Start with your floors, not the spec sheet
Most people shop by suction number and regret it. Decide what the robot actually has to clean first, then buy. Three buckets cover almost everyone.
Simple hard floors: a small-to-mid apartment with tile, laminate, or wood and not much clutter. You want dependable navigation and low upkeep, not horsepower. A mid-range mapping robot runs here for years with almost no attention.
Carpet and pets: rugs, shedding animals, daily debris. This bucket lives or dies on suction, brush design, and a self-empty bin, because the way it fails is hair welded around the roller and a thimble-sized dustbin you empty every other day.
Large or multi-floor homes: you need long runtime, accurate room-by-room maps that persist across floors, and ideally a dock that empties dust and washes mop pads itself so the robot vanishes into the background.
Pick your bucket before you read another number. It decides your shortlist, and getting it right beats any single figure on the box.
How to read the key specs
Navigation decides whether you love or resent the machine. Lidar (the spinning laser turret on top) maps in total darkness, drives in tight parallel lines instead of bouncing around, and lets you draw no-go zones and label rooms in the app. It is the one feature most worth paying up for. Gyroscope or bump-and-roam robots wander semi-randomly, skip patches, and strand themselves under chairs. Camera-based vSLAM sits between the two and adds obstacle recognition that steers around cables, socks, and pet accidents, the difference between a robot you run only when home and one you trust to run while you are out.
Suction is quoted in Pascals (Pa) and inflated shamelessly, so read it in tiers. Around 2,500-4,000Pa handles hard floors; 5,000Pa and up earns its keep on medium-pile carpet and ground-in pet hair. Past roughly 8,000Pa you are buying decibels and a bigger number, not cleaner floors. Brush design counts as much as suction: an anti-tangle rubber roller is what stops you sawing hair off the axle every weekend.
Mopping spans a damp rag dragged behind the robot to spinning or vibrating pads pressed down hard enough to actually scrub. On mostly hard floors, insist the pad lifts over carpet and the robot detects carpet on its own; skip that and you get soggy rugs. Battery only matters at scale. A cell that recharges mid-job and resumes where it stopped is the line between finishing a big home unattended and rescuing a dead robot at the far end of the hall.
Docks and the recurring-cost trap
The dock is where the money and the convenience now sit. A bare charging dock suits a small home. A self-empty dock vacuums the robot's bin into a sealed bag so you handle dust every month or two instead of daily, the single best upgrade for allergy and pet households. All-in-one docks add clean and dirty water tanks, automatic mop washing, and hot-air drying that kills the sour-mildew reek of a damp pad. They change daily life and they cost a fortune, and they eat real floor space, so measure the corner before you fall for one.
Now the trap: the sticker is only the down payment. Self-empty docks burn through disposable bags, mop pads fray, brushes and filters wear out, and a few all-in-one docks demand proprietary cleaning fluid. Before buying, total a year of bags, filters, side brushes, and pads and add it to the price. A cheaper robot that takes generic parts off any marketplace routinely beats a premium one chained to first-party refills. A station that wants a fresh bag and filter monthly quietly outspends the gap you thought you saved.
When to buy
This category drops in price faster and more predictably than almost any electronics, because it iterates yearly. A new flagship lands high and sheds 20-35% within its first six to nine months as the launch hype cools and stock deepens. The steepest cuts cluster on the named events: Black Friday and Cyber Monday in late November, the Prime Day window in mid-July, and January post-holiday clearance. In Korea, target the 11.11 and year-end festivals; in Japan, year-end plus Prime; across the EU, late November is the reliable floor.
The category-specific play is to ride the new-model dump. Roborock, Ecovacs, Dreame, and Eufy all refresh lineups roughly twice a year, around CES in January and again in late summer, and the outgoing flagship gets gutted the week its successor ships. Last year's top model on a Black Friday cut routinely beats this year's mid-ranger at full price, and all you forfeit is a small suction bump or a slightly cleverer dock. So unless you need it this week, set a price alert and wait. RipeBuy price histories show most models bottoming out in November and December, then drifting up through Q1, which means a "spring sale" is usually the normal street price wearing a fresh sticker.
When to pick an alternative
Let your hard constraint make the call. Pets and carpet: max suction, a tangle-free roller, a self-empty dock; ignore the fancy mop entirely. Mostly hard floors: spend on a pressure-scrubbing mop with auto-lift and carpet detection, not headline Pa. Large or multi-floor: demand lidar with persistent multi-floor maps and a battery that recharges and resumes. Tight on budget or space: a last-gen lidar robot on a plain charging dock does about 90% of the job for far less and skips the bulky station. Then time the buy. The smartest pick is the right machine bought at the bottom of its price cycle, not the flashiest one bought at launch.