Methodology

1. Scope

This methodology applies to:

• ERD/ARD product reviews (/vendors/[brand]/[model]/)
• Brand comparisons (/compare/[a]-vs-[b]/)
• Buying guide recommendations and “Editor's Pick” designations

It does not apply to news, opinion, sponsored content, or general technical explainers.

2. Scoring framework

We score each ERD/ARD across seven weighted criteria. Final score is a weighted average on a 100-point scale; products under 60 are not “Recommended.”

3. Data sources

In order of priority:

1. First-hand testing at a third-party lab, where feasible, against the test protocol below.
2. Manufacturer datasheets with specifications cross-checked against installed units in the field.
3. Published certification reports from BIS, TÜV, UL, or equivalent accredited bodies.
4. Field interviews with elevator OEMs, installers, AMC providers, and facility engineers — minimum three independent sources for any “verified” claim.
5. Peer-reviewed literature for theoretical claims about topology, chemistry, or control architecture.

Vendor marketing material is not a data source. Brochures, sales decks, and press releases inform our questions; they do not establish facts.

4. Test protocol

For physical tests we perform, at minimum:

• Switchover time measurement under abrupt grid loss, captured on a digital oscilloscope at the inverter output bus.
• Sine-wave THD measurement under full rated load and at 30% load.
• Cabin-light continuity test using a calibrated lux meter during the transition window.
• Sustained backup runtime on rated load, run to controlled cut-off.
• Inrush behavior test on lift-motor start-up while running on backup.
• Thermal soak test at the upper end of the manufacturer's stated ambient range.

Tested units are purchased at retail or supplied with documented chain-of-custody. We do not test units provided directly by vendors as “review samples” without independent verification that the sample matches retail stock — and any such tests are flagged in the review header.

5. Comparison pages

Head-to-head comparisons (/compare/[product-a]-vs-[product-b]/) use the same seven-criterion framework and the same per-product scores. Sort order on listing pages is alphabetical or score-descending — never paid placement, never weighted by affiliate revenue.

Where a comparison includes a product we have not yet tested first-hand, the page header carries a “Data source: vendor datasheet, awaiting field verification” tag and the product's score is bracketed.

6. Re-scoring and updates

Manufacturer firmware updates, model refreshes, or material specification changes trigger a re-score. The page header always shows: review date, methodology version used, and next scheduled review date. Re-scores that change a product's “Recommended” status are announced in the changelog at the top of the affected page.

7. Versioning

This methodology is versioned. Material changes (criterion weights, new criteria, removed criteria, test-protocol changes) are published with a public changelog. Old reviews continue to display the methodology version under which they were scored, with a link to the re-evaluation status.

8. Limits of our method

• We do not pretend to have tested every claim in this document on every product. Where data is from the vendor's datasheet or a third-party report rather than our own bench, the review page says so.
• Field-reliability scores depend on long-tail data; they are reviewed at 12-month intervals and may shift materially as more installations age.
• New products may be reviewed against datasheet specifications first, with a “data source: vendor datasheet, awaiting field verification” tag, until first-hand data is available.
• Total Cost of Ownership figures use a 10-year horizon and assumptions documented per review; we publish the model so you can substitute your own numbers.