How We Test & Score Vacuums
Every figure on this site carries its source. This page is the contract behind those labels: what each tier means, how scores are built today from specs and cited independent tests, and the protocol a BVC measurement will follow once we test units in hand.
01 Provenance Three Evidence Tiers
What Each Label on a Number Means
Hover over any spec or rating across our reviews to view its tier, source citation, and verification date. We never present estimates as lab measurements.
Published specifications directly from the manufacturer.
Verified against official spec sheets, manuals, and regulatory filings; checked dates are recorded.
Measurements published by verified independent testing labs.
Citations link directly to public, peer-verifiable tests where methodology is disclosed.
Empirical data measured in-house on physical retail units.
Requires a standardized test record with acquisition origin, firmware version, and ambient conditions.
02 Protocol Published in Advance
The Protocol a Future BVC Measurement Will Follow
This is the protocol we apply to every machine we test, so results can be compared like for like.
Current status: no unit has completed this protocol yet, so none of our 53 scores is a BVC measurement. Scores currently rest on manufacturer specifications and independent sources we cite, and every value carries a tier label showing which.
Debris Pickup
50 g fine flour, 30 g white rice, and 20 g collected pet hair weighed to ±0.1 g precision. Spread over 1.0 m × 0.5 m test zones on sealed oak and medium-pile carpet. Each run uses default cleaning mode with measured recovery.
- Hardwood & medium-pile carpet
- Flour, rice, pet hair batches
- Mean of 3 runs (±0.1 g scale)
Obstacle Avoidance
Power cables, dark socks, shoes with laces, bathroom scales, charger blocks, faux pet waste, and low-profile thresholds. A hazard is considered avoided only if the robot neither drags, tangles, nor pushes it off mark.
- 20 fixed obstacle types
- Zero drag & zero ingest criteria
- 5 repeated obstacle trials
Acoustic Sound Analysis
Acoustic sound pressure measured at 1.0 meter in an anechoic-tested baseline room under 35 dB(A) ambient noise. Readings taken across Eco, Normal, and Turbo modes, plus peak noise during dock auto-empty suction cycles.
- 1.0 m distance, ear height
- Standard, high & turbo modes
- Dock auto-empty cycle peak
Battery Runtime
Tested continuously with motorized brush rollers spinning on standard floor surfaces at 20–24 °C. We log the shortest run of two passes. Manufacturer "up to" figures are shown side-by-side to highlight real-world discrepancies.
- Motorized agitators active
- Shorter of two runs recorded
- Continuous real-world test
Dock & Maintenance
Continuous observation of dustbin evacuation seals, mop-roller hot water cleaning efficiency, pad bacteria odor prevention, and timed manual teardowns for hair untangling and HEPA filter rinsing.
- Sealed dustbin extraction
- 140°F–167°F hot wash checks
- Timed manual maintenance
Scoring Architecture
Criteria are directly grounded in evidence tiers. Hardwood, carpet, and pet hair carry the highest weight, followed by obstacle navigation and dock autonomy. The overall 10-point score represents real-world balance, not a blind average.
- 6 categorized criteria
- Weighted for home layouts
- Unbiased editorial rating
03 Scoring System Evidence-Weighted Breakdown
How the Score Card Reads
Six criteria on a calibrated 10-point scale, paired with an editorial verdict. The sub-scores provide the granular evidence; the overall score delivers the buying judgment.
Score Breakdown (6 metrics)
Floor-Specific Insights Over Generic Averages
A flagship vacuum boasting 10,000 Pa suction can still tangle on long pet hair or scatter dry debris across hard floors. Splitting scores into Hardwood, Carpet, Pet Hair, Edge Agitation, and App Control lets you select machines tailored to your specific layout.
Our overall rating weighs cleaning effectiveness highest, with dock automation and app reliability adjusting the final score.
04 Outside Voices Discourse, Sentiment & Trends
How Online Discourse and Public Sentiment Shape Our Evaluation
Our evaluation and ranking methodology considers online discourse, public sentiment, current trends and reader input alongside our own scoring. We use them to check our scores before we stand behind them, and we show our working on every review page.
Online discourse
Wherever they exist, we read published expert reviews and the commentary of established reviewers, and cite community or video discussion when a thread can be opened. Every source is dated and linked under “What Outside Reviewers Say”.
Public sentiment
We record owner ratings and review counts from public listings when the page can be read, such as ProductReview.com.au, and we name the page. We do not average star ratings into our score.
Current trends
We note when a model has been replaced by a newer one, dropped from a reviewer’s recommendations, or is an older design, each with a cited page, so a ranking does not quietly go stale.
Reader input
Corrections and reports sent through our contact page are checked against the cited sources before anything changes, and fixes are logged publicly.
How it affects a ranking
Outside voices test a score; they do not set it. When reviewers or owners disagree with us, or with each other, the review page says so. When fewer than two independent sources exist, the page says coverage is limited. Review counts are small and uneven, so we do not count followers, likes or mentions.
Coverage today
35 of 53 products have an Outside Voices section, drawing on 55 cited sources. Products with no publicly readable record are left empty rather than filled in, and forum threads and several retailer pages could not be read.
05 Ethics Full Transparency & Public Log
Our Commitments to You
Editorial integrity, a retesting policy, and public error correction govern everything we publish.
Retail Acquisition
Every tested unit’s review states whether it was purchased at retail, a loaner, or manufacturer-supplied. Manufacturers never see scores before publication and cannot edit reviews.
Firmware Versioning
Measured numbers carry the firmware version they were recorded on. Our policy is to schedule a retest after any over-the-air update that alters navigation or suction performance.
Automatic Retesting
Our policy is to review a model for retesting when its street price moves more than 20% or 12 months after it was last tested. This is a manual editorial check, not an automated trigger.
Corrections
Any inaccurate figure is corrected immediately in place, and the review shows the date its data was last verified.
Affiliate Independence
Commercial affiliate links never influence a vacuum’s score, tier placement, or editorial recommendation. Our editorial scoring stays completely independent.
