence Rating
Care/of primarily addressed the first layer, with some information about sourcing and formulation.
Its personalization quiz attempted the third without clinical validation equivalent to a medical assessment.
Safety needs its own rating
A product can have promising efficacy evidence but uncertain safety in certain groups.
A complete consumer system should display:
- Evidence of benefit
- Evidence of safety
- Known interactions
- Contraindicated populations
- Maximum studied duration
- Whether pregnancy data exist
- Whether long-term evidence exists
- Innovation Type: Benefit and Safety Scored Separately
Care/of did provide precautions, but its headline research grade focused mainly on efficacy and use.
The FDA framework allows broad structure-function language
US dietary supplements may make structure-function statements such as:
- Supports immune health
- Helps maintain energy metabolism
- Promotes relaxation
These claims are not approved by FDA before marketing in the same way as drug-treatment claims.
Companies are responsible for having substantiation and must use the required disclaimer stating that the claim has not been evaluated by FDA and that the product is not intended to diagnose, treat, cure or prevent disease.
This environment encourages language broad enough to remain legally defensible while still sounding beneficial.
Market Signal: Regulatory Structure Rewards Generalized Benefit Language
Care/of's evidence labels created a voluntary layer of specificity that regulation did not force it to provide.
Evidence transparency did not save the business
Bayer acquired a majority interest in Care/of in 2020, when personalized nutrition appeared to offer a major growth opportunity.
Less than four years later, the business closed.
That does not prove consumers disliked the evidence-grading system.
The shutdown may reflect a wider combination of:
- Subscription economics
- Customer-acquisition cost
- Retention
- Competitive pressure
- Operational complexity
- Bayer's portfolio priorities
- Funding decisions
Public reporting confirmed the closure but did not provide a complete audited explanation of which commercial factors were decisive.
Commercial Signal: Trust Innovation Does Not Guarantee Business Sustainability
Care/of could be influential as an idea while unsuccessful as a company.
The system's disappearance leaves a real market gap
Most current supplement retailers still organize products around:
- Benefit
- Ingredient
- Format
- Life stage
- Consumer goal
They rarely allow shoppers to filter by:
- Strength of evidence
- Number of randomized trials
- Consistency of results
- Relevance to the exact dose
- Independent replication
- Safety confidence
- White Space: Evidence-First Supplement Retail
A future platform could let consumers sort products by:
- High-confidence benefit
- Moderate-confidence benefit
- Early-stage research
- Traditional use only
- Evidence of no meaningful effect
- Insufficient data
That would be more useful than displaying only five-star customer reviews.
Negative and mixed evidence should be visible
A scientifically responsible database should not only reward ingredients with positive studies.
It should display when:
- Large trials found no benefit
- Reviews concluded evidence was insufficient
- Results were inconsistent
- Benefits appeared only in deficient participants
- Surrogate biomarkers improved without clinical outcomes
- Safety signals emerged
- Innovation Territory: Evidence Grades That Can Go Down
Care/of's earliest system reportedly included Mixed Research, a particularly useful label because it acknowledged disagreement rather than placing every ingredient on a simple weak-to-strong ladder.
The later public summaries emphasized categories such as Good, Emerging and Limited Research.
"Mixed" should arguably have remained central.
The most honest grade may be "it depends"
Scientific evidence often differs according to:
- Population
- Baseline status
- Dose
- Duration
- Outcome
- Product form
- Health condition
Vitamin D may have very strong evidence for treating deficiency but weak evidence for improving energy in a vitamin-D-sufficient adult.
Iron may be essential and effective for iron-deficiency anemia but inappropriate for someone with adequate iron stores.
Magnesium may be biologically essential while having uncertain effects as a universal sleep aid.
Evidence Signal: Claim-Specific Grades Are Better Than Ingredient-Wide Grades
Care/of showed that admitting uncertainty can increase trust
Supplement marketing often assumes that confidence sells.
Care/of tested the opposite idea.
It suggested that a brand could say:
This ingredient is strongly supported.
This one is promising.
This one is mostly traditional.
This one has limited modern research.
That honesty may make the strongly supported products feel more credible.
Consumer Psychology Signal: Selective Uncertainty Makes Strong Claims More Believable
When every ingredient is described as transformational, none of the claims has much informational value.
The best future system would be independently auditable
A robust evidence-grading platform should publish:
- The exact claim being graded
- Search dates and databases
- Included and excluded studies
- Study-design hierarchy
- Risk-of-bias assessment
- Funding and conflict disclosures
- Dose and formulation matching
- Population relevance
- Effect size
- Safety findings
- Update schedule
- Identity of reviewers
It should also provide version history:
- Previous grade
- New evidence
- Reason for change
- Date updated
- Innovation Type: Version-Controlled Consumer Evidence
This would turn supplement evidence from marketing copy into public infrastructure.
Care/of's real innovation was not proving that supplements work
The brand did not solve the scientific uncertainty surrounding the category.
It did something more modest and still unusual.
It made the uncertainty visible.
The system helped consumers see that:
Vitamin D and shatavari do not have equivalent evidence bases.
Traditional use and modern clinical research are different.
A recommendation may contain ingredients at several levels of scientific maturity.
Citations can be inspected rather than merely implied.
Its limitations were equally important:
The full grading methodology was not public.
Grades could be broader than the exact benefit claimed.
Ingredient evidence did not validate the personalized recommendation.
The company graded products from which it earned revenue.
The business eventually closed.
Care/of should therefore not be remembered as the company that scientifically proved personalized vitamins.
Care/of should be remembered as the company that asked a more useful question than most of the category:
- How strong is the evidence—and are we willing to tell the customer when it is weak?
- Brand Radar Signal Tags
- Brands and Organisations
Care/ofBayerCare/of Scientific Advisory BoardHarvard T.H. Chan School of Public HealthTufts USDA Human Nutrition Research Center on AgingU.S. Food and Drug Administration
Evidence Labels
Very Strong ResearchStrong ResearchGood ResearchEmerging ResearchLimited ResearchMixed ResearchTraditionally Used
Products and Ingredients
Vitamin DShatavariMagnesiumIronPersonalized Vitamin PacksBotanical SupplementsMineral Supplements
Innovation Types
Consumer-Facing Evidence HierarchyResearch Disclosure at Point of SaleEvidence-First Supplement RetailClaim-Specific Evidence RatingThree-Layer Supplement Evidence RatingBenefit and Safety ScoringVersion-Controlled Consumer Evidence
Trust Mechanisms
Peer-Reviewed CitationsResearch SummariesScientific Advisory BoardIngredient Sourcing TransparencyAcknowledgement of Limited EvidenceTraditional-Use DisclosurePublic Uncertainty
Evidence Signals
Vitamin D Rated Very StrongShatavari Rated Limited and Traditionally UsedMagnesium Sleep Evidence Remains Low CertaintyIngredient Evidence Does Not Validate Recommendation AlgorithmNo Complete Public Rubric LocatedNo Independent Grade Audit Located
Consumer Benefits
Easier Research InterpretationComparison of Strong and Weak ClaimsReduced Scientific AmbiguityAccess to Primary ReferencesMore Critical Supplement Selection
Risk Signals
Seller Acting as Evidence GraderIngredient-Level Grade Too BroadCitation Count Mistaken for QualityTraditional Use Mistaken for Clinical ProofQuiz Mistaken for DiagnosisDose Not Matched to DeficiencyEvidence Grade Mistaken for Personal NecessityClosed Brand Presented as Active
Commercial Signals
Bayer Majority Acquisition in 2020Care/of Closure in 2024Subscription Nutrition EconomicsTransparency Innovation Without Commercial SurvivalHistorical Brand Case Study
Sources
Care/of evidence-rating system
NutraIngredients—Care/of launch model: Reports that the original system used Very Strong Research, Mixed Research and Emerging Research labels and quotes the founders' intention to be honest about the state of the science.
Philadelphia Magazine—Care/of transparency: Describes the scientific advisory board's role and the early range from Emerging to Very Strong and Mixed Research.
Medical News Today—Care/of research categories: Records the later categories including Traditionally Used, Very Strong, Strong, Good, Emerging and Limited Research.
Healthline dietitian review: Confirms that product pages included grades, summaries and references; identifies vitamin D as Very Strong Research and shatavari as Limited Research and Traditionally Used.
Ownership and closure
Bayer acquisition announcement: Confirms that Bayer acquired a majority stake in Care/of in November 2020. https://www.crainsnewyork.com/health-pulse/personalized-vitamin-company-careof-will-lay-143-workers-summer
NutraIngredients—Care/of closure: Reports that the Bayer subsidiary would wind down and lay off 143 employees following a loss of funding. https://www.nutraingredients.com/Article/2024/06/07/Care-of-to-close-operations/
TechCrunch—Subscription shutdown: Reports that Care/of cancelled subscriptions and stopped accepting orders in June 2024.
Ingredients Network—Brand closure: Confirms that the former website announced the shutdown and cessation of new orders.
Independent evidence spot checks
Vitamin D health-outcomes review: Distinguishes well-established deficiency and skeletal benefits from inconsistent broader preventive claims.
Vitamin D clinical review: Reviews established and proposed uses while illustrating why claim scope matters.
Magnesium and insomnia systematic review: Finds only low- to very-low-certainty evidence of improvement in older adults.
Magnesium supplementation review: Reports mixed sleep findings across the available intervention literature.
Traditional-medicine research guidance: Explains the role and limitations of documented traditional use as evidence.
Evidence-grading context
GRADE overview: Explains that formal evidence assessment considers confidence in estimates and separates certainty from recommendations.
Levels of evidence and risk of bias: Describes why study design alone is insufficient without assessment of methodological quality.