Full public methodology
How we test &
review products
Every scoring criterion, every editorial rule, and every disclosure requirement — published in full. If a review is on this site, this document governs how it was made.
Document version
v2.4
Last updated May 2026
Current
Scoring framework
Every store is scored across 5 weighted dimensions on a 0–10 scale. The final score is a weighted average displayed to one decimal place. No single criterion can override the others.
01
Product Quality
30%
Signals evaluated
Accuracy of product descriptions versus items actually received (AI-analysed against verified buyer reports)
Material and durability assessments from long-term verified owners (90+ days post-purchase)
Consistency across product lines — not just flagship items
Authenticity signals: counterfeit complaints, grey-market sourcing flags
For hands-on reviews: editor purchases and tests the product independently for a minimum of 14 days before scoring this dimension. See §3 Hands-on vs research-only.
Customer Service
25%
Signals evaluated
Median first-response time across live chat, email, and phone — measured by test contacts
Issue resolution rate: percentage of complaints resulting in satisfactory resolution (AI-measured across reviewed cases)
Escalation handling: how the store responds when initial contact fails
Verified shopper NPS signal from community reviews
Our editors conduct real test contacts — submitting a support ticket or live chat — before assigning a final customer service score. Results are logged and date-stamped.
Shipping & Fulfilment
20%
Signals evaluated
Promised vs actual delivery window accuracy (AI-measured from verified delivery reports)
Packaging quality and item condition on arrival
Carrier tracking reliability and proactive communication on delays
International shipping transparency where applicable
Return Policy & Fairness
15%
Signals evaluated
Return window length and stated policy clarity
Ease of actual return process (verified through buyer reports and test purchases)
Refund turnaround time from item receipt to credit issued
Hidden exclusions, restocking fees, or return-shipping cost traps
Pricing & Value
10%
Signals evaluated
Price competitiveness versus comparable products from 3+ other stores
Hidden fees: processing, mandatory upsells, surprise charges at checkout
Sale price accuracy — we verify whether "was" prices are genuine historical prices
Pricing is weighted lowest because "cheap" does not mean "trustworthy." A low-priced store with poor product quality and no return policy scores poorly overall.
3-phase review process
Every store review — regardless of size or affiliate status — passes through all three phases in sequence. No phase may be skipped. The phase each review reached is noted in the review footer.
02
1
AI Phase
Data Collection & AI Pattern Analysis
Our AI engine ingests verified shopper reviews, return-policy documents, pricing history (90-day lookback), and customer-service response-time data for the store. It flags statistical anomalies — sudden spikes in negative sentiment, policy changes, pricing inconsistencies, and claim-versus-reality gaps — that a human reviewer would take weeks to surface manually.
The AI does not write any review content. It produces a structured research briefing delivered to the assigned human editor.
The AI does not write any review content. It produces a structured research briefing delivered to the assigned human editor.
Research briefing doc
Anomaly flag list
Sentiment trend chart
Raw data log (internal)
2
Human Phase
Expert Review & Independent Verification
A named category specialist receives the AI briefing and independently verifies the key claims. This includes: contacting customer support as a real shopper, cross-checking pricing claims against archived pages, purchasing (or requesting test units of) products where the review type is "hands-on," and challenging any AI flags that don't hold up to scrutiny.
The human editor can raise or lower any criterion score — and must document the reason in an internal editor log. AI scores are starting points, not final verdicts.
The human editor can raise or lower any criterion score — and must document the reason in an internal editor log. AI scores are starting points, not final verdicts.
Editor notes log
Test purchase receipts (where applicable)
Support test transcript
3
Sign-off Phase
Editorial Sign-off & Conflict of Interest Check
The editorial lead reviews the draft for accuracy, bias, AI-label compliance, and affiliate-link disclosure before publication. At this stage, scores are locked — meaning they are committed to the database and cannot be changed without a new editorial review cycle. Affiliate links, if applicable, are placed only after this lock.
The editorial lead also checks: does the reviewing editor have any financial or personal relationship with this store? If yes, the review is reassigned.
The editorial lead also checks: does the reviewing editor have any financial or personal relationship with this store? If yes, the review is reassigned.
Score lock timestamp
Disclosure sign-off
Conflict check log
The phase each review reached is listed in the review footer alongside the editor's name and the lock date. A review currently in Phase 2 is labeled "In progress — research phase" and is not publicly scored until Phase 3 is complete.
Hands-on vs research-only — disclosure
Not every review involves a physical product test. We label this clearly on every review page so readers know exactly what level of verification was performed.
03
Hands-On Review
The editor independently purchased and tested the product or placed a real order with the store. Physical inspection, performance testing, and real customer-service contact were completed before scoring.
- Product purchased independently (receipt on file)
- Minimum 14-day use period for product reviews
- Real support contact completed and logged
- Return process tested where feasible
- Editor bio and purchase declaration on file
Research-Only Review
No physical purchase was made. The review is based on AI pattern analysis of thousands of verified shopper reviews, policy documents, pricing data, and secondary research verified by the human editor.
- AI analysis of 500+ verified buyer reviews minimum
- Policy documents reviewed and cross-referenced
- Pricing data verified via historical archive
- Editor verifies AI flags via secondary sources
- "Research-Only" label prominently displayed
Important: Research-only reviews carry a clear label on the review page. We never present a research-only review as if physical testing occurred. If you see a review that misrepresents this, please submit a correction.
| Review category | Default type | What the editor does |
|---|---|---|
| Electronics & gadgets | Hands-on | Purchases device, tests for 14+ days, runs benchmarks or durability assessments |
| Beauty & personal care | Hands-on | Uses product for 30+ days; cosmetic chemistry editor checks formulation claims |
| Apparel & fashion stores | Both | Test purchase made; fit, material, and return process verified in-person |
| Home & kitchen | Hands-on | Appliances tested for function and durability; large items may use verified owner panels |
| Digital / software stores | Research-only | AI analysis of verified user reviews; pricing and refund policy verified manually |
| Large furniture / outdoor | Research-only | Research-based, using long-term verified owner reports (6+ months post-purchase) |
| Marketplace / multi-seller | Both | Platform review is research-only; individual seller spotlights are hands-on |
AI-assisted analysis
How artificial intelligence is used in our research pipeline — and the hard limits we place on it.
04
Verified review pattern analysis
Our AI reads thousands of verified buyer reviews per store — detecting sentiment trends, flagging claim-versus-reality gaps, and identifying recurring complaints that surface only in aggregate. The model is trained to ignore incentivised or unverified reviews.
Policy change monitoring
AI monitors return policies, shipping terms, and pricing pages for changes between review cycles. When a material change is detected, a review is automatically flagged for priority human re-evaluation — before the 90-day cycle would normally trigger it.
Pricing history analysis
For every store with a pricing dimension, AI tracks 90-day price history to verify whether advertised "sale" prices are genuine discounts or inflated "was" prices. This data feeds directly into the Pricing & Value dimension score.
Research briefing generation
AI compiles all of the above into a structured briefing document delivered to the assigned human editor. This document contains flagged anomalies, supporting data, and a preliminary dimension breakdown — never a finished review.
Human override is always authoritative
The human editor may override any AI-flagged signal or preliminary score dimension. When they do, they must document the reason in the internal editor log. This log is retained and available for internal audit. AI findings that a human editor cannot independently verify are excluded from the final score — not averaged in.
Every section of a published review that was AI-assisted carries a visible AI badge. Sections written entirely by the human editor have no badge. There is no unmarked AI content on this site.
Editorial independence policy
Non-negotiable rules governing every review. These apply regardless of affiliate status, advertiser relationships, or editorial convenience.
05
Editorial Independence Pledge — v2.4
- No store, brand, or individual can pay to be reviewed, ranked higher, or reviewed faster
- Affiliate links are placed only after editorial scores are locked and the review is published
- Products received for free are disclosed; free receipt confers no score benefit
- Every published review carries a named editor, their credentials, and the lock date
- Editors with any financial or personal relationship with a reviewed store are recused
- Score changes require a new editorial review cycle — not an internal edit
- Advertisers cannot influence editorial coverage or suppress a negative review
- Reader corrections receive a response within 5 business days; confirmed corrections are published
FTC & affiliate disclosure
Required disclosures under FTC 16 CFR Part 255 and Amazon Associates Operating Agreement. All disclosures appear on every relevant page — not only in a single sitewide policy document.
06
Amazon Associates — Required Statement
Trusted Buyer Report is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com. As an Amazon Associate, Trusted Buyer Report earns from qualifying purchases. This does not affect the price you pay. Affiliate commissions support our editorial team but are never applied before scores are locked and published.
Page-level disclosure
Every page containing an affiliate link carries a clear, prominent disclosure above the fold — not only in the footer or a separate policy page. This satisfies FTC requirement for clear and conspicuous placement.
Product-received disclosure
If a product was received free or at a discount for review, this is stated on the review page with a disclosure box near the top of the content — not buried in a footnote.
AI-content disclosure
Every AI-assisted content section is labeled with an AI badge visible inline. Readers are never left to assume human authorship of AI-generated material.
Score-before-link rule
Affiliate links are added to a page only after the editorial score is locked in our database. No score may be modified after the affiliate link is live without a new review cycle.
- Affiliate relationships disclosed on every page that contains affiliate links — not only in a privacy policy or terms page
- "As an Amazon Associate I earn from qualifying purchases" — present on all Amazon affiliate pages
- Material connections (free products, press passes, brand trips) disclosed on the relevant review — not just on an About page
- AI-generated or AI-assisted content labeled at point of use — not only in a site-wide AI ethics document
- All disclosure language reviewed against FTC 16 CFR Part 255 guidelines
- Disclosures are in plain English — not legal boilerplate obscured behind links
For the complete legal affiliate disclosure, see our full affiliate disclosure page →
Update & freshness policy
A review that was accurate 18 months ago may be dangerously out of date today. Our update schedule ensures ratings always reflect current store reality.
07
Every active store listing is re-reviewed on a scheduled 90-day cycle. "Active" means any store with a public score on this site. Archived stores that have closed or gone dormant are clearly labeled as archived and excluded from the update cycle.
Scheduled cycle
All active store reviews are flagged for re-review after 90 days. The next review date is displayed on every store page. A review remains marked "current" until that date passes.
90
Day max age
4×
Per year, per store
Urgent out-of-cycle
Any verified report of a significant store change (ownership, return policy, quality drop, pricing deception) triggers an immediate priority review — regardless of when the last update was.
48h
Flag-to-review SLA
5
Business days to publish
Community signals also feed into our update triggers. A statistically significant drop in verified community review sentiment — measured against the store's 90-day baseline — automatically generates an urgent review flag in our editorial queue.
The "Last reviewed" date shown on every store page reflects the most recent Phase 3 sign-off date. If a score is being updated, the page displays a "Review in progress" banner until the new score is locked.
Corrections policy
Factual accuracy matters. We respond to every correction submission and publish confirmed corrections transparently in-article and in the public corrections register.
08
Response SLA
Every correction submission receives an acknowledgement within 2 business days and a decision — confirmed, declined (with reason), or under investigation — within 5 business days.
2
Days — acknowledge
5
Days — decision
Public register
All confirmed corrections are published in our public corrections register with: the original claim, the corrected claim, the date of correction, and the editor who made the change.
- Corrections are noted transparently at the bottom of the corrected article — not silently overwritten
- Score corrections that result from a confirmed factual error require a full Phase 2 + Phase 3 re-evaluation before a new score is published
- Declined corrections receive a written explanation of why the original content stands
- If a correction reveals a conflict of interest, the review is reassigned and re-evaluated from Phase 1
Found something wrong?
We take every correction seriously. Please include the URL of the article and the specific claim you believe is incorrect.
Submit a correction
Smarter shopping, weekly
New store reviews, buying guides, and research-backed insights delivered to your inbox. No spam, no fluff.