The Four Components
Our Overall Rating is a weighted blend of four distinct signals. We do not publish the exact weights, but we explain the inputs and normalization approach below.
1. Owner Sentiment
This is the mean rating assigned to each owner comment tied to a specific model. We normalize this against the dataset average to ensure that a model with 10 comments isn't unfairly penalized compared to one with 100.
2. Recommendation Volume
We count the number of owner comments that explicitly recommend the model or use positive language ("love it," "game changer"). This is log-scaled so that viral models don't saturate the scale while still rewarding popularity.
3. Specs & Features
A capability score derived from populated fields: suction power (Pa), battery runtime (minutes), mop type, self-empty/self-wash capabilities, and obstacle avoidance. Missing specs are treated as "not rated" rather than zero, and the remaining components are reweighted.
4. Budget
Relative price position among models with an observed price. A $300 model with the same specs as a $800 model earns a higher budget score. Prices are snapshots and may change.
Evidence Sourcing
We source owner comments from public community threads (Reddit, forums). Each comment is linked to its origin. We do not scrape private reviews or incentivize feedback.
- Model-Level: Comments that mention a specific SKU or model name.
- Brand-Level: Comments that discuss the brand generally without naming a model. These feed into Brand Pages but not Model Ratings.
Known Limits
We are transparent about where our data is thin.
- Price Volatility: Prices are point-in-time observations. We do not guarantee current retail pricing.
- Long-Term Durability: Our dataset reflects recent owner experiences. We do not lab-test for 5-year lifespan.
- Complaint Bias: Owners are more likely to post when something goes wrong. Complaint themes are directional signals, not definitive proof of failure.
Independence
ZenVacuum does not accept payment for placement. Affiliate links may appear on retailer sites, but commissions never influence our ratings, rankings, or verdicts. If a product is bad, we say so, regardless of commission potential.