Performance
Sustained throughput under thermal throttling. We run Geekbench, 3DMark Wild Life, and a custom 30-minute video export loop, then chart the curve across the full battery drain.
Weight 25%
Brooklyn · The Benchmark Lab
—
Issue No. 142
The Benchmark Lab.
How 14 days, six rubrics, and a zero-paid-placements policy turn a gadget into a Gizazine scorecard — and why every chart we publish is reader-funded.
§ 01 — The Contract
Twelve standards. One promise. Zero paid placements.
Every scorecard Gizazine publishes is signed off against a public editorial ethics policy of twelve published standards. The shortest version: we buy every device we review at retail, we test it for at least fourteen days, we publish our methodology, and we accept no compensation from a manufacturer for coverage. The lab exists because reader funding bought it.
The full policy covers how we handle review units, how we disclose embargoes, how we treat first-look reporting, and what happens when a manufacturer sends a defective replacement. Standards eleven and twelve are the ones readers ask about most: every Gizazine scorecard is funded by subscriptions and memberships, not by a brand's media budget, and any gift, loan, or hospitality over fifty dollars in value is published in the article's footer within seventy-two hours.
You can read the entire policy as a single page — including the appeals process when a maker disputes a finding — at our editorial standards page. The lab's job is narrower: it produces the raw measurements, teardown photography, and benchmark logs that the policy asks for.
- 01Independent purchase. Every device is bought at retail by a Gizazine staffer or ordered through a publicly verifiable third party.
- 02Fourteen-day minimum. No device is scored before fourteen calendar days of daily use, including at least one weekend.
- 03Six-rubric scoring. Performance, build, battery, software, repairability, and long-term value — each scored 1–10 with a written justification.
- 04Public methodology. Test scripts, benchmark runs, and teardown photos are published with each review, not summarized.
- 05Zero paid placements. No manufacturer payment influences coverage. Reader subscriptions fund the lab.
- 06Disclosure under seventy-two hours. Every gift, loan, or hospitality over $50 USD is listed in the article footer.
- 07Embargo honesty. We disclose embargo terms on first-look reporting and never waive edit-control in exchange for access.
- 08Correction log. Errors are corrected in-line, dated, and never silently overwritten.
- 09Repairability before disposability. Every teardown documents the screws, adhesives, and connectors a future repair would meet.
- 10No astroturfing. Staff and contributors do not pose as customers in product forums or in retailer reviews.
- 11Reader-funded operations. Membership and newsletter revenue — not brand media budgets — pay for the lab's equipment.
- 12Open appeals. Makers may submit a written rebuttal; we publish it alongside the original review.
§ 02 — The Rubric
Six dimensions. One number per device.
A Gizazine score is not a vibe. It is the sum of six weighted measurements, each captured on the same hardware under the same conditions, then sanity-checked against the raw data we publish alongside the article.
Build & Materials
Drop, scratch, and ingress tests against the manufacturer's own claims. We use a calibrated drop rig and a Taber abrasion tester for every chassis claim.
Weight 15%Battery & Thermals
Real-world screen-on time at 200 nits, plus a thermal map under sustained load. We surface the gap between marketing "up to" hours and what we actually measured.
Weight 20%Software & Longevity
Update commitments, app sandboxing, default permissions, and the cost of living with the device for two years. We track every OS patch the vendor ships.
Weight 15%Repairability
Every device goes on the bench. Screw count, adhesive type, modular connectors, and parts availability from the manufacturer and the iFixit ecosystem.
Weight 15%Long-Term Value
A blended score weighting the cost of the device against expected useful life, repair cost, and resale floor. The number our editors argue about most.
Weight 10%
§ 03 — By the Numbers
What "reader-funded" looks like in practice.
80+
Devices torn down in the lab every year, with full teardown photography and screw-by-screw documentation published for each.
14
Minimum days of daily-use testing before a scorecard is signed. Weekends are non-negotiable; launch-day verdicts are forbidden.
7
Consecutive years of publishing the public methodology — every rubric, every weight, every script, every correction.
§ 04 — The 14-Day Test Cycle
What happens between unboxing and the score.
A device does not enter the Gizazine scorecard on launch day. It enters a fourteen-day cycle that begins the moment the courier scans it as delivered. The cycle is the same whether the device is a $39 pair of earbuds or a $3,499 workstation.
-
Day 1
Intake & serial logging
The courier scan is photographed. Serial number, MAC, IMEI, and firmware version are entered into the lab log. The device is assigned a unique test ID that follows it through the cycle and is published at the top of the eventual review.
-
Days 2–3
Independent purchase verification
The order is matched to a public receipt. Review units, loans, or PR samples are quarantined into a separate "loan review" track with a banner disclosure. The default track is the retail purchase track.
-
Days 4–11
Daily-use log & weekend stress
Two editors rotate the device through a written daily-use log: commute, desk, gym, kitchen, couch. The weekend window forces at least one non-workload stress case — travel, family use, accidental drops.
-
Day 12
Benchmark pass & thermal map
Geekbench, 3DMark, our custom export loop, and a sustained-load thermal map. Raw CSV files and FLIR imagery are attached to the article under a public data link, not hidden behind a paywall.
-
Day 13
Teardown & repairability audit
The device goes on the bench. Screws are counted, adhesive type is logged, modular connectors are photographed, and an iFixit-style repairability score is produced. The teardown gallery ships with the review.
-
Day 14
Scoring & editor sign-off
Two editors score independently against the six rubrics. Disputes go to a third editor. The composite score, weighted average, and dissenting notes are all published — including when the team disagrees.
§ 05 — Recent Teardowns
All teardowns →
Off the bench this quarter.
Inside the Aether MX-7: a driver you can actually replace.
A single Torx T6 holds the entire driver assembly. The battery is soldered but reasonably accessible. Repairability score: 7.4 / 10.
The Hub Mini 3 hides its radios under a copper lid.
Three radios, one antenna trace, and a thermal pad we suspect is doing real work. The lid is glued, not clipped.
The Keyfarer K2 is finally a keyboard you can service.
Hot-swap sockets across the entire board, a separate daughterboard for the USB-C, and a foam mod that doesn't void the warranty.
The Lumix S9 is a beautiful mess of adhesives.
Eleven screws, seven adhesive strips, and a sensor mount that requires a proprietary spanner. The teardown is also a cautionary tale.
The lab publishes every teardown as a public gallery. Browse the full archive, or read the methodology behind how we score what we find.