Compare evidence against your constraints. Do not average unlike claims into a magic score, and do not treat missing information as a zero or a win.
Gate 1: can it physically do the route?
Check height, thresholds, rug transitions, floor sensitivity, dock envelope, outlet, approach path, and multi-floor handling. Fail a hard constraint before discussing features.
Gate 2: choose the cleaning architecture
Compare brush access, hair handling, bin path, carpet strategy, mop mechanism, water controls, edge reach, and the exact dock functions. Ignore isolated suction numbers unless test conditions are comparable.
Gate 3: price the ownership system
Add bags, filters, brushes, mop parts, solution, battery allowance, repair risk, and maintenance time. Use the calculator with low and high replacement scenarios.
Gate 4: software and privacy fit
Compare map controls, schedule logic, offline fallback, camera acceptance, data handling, household access, account deletion, and update expectations.
Keep an evidence column
Label every row as manufacturer claim, sourced fact, editorial analysis, hands-on observation by a named publisher, or unknown. Record the exact source and checked date. Do not copy a claim from one SKU to a product family.
Make the final decision reversible
Buy through an authorized channel with a realistic return window. During that window, test the worst threshold, darkest room, busiest chair area, pet-hair zone, rug transition, and dock schedule.
Before you buy
- Hard home-fit gates pass
- Exact bundle is identified
- Unknowns remain visible
- Annual parts basket is priced
- Privacy tradeoffs are accepted
- Return-window test plan is written
How to use this guide
This guide is a durable editorial framework. It explains system tradeoffs without assigning product scores or pretending a manufacturer claim is a shared lab measurement. For a model-specific decision, record the exact SKU, region, source URL, checked date, and what remains unknown.