Review methodology

What “reviewed” means

The current product records are researched profiles and formula assessments. We verified cited identity, formula and label records and interpreted them against ingredient evidence. We did not personally test these products. “Reviewed” never means hands-on use unless the page separately documents the test, dates, conditions and limitations.

We maintain four distinct evidence states:

1. Product profile: structured identity and context, without a hands-on test claim.
2. Formula assessment: analysis of a dated, market-specific ingredient declaration and its limitations.
3. Short-term use note: limited observations over a stated period, without long-term conclusions.
4. Structured test: a documented protocol, sample size, conditions, conflicts and limitations.

The page label must match the work performed. The current dataset uses the first two states only.

How products are selected

Editorial relevance comes first: the product should answer a real comparison question in the dandruff cluster and have a stable Indian-market identity. We prefer official manufacturer information, then manufacturer packaging or documents, regulatory material and reputable Indian retailers. Availability at one merchant is not an inclusion guarantee.

We do not force active-category coverage when reliable formula data is absent. Affiliate status does not improve a product’s chance of inclusion.

Formula verification and conflicts

Each formula is tied to a market, represented pack, stable evidence IDs and sources. Each observation records its evidence type, captured and verified dates, verifier, transcription and confidence. A full official declaration supports higher formula confidence; an active-only “shampoo base q.s.” declaration remains partial.

Online verification and physical-pack verification are separate. A manufacturer or retailer page is remote evidence even when it shows packaging. “Physically verified” is used only when an editor genuinely inspected an identified pack and linked the required physical-pack evidence and asset metadata.

If sources disagree, we do not blend them into a plausible-looking list. The selected current formula and the disputed declaration remain separate records. An undated conflict lowers confidence. When succession is established, the older formula can be retained as superseded or archived rather than erased. Different lists do not establish an exact reformulation date.

Evidence hierarchy

For formula identity, we prefer an identified physical pack, then manufacturer documents or regulatory records, manufacturer pages or direct correspondence, retailer packaging images, retailer pages, and archived/contextual sources. A documented exception may change that order for a specific record. Retailer imagery is never relabelled as physical possession.

Manufacturers establish product identity, pack declarations, directions and warnings. They do not establish clinical effectiveness merely by making a claim. For ingredient and medical context we prefer clinical guidelines, government or public-health material, systematic reviews, peer-reviewed trials and official monographs. Retailer copy is corroborating evidence only and is not treated as scientific support.

What we compare

Criteria are declared before the conclusion. They may include active strategy, evidence relevance, formula completeness, fragrance, cleansing and conditioning signals, manufacturer directions, cautions, represented pack and verification date. Formula analysis can describe what a declaration suggests; it cannot reproduce real-world hair feel or establish personal suitability.

Merchant destinations

Merchant destinations remain separate from product and formula identity. Editors manually check that a URL resolves to the represented product and record the date. We do not copy price, stock, seller, rating or review data. Affiliate and non-affiliate links are both representable and labelled; the current destinations are non-affiliate.

Reverification and corrections

Evidence types have different default intervals: manufacturer pages are normally rechecked within 180 days, retailer pages within 90, retailer pack images within 120, and identified physical packs or formal manufacturer/regulatory documents within 365. Staleness is a warning, not deletion. Editors run npm run data:report, npm run formula:freshness and targeted npm run formula:compare before publication and during maintenance.

Possible errors can be submitted through the formula corrections and evidence pathway. Include the exact product, market, pack, photographed or official declaration where lawful, source URL and observation date. Material corrections update the page date; distinct formulas create new version records.

What we do not do

We do not award ratings from packaging copy, fabricate use experience, extrapolate one person’s result to everyone, or copy a retailer score. A missing formula or verification date lowers confidence rather than inviting invention.

The noindex demonstration product exists only to exercise isolation safeguards. It is not sold, tested, recommended, searchable or linked to a merchant.