The Audit Ledger

We Calculate the True Forensic Score of Consumer Hardware

Affiliate sites sell you things. We audit the claims. Our scoring methodology reverse-engineers synthetic benchmarks, strips out review-farm bias, and normalizes hardware performance against historical pricing curves. No "seamlessly unlocking your potential." Just raw data, tear-downs, and forensic analysis.

View The Audit Index
Torn down laptop motherboard showing forensic details
74% of 'Best Mattress' guides are owned by 3 holding companies     //     Avg GPU MSRP inflation: +41% since 2019     //     Appliance lifespan drops 1.4 years per decade     //     Fake Review Network Takedowns: 1,492 this quarter

The Forensic Scoring Engine

A score is only as good as the math behind it. We don't use 'expert opinions' as a crutch. We use normalized data.

1. Affiliate Stripping

We mathematically penalize sites that have a 1:1 ratio of positive reviews to affiliate outbound links. If they never review a bad product, their good reviews are void.

2. Benchmark Normalization

Manufacturers cherry-pick benchmarks. We run 14-point regression analysis on synthetic loads to find thermal throttling points the press materials ignore.

3. Price-to-Performance Delta

A good product at a bad price is a bad product. We track MSRP vs street price over 36 months to calculate the actual depreciation curve.

4. Warranty & Repairability

Glued-in batteries drop the final score by a flat 20%. If iFixit gives it a 3/10, we mathematically cap the product's maximum achievable forensic score.

Recent Audits

ProductCategoryStated ClaimForensic ScoreVariance
Apple M3 Max 16-inchLaptops18hr Battery8.4 / 10-4%
LG C3 OLED 65"Displays800 nits peak9.1 / 10+2%
Dyson V15 DetectAppliances60min runtime6.2 / 10-24%
Sony WH-1000XM5Audio30hr ANC7.8 / 10-9%
Nvidia RTX 4080 SuperComponents4K 120fps8.0 / 10-6%
"A good product at a bad price is a bad product. MSRP is a fiction created for launch-day reviews."
— The Editorial Board

The 5 Red Flags of a Fake Review Farm

The internet is a dead mall of SEO spam. Here is exactly how to spot a site that has never actually touched the product they are reviewing.

Server rack representing review farms
  • 01. The 'Pros' list perfectly matches the manufacturer's spec sheet.
  • 02. The 'Cons' are minor inconveniences ('Only comes in 2 colors').
  • 03. There is no original photography, only press-kit renders.
  • 04. The review was published on the exact minute the embargo lifted, but claims '3 months of testing'.
  • 05. Every single alternative suggested in the 'Also Consider' section costs more and has a higher affiliate commission.

Review Authenticity Scorer

Paste a URL to a review. We'll run heuristic analysis to determine if the author actually used the product.

Open Tool ->

The Mattress Cartel

If you search for 'Best Mattress 2024', the top 10 results are likely owned by the same holding companies that own the mattress brands themselves. This is vertical integration masking as independent journalism.

We audited 4,000 mattress reviews across 50 sites. 92% of 'Top Picks' coincided with brands that offered a CPA (Cost Per Acquisition) over $150. We rebuilt the ranking based solely on fiberglass content, foam density (lb/cu ft), and warranty non-void clauses.

14,092
Products Audited
8.2 yrs
Avg Warranty Claim Time
-22%
Performance vs Marketing Claim
4.1
Avg iFixit Repair Score

How We Test (The Reality Check)

Marketing departments test in a vacuum. We test in a 72-degree room with dust, background processes running, and realistic user error.

1.Procurement

We buy 80% of our test units at retail. Review samples must come with a signed contract waiving embargoes.

2.Tear-down

Before turning a device on, we check for glued batteries, proprietary screws, and thermal paste application.

3.Regression Testing

We ignore manufacturer-provided benchmark guides and run synthetic loads designed to force thermal throttling.

What we mathematically penalize

  • + 1:1 ratio of positive reviews to affiliate outbound links.
  • + Glued-in batteries and proprietary components.
  • + Cloud-dependent IoT devices without local control.
  • + MSRPs that drop more than 20% within 90 days.

Methodology FAQ

Do you accept review units?

Yes, but with a signed contract stating we will not honor embargoes if a critical flaw is found, and we reserve the right to destroy the unit during testing.

How do you calculate the 'Affiliate Bias Penalty'?

We crawl the reviewer's historical data. If their average score is above 8.5/10 across 100 products, their scores are mathematically deflated in our aggregate index by a factor of 1.2.

Why is the Forensic Score almost always lower?

Because marketing departments test in a vacuum. We test in a 72-degree room with dust, background processes running, and realistic user error.

The Raw Data

Get our monthly dump of forensic audit scores and price-history datasets in CSV format. No marketing, just data.

Subscribe via RSS