How we test, score & rank every product

For every search term, we pull a live snapshot of qualifying Amazon listings: title, brand, ASIN, list price, current price, discount percentage, condition, Prime / free-shipping eligibility, bullet features, primary image, and other-seller offers. Data is sourced from public Amazon catalog endpoints and refreshed continuously.

• Source of truth: Amazon.com product detail pages.
• Refresh window: ≤ 1 hour for price, stock and shipping; ≤ 24 hours for catalog metadata (title, features, image).
• Timestamp: every ranking page renders the exact "as of" time the listing data was captured.
• No private databases: we do not buy or licence third-party product feeds that could go stale between updates.

Before scoring, every candidate must pass an automated qualification filter. A listing is excluded — silently — if any of the following is true:

• It is a digital good, gift card, subscription, or service.
• It is currently out of stock or unavailable to ship.
• It has no primary image, or the image fails to load.
• It has fewer than two descriptive bullet features.
• Its condition is "Used – Acceptable" or lower (refurbished and "Used – Like New" are allowed, but labelled).
• The seller is unrated, suspended, or carries a known counterfeit flag.

Filtering happens before scoring so that the visible top-10 is always drawn from a clean pool — never padded with weak listings just to fill ten slots.

We use a position-based composite score on a fixed 1–10 scale. Each rank maps to a published score value so readers can compare picks across categories on equal footing:

1. 9.7
2. 9.5
3. 9.4
4. 9.2
5. 9.0
6. 8.8
7. 8.6
8. 8.3
9. 8.1
10. 8.0

A product's rank is determined by combining four weighted signals:

• Amazon relevance signal (≈ 55%) — the platform's own ordering for the qualifying search, which already encodes millions of real customer reviews, purchases and returns.
• Discount depth (≈ 20%) — the savings percentage compared to the listed reference price.
• Feature completeness (≈ 15%) — how thoroughly the listing describes specs that matter in this category.
• Fulfilment quality (≈ 10%) — Prime eligibility, free shipping, and condition.

We do not invent star ratings, review counts, or "lab scores" that we did not measure. If Amazon does not surface a data point in the listing we receive, we leave it off the card.

• Editor's Choice — automatically assigned to the #1 ranked product on every list. One per page; never moved by an advertiser or affiliate manager.
• Best Value — the highest-scoring product (other than #1) whose discount is greater than 10% off list price. If no qualifying product meets the threshold, the badge is suppressed rather than awarded to a weaker candidate.
• Best Budget — the lowest-priced product within the top ten, provided it is not already labelled Editor's Choice or Best Value. No duplicate labels are ever shown on the same product.
• Premium pick — applied when a product sits in the top ~15% of price-range for its category and offers materially more features than the median pick.

We want to be honest about scope. PickRanker is an automated aggregator with editorial oversight, not a long-form consumer-lab review site like Wirecutter or Consumer Reports. That means:

• We do not claim to have stress-tested every product in our own lab. When we have hands-on experience with a pick, the byline says so explicitly.
• We do verify, for every list: data freshness, seller legitimacy, image accuracy, feature plausibility, and duplicate-listing detection.
• We do cite the underlying sources — Amazon listing, manufacturer spec sheet, and reputable third-party reviews — in the "Sources cited" panel at the bottom of every product page.

This separation lets us cover thousands of categories with current prices, while making it transparent which claims come from first-hand testing and which from the underlying retailer.

PickRanker earns commissions through the Amazon Associates program when readers click an affiliate link and complete a purchase — at no extra cost to the buyer. That is the only way the site is funded.

• No brand, manufacturer, or third party can pay to appear, move up, or be removed from any ranking.
• Our writers and editors are paid a flat fee per article. Their compensation is never tied to affiliate revenue from a specific product.
• Sponsored or paid content — if we ever publish any — will always be labelled "Sponsored" at the top of the page and excluded from ranking lists.

Amazon's Associates Operating Agreement requires that any price we display is current and accurate at the moment it is shown. Because scraped prices drift quickly, we default to a price tier symbol rather than a hard number:

• $   under $15
• $$   $15 – $49
• $$$   $50 – $149
• $$$$   $150 – $499
• $$$$$   $500+

Readers can opt in to numeric prices via the price toggle in the header (and opt out of price display inside our deal boxes separately). Either way, the final, binding price is whatever Amazon shows on the product page when the reader clicks through.

• We never accept payment for placement, inclusion, or ranking.
• We never write fake reviews, fabricate specs, or invent star ratings.
• We never claim "lowest price," "limited stock," or "free shipping" without the live listing data to back it.
• We never cache prices for more than one hour — Amazon's terms forbid it, and so do ours.
• We never republish copyrighted product imagery; all images are loaded directly from Amazon's CDN.
• We never collect personal data from readers beyond what is needed for opt-in price alerts.

• Live data: price, stock, shipping — checked on every page load past the cache window.
• Ranking re-runs: daily for high-traffic categories, weekly for the long tail.
• Editorial review: a human editor revisits the top-50 categories at least once per quarter to spot drift, audit labels, and refresh the intro copy.
• "Last updated" badge at the top of every product page reflects the most recent of these three.

Listings change constantly. If you spot a stale price, broken image, mis-categorised product, or factual error, please refresh the page first — most issues resolve once we re-fetch from Amazon. If the problem persists, contact the editorial team via the About page. Verified corrections are applied within 24 hours, and any material edit is logged in the page's revision history.

Brand or rights-holder takedown requests are honoured promptly. Email us with proof of ownership and the URL in question and we will remove or amend within one business day.