Evidence: Mixed — the evaluation workflow is a versioned E5 method validated with synthetic fixtures, while its labeling, evidence-transfer, and advertising boundaries are independently cross-checked at E2
Risk scope: R0 — editorial methodology and disclosure rules; no diagnosis, treatment, product-safety verdict, or individualized guidance
Direct observation: Not performed for this methodology page
Expert review: Not required for this R0 method; Article-level R2 claims still require exact-scope E4 review
Method checked: July 24, 2026
Key limitation: A method can organize evidence and refuse unsupported conclusions; it cannot manufacture missing product evidence, testing, price data, expert review, or publication authority
Next review: By July 24, 2027, or sooner after a material evidence-model, scoring-schema, disclosure, price-method, CMS, affiliate-policy, or regulatory-boundary change
Quick answer
Formula Clarity evaluates the exact product, formula version, market, claim, evidence, package, price context, and commercial relationship—not the brand name alone. The method uses ordered blocking gates instead of a hidden total score. Missing identity, weak evidence, unreviewed R2 claims, unsupported testing language, unmatched comparisons, stale price inputs, or undisclosed commercial ties cannot be offset by strengths elsewhere. A passing method result supports only the bounded conclusion recorded for that exact evidence set.
If you need the foundational label-reading sequence before applying this method, see how to read an INCI list step by step.
For a concrete example of applying this method to one product, see the CeraVe Moisturizing Cream analysis.
For a second exact application of the method, see the Vanicream Moisturizing Cream analysis.
For the purchase-channel side of a product evaluation, see Brand Store vs Authorized Retailer vs Marketplace: Authenticity, Returns, and Price Trade-offs. It separates the website, seller of record, fulfiller, price, shipping, and return terms without treating a marketplace remedy as product authentication.
Why a published method matters
Beauty-product pages often compress unlike things into one verdict: an ingredient study, a current ingredient list, a brand claim, a package feature, a price, and a personal impression may all appear beside the same rating. That presentation can make the evidence look more unified than it is.
Formula Clarity uses separate records because each object answers a different question. An ingredient declaration can identify names. A regulator can explain a labeling rule. A formulation study can support a finding about the exact material and method tested. A controlled first-party observation can document a package or texture under a stated protocol. A price snapshot can describe one seller, market, size, currency, date, and set of terms. None of those objects automatically proves every other one.
The method therefore has two jobs:
- preserve the identity and limits of each evidence object; and
- stop the analysis when a decision-critical gate has not been met.
It is not designed to make every product rankable. Sometimes the correct output is that identity is incomplete, evidence is limited, products are not comparable, a health-sensitive conclusion needs qualified review, or a price verdict cannot be made from the available snapshot.
The exact object comes first
An evaluation begins with an exact purchase object, not a brand reputation or family name. The minimum identity record includes:
product and variant
market or jurisdiction
package size or configuration
formula or ingredient-list version
claim wording and checked date
relevant directions and warnings
seller and price context when value is discussed
A product can change formula, package, size, claim, seller terms, or market presentation while keeping a similar name. When the exact version is unclear, the method returns not_evaluable_identity_missing for conclusions that depend on it. An older formula cannot silently supply evidence for a newer one.
This identity gate also prevents a common comparison error: pairing one product’s current U.S. formula with another product’s older or regional formula. A comparison must use sufficiently matched markets, variants, sizes, and dates or state why the difference does not change the specific conclusion.
Evidence is claim-specific
Formula Clarity uses an evidence ladder, but the ladder is not a universal product score:
| Level | What it means in this method | What it does not mean |
|---|---|---|
| E0 | An idea or unanswered question | Publishable evidence |
| E1 | A source, label, registry, or claim was captured and dated | Independent confirmation |
| E2 | The material point was cross-checked with fit-for-claim evidence | Direct product testing |
| E3 | An exact observable trait was recorded under a controlled protocol | Clinical efficacy or universal tolerability |
| E4 | A qualified reviewer approved the exact health-sensitive scope | Stronger underlying evidence than actually exists |
| E5 | A method, dataset, or tool passed versioned validation | Permission to skip E4 for R2 claims |
The required level follows the claim. A simple dated brand statement may remain E1. A material buyer conclusion normally needs E2. Testing language requires a real E3 record for the exact observable trait. A retained R2 claim requires relevant E4 review of the exact wording and version. Original calculations or methods require E5 validation, but E5 does not turn medical interpretation into an editorial-only decision.
When evidence does not meet the claim-specific minimum, the output is evidence_limited. The page can accurately attribute what a source says, but it cannot upgrade that statement into independent proof.
Ingredient and formula evidence stay separate
An ingredient list supports declared identity and limited ordering context. It normally does not disclose exact concentration, raw-material grade, purity, pH, processing, delivery system, stability, or finished-product performance. Formula Clarity does not reverse-engineer those properties from ingredient order.
The U.S. ingredient-declaration rules also include ordering exceptions for ingredients at one percent or less and for color additives. That is another reason not to treat list position as a visible percentage scale. See FDA’s Cosmetic Ingredient Names, FDA’s labeling summary, and 21 CFR 701.3.
A study transfers only as far as the tested object supports. The method records the ingredient or material, concentration, vehicle, comparator, population or model, duration, application schedule, endpoint, measurement method, funding, conflicts, and whether the sold product is actually the tested product. A shared ingredient name is not enough to claim equivalent delivery or results.
Claims and classification are checked before conclusions
A front-label statement or product-page statement is a brand claim unless independent evidence confirms it. Formula Clarity captures the exact wording, claimant, market, date, and version before assessing substantiation. FTC’s advertising-substantiation policy is relevant because objective claims need an appropriate evidence basis; it does not make every marketing phrase independently verified. See FTC’s Policy Statement Regarding Advertising Substantiation.
Claims also affect regulatory scope. FDA explains that cosmetic claims must be truthful and not misleading, and that claims involving treatment or effects on body structure or function can change a product’s regulatory character. The method records this boundary rather than issuing a legal classification for an unreviewed product. See FDA’s Cosmetics Labeling Claims.
Any claim that enters diagnosis, treatment, pregnancy, prescriptions, procedures, serious reactions, toxicity, ingestion, or individualized safety is routed to the R2 gate. Without a verified reviewer and an exact-scope E4 record, the output is blocked_r2_without_e4. A disclaimer cannot make the claim publishable.
Direct observation has a narrow role
Formula Clarity may eventually record observable traits such as package dimensions, dispensing behavior, measured mass or volume where the protocol is valid, texture, finish, or pilling under stated conditions. The record must identify the exact product, protocol, environment, date, operator, and result.
Observation does not prove efficacy, long-term stability, safety, or typical experience for every user. If an article says “we tested,” “we tried,” or reports a first-party result without the exact E3 record, the observation gate returns blocked_observation_claim.
Manufacturer descriptions and retailer copy are not first-party observations by Formula Clarity. Product access, samples, gifts, returns, and disposal must also be disclosed separately when they exist.
Comparison uses predeclared dimensions
A comparison begins by defining the buyer question and the candidate universe before judging products. Inclusion and exclusion rules, evidence cutoffs, missing-data behavior, and comparison dimensions must be recorded in advance.
Typical dimensions may include:
exact formula and market identity
claim and evidence quality
relevant ingredient or formula context
package and storage information
verified observable traits
size and use assumptions
price scenario
commercial relationship
material unknowns and stop conditions
The method does not assume that every dimension applies to every page. It does require equivalent treatment of paying and non-paying products. A product cannot receive a favorable conclusion merely because it has an affiliate program, a higher commission, more retailer availability, or more attractive marketing assets.
When variants, markets, evidence dates, product roles, or decision criteria are not sufficiently matched, the output is not_comparable. The analysis may still describe each object separately.
Price and value are their own evidence layer
A price statement is incomplete without seller, market, currency, size, checked date, stock condition, shipping, tax treatment, membership or coupon conditions, and the arithmetic used. A temporary promotion is not a permanent product attribute.
Value is scenario-dependent. The method may normalize price per unit, per expected use, or across a declared usage scenario only when the inputs and assumptions are shown. Price does not prove efficacy, and a higher ingredient percentage does not automatically justify a premium.
When a page requests a price or premium verdict without an approved current price method and dated snapshots, the output is blocked_price_method. Unknown shipping, taxes, or usage assumptions remain unknown rather than being converted to zero.
Commercial relationships never change the evidence standard
Affiliate links, gifted access, sponsorship, samples, or other material connections must be disclosed before the relevant decision link or claim. Commercial eligibility is not an inclusion, scoring, or ranking criterion.
The method keeps two questions separate:
- Does the evidence support the conclusion?
- Is there a commercial relationship that must be disclosed and implemented correctly?
A relationship may affect disclosure and link handling, but it cannot improve evidence quality. Undisclosed material connections or commercial influence on inclusion produce blocked_commercial_disclosure.
No hidden universal total score
Formula Clarity does not use one secret number that lets strengths compensate for failures elsewhere. A product with excellent documentation cannot compensate for the wrong formula version. A low price cannot compensate for missing evidence. A polished package cannot compensate for an unreviewed R2 claim. An affiliate opportunity cannot compensate for a failed comparison.
The ordered gates are:
identity
→ evidence
→ risk
→ observation
→ comparison
→ price
→ commercial
→ publication
The first failed decision-critical gate controls the result. Passing later gates does not erase the failure. Noncritical unknowns can produce comparable_with_limits; complete supported inputs can produce decision_supported. Neither output means “best for everyone.”
What the decision record contains
A method output stores more than a verdict:
exact object and version
buyer question
criteria and candidate universe
evidence level per material claim
risk tier and reviewer status
observation status
comparison parity
price method and checked snapshots
commercial relationship and disclosure status
material unknowns
blocking gate or supported conclusion
refresh triggers and correction path
The reader-facing conclusion should explain why the result was reached and what could change it. Internal claim IDs, private correspondence, credentials, tokens, preview URLs, and restricted evidence do not enter the public body.
Synthetic examples used to validate the method
The method is tested with synthetic fixtures rather than favorable real-product examples. The fixtures confirm that:
- missing formula identity stops the evaluation;
- evidence below the required level stays limited;
- retained R2 wording without E4 is blocked;
- testing language without E3 is blocked;
- unmatched variants are not treated as comparable;
- a price verdict without a price method is blocked;
- an undisclosed material connection is blocked; and
- a complete, non-medical, non-commercial evidence set can reach a bounded supported decision.
These fixtures validate decision logic. They do not prove that any retail product passes.
What this method refuses to claim
This method does not promise that Formula Clarity has purchased or tested a product unless a page links to the exact observation record. It does not convert ingredient evidence into finished-product efficacy. It does not certify regulatory compliance, authenticity, safety, or suitability for an individual. It does not use ratings as proof of typical outcomes. It does not treat “not found” as “false.” It does not publish an R2 conclusion without relevant E4 review. It does not treat affiliate availability as evidence.
The method also does not make CMS status equal publication. A WordPress Draft, private preview, scheduled timestamp, or editor screen is not a public acceptance record. Public status requires the exact release authorization, canonical and indexability checks, required disclosures and links, and logged-out acceptance after go-live.
Corrections, updates, and method changes
A material formula, package, claim, price, seller-term, affiliate, source, reviewer, or method change opens a refresh transaction. The affected conclusion is re-evaluated under the current method. A date-only save does not count as a meaningful update.
Method changes require a new version, updated schema or rules, fresh fixtures, validation, and a change record. Historical outputs retain the method version that produced them. A newer method does not silently rewrite an older accepted decision.
Methodology and limitations
This page describes Formula Clarity’s internal evaluation method and public refusal rules. The method specification, machine-readable schema, fixtures, and validator are versioned in the project repository. External boundaries are based on U.S. labeling and advertising sources checked for this version. The method does not provide medical, legal, regulatory-compliance, laboratory, or individualized buying advice. It can improve consistency and expose missing evidence, but it cannot eliminate source error, undisclosed reformulation, counterfeit products, market variation, or the need for qualified review.