Most comparison sites tell you what they picked. This page tells you how, in enough detail that you can disagree with us.

Every product on BuyeReviews goes through the same process: we decide what matters for a category before looking at any products, check claims against independent sources, score the results, and audit our own scores for bias. What follows is each of those steps, including the parts where the honest answer is that we do not know.

The two numbers on a page

The Buyer Score measures how a product is working out in the real world for people who already own it, and it determines the order in which products appear. We build it from the signals buyers leave behind: ratings, written reviews, sales volume, and price relative to the category. Together, these answer one vital question: is this item living up to its promises in actual homes? What it cannot tell you is whether a product is safe, whether its claims are true, or what it costs to produce and dispose of. (People rarely rate a supplement on whether its dose is backed by human clinical trials.)

The Integrity Score (or Standards Score) reflects what we found when we evaluated the product against the graded topics for its category. It is a weighted composite rather than a simple average, because the criteria that most reliably separate a verified product from a claims-only one carry double weight.

Step one: We decide what matters before looking at products

Each category page starts with eight to twelve graded topics, written specifically for that product type and fixed before a single product is scored. Setting the criteria in advance is the point: it is much harder to bend a standard when you wrote it before knowing who wins.

Four core topics appear in every category set:

• Health & safety impact: What the product exposes you to in normal use, at the actual dose or specification present.
• Environmental impact: What it costs to manufacture, operate, and dispose of.
• Claim honesty: Whether what is printed on the label matches what is in the box.
• Independent verification: Whether a third party outside the manufacturer has verified the claims.

The remaining topics are category-specific and focus on real-world failure modes or hazards rather than marketing feature lists. For a humidifier, that means ease of cleaning—because the real failure mode is mold and bacteria growth inside it. For an air purifier, it means whether it emits ozone. For a supplement, it means contaminant testing, because the risk lies in the impurities alongside the ingredient rather than the ingredient itself.

Every topic belongs to one of three pillars—Health, Ecology, or Evidence—which is how they are presented in the Health and Ecology section of each page. The one or two topics per category that most cleanly separate a verified product from a claims-only one are marked as priority topics and count double in the Integrity Score.

Once established, the category standard does not change. Every product on that page is measured against the same yardstick, including products added months later. If a new product genuinely requires a topic the page lacks, we regrade the entire category page rather than evaluate one product by a different standard.

Step two: We verify before we score

We do not grade from memory, nor do we grade from marketing copy. Before scoring a product, we verify:

• Its true composition or specification using manufacturer documentation or primary retailer listings—never secondary summaries.
• Every claimed certification against the official registry of the issuing body. (Certifications lapse, and a logo printed on a box often outlives the certificate behind it.)
• Whether independent testing exists for the product or its specific category.
• Recalls, regulatory enforcement, warning letters, and California Proposition 65 notices.
• The actual manufacturer when branding is opaque, including whether the identical item is sold under alternate names.
• Whether claimed health effects are backed by human evidence at the actual dose present, rather than at doses used in loosely related studies.

Every fact that moves a score has a primary source recorded alongside it.

Step three: How scores are assigned

Scores run from 0 to 100 on every topic, structured into three distinct tiers:

• High: Confirmed by an independent source outside the manufacturer.
• Medium: Plausible and consistent with category norms, but not independently verified.
• Low: Requires positive evidence of a problem—a self-contradicting listing, a Prop 65 warning, an active recall, or a claim that has been tested and refuted.

The Low tier is critical: a lack of public information is not evidence of bad performance, so an unknown product is graded as unverified rather than defective. This prevents small brands from being penalized for being small and prevents familiar brands from being rewarded simply for being recognizable.

A claim resting solely on the manufacturer's word cannot enter the top band, no matter how reasonable it sounds. Confidence is not evidence.

Step four: Managing estimates and missing data

Manufacturers do not publish everything. When a necessary figure is missing, we estimate it, clearly mark the estimate on the page, and cite the baseline used:

• Peer comparison: Other products on the same page disclose the metric, establishing a clear range.
• Category benchmark: Independent testing has established an industry-wide standard range for this product type.
• Engineering limits: Physics or device-class mechanics constrain the result regardless of brand claims.
• Manufacturer claim: Based solely on the maker's statement, and scored accordingly lower in confidence.

Every product displays a Confidence Level based on how much data had to be estimated and whether missing details fall on high-priority topics. Where confidence is moderate or low, we state this plainly at the top of the score section and identify what is unknown. We never present a product with insufficient verification as a confident recommendation.

Step five: Internal score audits

A score must hold up even if the brand name is anonymized. If a product would grade differently purely because of who makes it—while keeping the evidence identical—the score is flawed and is corrected before publication. For every topic, we compare the highest and lowest scoring products to verify that the gap reflects actual differences in evidence rather than brand reputation.

Popularity, price, review volume, and product placement never affect the Integrity Score. While those are useful market signals, they belong exclusively in the Buyer Score. They cannot tell you what is inside the box, so they are kept strictly out of the verification metrics.

Products added to existing pages are audited directly against current entries. Scoring products in isolation causes ratings to drift over time, and drifted ratings make comparison tables useless.

What we will not list

Certain products are removed from our pages entirely, regardless of sales volume: items subject to active recalls, products facing regulatory enforcement over safety or false claims, items with documented hazards in normal use, and products containing banned ingredients.

When an entire product category carries risks for specific groups, we flag it clearly. Category pages outline who should avoid the product type entirely, which medical conditions or medications it may interfere with, and which home environments it suits poorly. These same warnings appear on individual product listings under Who this product is not for.

What our ratings cannot tell you

• We are not a testing laboratory: We do not conduct internal lab tests. We cross-verify manufacturer claims against existing scientific and regulatory evidence, and we state clearly where that evidence ends.
• Scores compare products within a single category, not across categories: An Integrity Score is calibrated against peers on that specific page. Matching numbers across two different categories do not imply equal performance.
• Evidence evolves: Formulations change, certifications expire, and new research is published. A score reflects what was verifiable at the time of our audit.
• A high Buyer Score is not a safety guarantee: Many products that consumers love have never undergone independent safety or claims verification. This is why we keep user satisfaction and technical integrity completely separate.

Keeping pages current

We review active pages routinely and regrade products whenever material changes occur: specification updates, altered formulations, lapsed certifications, new recalls, or fresh independent research. Discontinued items are removed.

When we get it wrong

If you spot an error or have access to newly published documentation from a manufacturer, please let us know. We would much rather correct a mistake than leave inaccurate data on the page.