Every report says how it was made. When we haven't used a product ourselves, we say so in the first paragraph — and show you the owner data instead.

How We Review

GroundedPick helps US shoppers make smarter decisions. Here is a plain-English account of how every rating, coupon result, and article is produced — including what we test ourselves and what we research from aggregated data.

📊
Data-first
Every store rating aggregates hundreds of verified owner reviews, not editorial opinion alone.
🔬
Tested codes
Editors place a real cart before publishing any coupon code. Expired codes are removed within 48 hours.
✏️
Human-reviewed
AI-assisted drafts are fact-checked by a named editor before publication. The editor's name appears on every page.

What each review badge means

Every article on GroundedPick carries one of three badges so you always know how the information was gathered.

🔬 Tested by our team

We purchased the product and used it hands-on. The article includes original photos and first-hand performance data.

📊 Research-based

We analyzed 200+ verified owner reviews, manufacturer specs, and long-term reliability data. We did not use the product ourselves — this is stated clearly in the article.

⚡ Quick pick

A fast recommendation based on fewer than 200 reviews and publicly available specs. Best used as a starting point, not a final verdict.

Why we're upfront about this Most of our articles are Research-based. Rather than pretend we've used every product we cover, we invest that effort in aggregating real owner data at scale — surfacing patterns across hundreds of long-term reports that a single hands-on reviewer cannot replicate. As our team grows, more articles will be upgraded to Tested status.

How research-based articles are produced

The majority of GroundedPick articles carry the Research-based badge. This is what that process looks like from start to finish.

1

Review aggregation

We collect 200–500+ verified buyer reviews from Amazon, specialty retailers, and community forums (Reddit, brand-specific communities). Only verified-purchase reviews are counted at full weight.

2

Pattern analysis

We identify recurring praise and failure patterns across the review set, focusing on reports from owners who have used the product for 3 months or longer to surface durability data.

3

Spec verification

Manufacturer claims are cross-checked against independent lab reports, third-party expert sources, and community stress-tests where available.

4

Price & value tracking

We track price history over time using third-party price tools, so value assessments reflect real market conditions — not just the launch price.

5

AI-assisted drafting

A large-language model drafts a structured outline from the aggregated data. It is instructed not to fabricate facts, citations, or test results.

6

Editor fact-check & publish

A named editor verifies every claim against source data, adds original analysis, applies the research disclosure, and publishes with a byline and last-updated date.

What this process does and does not include Research-based articles do not include hands-on use by our team. We do not handle, wear, or operate the products we cover in these articles. What we provide instead is a structured analysis of real owner experience at scale — including long-term failure rates, common defects, and value trends that a single short-term test cannot capture. Every Research-based article states this clearly near the top.

How store ratings work

Each store page aggregates owner reviews from multiple verified sources. We collect review text, star ratings, verified-purchase flags, and submission dates, then apply a recency-weighted algorithm to produce the Reader Rating shown on each card.

1

Data collection

We pull owner reviews from multiple third-party sources and our own reader-submission form.

2

Verification weighting

Reviews with a verified purchase or receipt are weighted at 1.4×. Unverified anonymous reviews count at 0.8×.

3

Recency weighting

Reviews from the past 6 months count fully. Older reviews decay linearly to 50% weight at 36 months.

4

Moderation & publish

A final moderation pass removes off-topic, spam, or policy-violating content before the rating appears publicly.

How coupon codes are tested

Every active coupon code on this site was tested by a human editor who placed a real cart on the retailer's checkout page. We do not publish codes we cannot verify.

  • Codes are tested on a real browser, not a headless bot, so we catch CAPTCHA-gated checkouts and region restrictions.
  • The editor records the cart subtotal, discount applied, and final price.
  • The "Last verified" timestamp reflects when the most recent live test passed.
  • Expired or failing codes are removed within 48 hours of a failed sweep.
  • Codes that only work for new customers are labeled "new customers only."

Legitimacy signals

Our store legitimacy score combines positive and negative signals from multiple data sources.

Signal Direction Source
Verified customer reviews (6 mo.)PositiveOwner reports + our data
Active return policy pagePositiveSite crawl
SSL certificate (HTTPS)PositiveTLS check
Physical address on contact pagePositiveSite crawl
BBB accreditationPositiveBBB API
WHOIS domain age < 6 monthsNegativeWHOIS lookup
Chargeback rate spike (reported)NegativeCommunity reports
Unresolved BBB complaints (12 mo.)NegativeBBB API
Multiple confirmed scam reportsNegativeCommunity reports

Scoring criteria

Product and store reviews use a 1–5 editor rating composed of weighted criteria. Criteria vary by category; the general framework is:

CriterionWeightHow it is measured
Value for money25%Price vs. comparable alternatives, tracked over time
Quality / reliability25%Owner-reported defect rates and long-term durability patterns
Customer experience20%Aggregated service ratings, return policy clarity, support responsiveness
Feature completeness20%Checklist of category-specific must-have features
Trustworthiness10%Legitimacy signals (see above), BBB score, WHOIS domain age

The final editor rating is rounded to one decimal place. A "Not Recommended" verdict is applied when any criterion scores below 1.5, regardless of the weighted total.

Corrections policy

We correct errors promptly and transparently:

  • Minor corrections (typos, outdated prices) are made silently and the last-updated date is refreshed.
  • Substantive corrections (wrong rating, incorrect product claim) appear in the article's change log with the date and a description of what changed.
  • If a correction materially changes our recommendation, the lead paragraph is updated to reflect the current verdict.

To report an error, email our corrections team.

Advertising disclosure

Some links on GroundedPick are affiliate links. When you click an affiliate link and make a purchase, we may earn a small commission at no cost to you. This commission does not influence our ratings or recommendations.

Sponsored placements (stores or products that pay for featured placement) are labeled "Sponsored" or "Featured" and are never mixed into editorial rankings without disclosure.

Our editorial independence guarantee Advertisers cannot purchase positive reviews, altered star ratings, or removal of negative information. The editorial team makes all rating and recommendation decisions independently of the commercial team.