Category: Robot Vacuums

Here’s the maintenance a robot vacuum actually needs, stripped of brand fluff: empty the bin after every run (daily if you have a shedding pet), clear hair from the main brush and check the side brushes weekly, wipe the sensors and charging contacts weekly, and clean or swap the filter every one to three months depending on how much dust and fur it’s pulling in. That routine takes under fifteen minutes a week for most households. Everything past that point – the recurring “clean the filter” warning that a new filter doesn’t fix, the dock that smells the moment you open it, the roller that jams solid mid-run – has a specific, mechanical cause, and that’s what the rest of this page is actually about.

There’s a one-page download below that turns this into a checklist you can hold up against your own machine: what to check, how often, and whether the interval changes if you have pets.

Download the Robot Vacuum Maintenance Schedule (CSV)

When a “clean the filter” warning isn’t actually about the filter

This is the failure that wastes the most time, because the fix everyone tries first is the wrong one. The warning says the filter is blocked or wet. You pull it, clean it, dry it fully, put it back – and within a run or two, the same warning is back.

The filter you can see and remove is only half of the air path. Past it, air still has to move through the motor’s own intake and out through an exhaust vent, and both of those have their own small filter and a foam pad that nobody thinks to check because neither one is visible without opening the housing further than a normal cleaning session goes. A blockage there produces the identical on-screen warning as a genuinely dirty dustbin filter, because from the machine’s perspective the symptom – reduced airflow across the fan – is the same either way. Cleaning the dustbin filter over and over does nothing for a blockage that isn’t in the dustbin filter.

If a filter warning survives a real clean-and-dry cycle, the next step isn’t a second filter – it’s opening the motor housing far enough to find that second filter and the foam pad behind it, then cleaning both. If the warning still comes back after that, the motor’s fan assembly itself has likely failed, which is a repair or a replacement part, not a cleaning task. One documented case followed almost exactly that arc: a laser-navigation sensor failure in the first year (fixed under warranty, a hardware fault with nothing to do with cleaning), followed roughly a year later by a filter warning that survived repeated cleaning and full filter replacement, which turned out on paid inspection to be a failed fan motor. Two unrelated hardware failures on one unit, a year apart – worth knowing before you assume every warning light is a chore you’re behind on.

Two different ways brands fight hair-wrap – and where each one still loses

Every brand solves hair-wrap one of two ways, and knowing which one your machine uses tells you what to actually check when it fails.

Cutting designs build a blade, or a row of blades, into the roller itself, so hair gets sliced as it wraps instead of accumulating. This works well on long hair – human or pet – but a short blade positioned only at the two ends of the roller can’t keep pace with a heavy volume of hair across the full width; a blade that runs the roller’s whole length clears more before it needs a hand-cleaning session. Neither version reliably slices short, fine pet fur, which tends to pack into the bristle base rather than wrap in strands long enough to catch a blade.

Airflow designs skip cutting entirely. A tapered or cyclonic roller shape uses the air pressure from the machine’s own suction to comb hair down toward one end and straight into the dustbin, rather than letting it wind around the shaft. There’s nothing to dull or jam, but it depends entirely on suction being uninterrupted – a partially clogged filter reduces the airflow this design depends on, so a filter problem here shows up as a hair-wrap problem too.

A third, hybrid approach – a rubber-and-bristle roller with a raised spiral pattern – reduces wrap significantly at first, but the seam where the bristle strip meets the rubber base is a real gap, and it’s specifically where stray strands accumulate over months of use as that seam wears slightly wider. None of these three approaches is “solved” for every situation; they trade one weak point for a different one.

One data point worth keeping in mind if you’re shedding-pet-heavy: in one household test, hair jammed a mid-tier “anti-tangle” roller solid within a single cleaning run – not a gradual buildup over weeks – and the robot had to be manually cleared before it could resume. Whatever anti-tangle claim is on the box, a single bad run with a heavy shedder is within the range of normal, not a sign the unit is defective.

The dirty-water tank timeline is faster than it looks

If your robot mops, the dirty-water tank is the single fastest-developing odor source in the whole machine, and the timeline is shorter than most people assume: leave it sitting even one day and it already smells; by roughly a day and a half, the water has visibly started to separate and ferment. This isn’t a defect in any one brand – it’s what happens to any standing water that’s picked up floor grime and organic residue at room temperature.

The part that gets missed on a “quick” cleaning: the tank’s float mechanism – the small buoyant piece that rises and falls with the water level to signal when the tank is full – sits directly over a gap in the tank floor that a rinse alone doesn’t reach. The float has to be physically lifted out to get at the grime underneath it, and that hidden pocket is a common reason a tank still smells after what looked like a full cleaning.

If you don’t want to disassemble anything, there’s a fallback that flushes the internal water path without opening the machine: run the mop function repeatedly over a wet section of hard floor (a bathroom floor works well), emptying the dirty-water tank each time, until what comes out is clear rather than gray. It won’t replace an occasional real teardown, but it clears out the water lines themselves.

What a pet household actually breaks – beyond the roller

Three failure modes show up specifically in pet households and rarely get mentioned in general buying advice:

Liquid and soft messes spread rather than lift. A disc-shaped mop pad has a small, fixed contact patch. Point it at something with any liquid content – a knocked-over water bowl, a fresh accident – and the pad’s motion drags that small area of mess into a much larger streak as the robot moves through it, rather than lifting it the way a human would with a paper towel. This is a structural property of a round pad design, not a malfunction, and it means the right response to a wet mess is picking it up by hand first and running the robot afterward, not trusting the robot to handle it.

Large, light debris gets flung past the cleaning path entirely. Clumping litter granules in particular are light enough, and just large enough, that a spinning side brush can fling a piece outward before it ever reaches the main brush or the suction inlet. The robot completes its run and reports a clean pass while a few granules sit visibly nearby, because from the robot’s own sensors, nothing went wrong – the debris simply never entered the path it was measuring.

A dirtied mop pad can re-contaminate floor it hasn’t cleaned yet. On models that auto-return mid-run to rinse a mop pad the machine has flagged as dirty, the return trip only avoids spreading that mess if the pad lifts clear of the floor before the robot moves. If the lift doesn’t engage in time – more likely on hybrid vacuum-mop models than on dedicated mopping robots – the same dirty pad gets dragged back across floor it was supposed to leave clean on the way to the dock.

None of these three are cleaning failures you can fix with a better schedule. They’re what to expect from the hardware, and the practical response is the same in each case: clear known problem spots (litter box perimeter, food and water bowls, anywhere pets have accidents) before a scheduled run, rather than after one goes wrong.

The number on the box assumes a household that doesn’t have your pet

Every auto-empty dock ships with a headline “goes this many days between empties” claim, and as of September 2026, three current auto-empty docks each publish a different figure on their own brand’s site: Roborock’s support page for its current RockDock-family docks says to replace the disposable dust bag every 7 weeks; iRobot’s page for the Roomba Combo 2 Essential’s AutoEmpty dock says its bag holds up to 60 days of dirt and debris; Narwal’s own maintenance guide says the Flow 2’s dock bag lasts up to 120 days. All three numbers come directly from each brand’s own published page, checked today – not from a single independent lab running all three side by side, so treat them as each manufacturer’s best-case claim for its own model rather than a controlled comparison.

Converting those into a shared unit makes them easier to compare against your own household: 7 weeks is 7 weeks by definition; 60 days works out to roughly 8 and a half weeks (60 / 7); 120 days works out to roughly 17 weeks (120 / 7). That’s better than a two-times spread between the shortest and longest claim, on the same basic feature.

None of those day-counts are built around a shedding animal. Narwal’s own guide is explicit about the adjustment on the filter specifically – every two to three months for normal use, every four to six weeks with pets, call it roughly half the interval. Applying that same rough halving to a dock’s day-count is a reasonable adjustment, not a published number: a dock rated for 60 days in a pet-free home is a starting estimate closer to a month with a shedding dog or cat in the house, not eight and a half weeks. Treat the box’s number as the ceiling for a light, pet-free household, and expect to check well before it if you have fur to account for.

When to stop cleaning and start troubleshooting

Every fix on this page assumes the machine is fundamentally working and just needs upkeep. That assumption breaks down in one specific, testable way: if a symptom survives a real cleaning cycle – the actual filter removed, cleaned, and fully dried; the roller actually freed of hair; the sensors actually wiped – done correctly, twice, the problem almost certainly isn’t dirt anymore. At that point the next step is checking the specific hidden points behind the visible ones (the motor’s own intake filter and foam pad, in the case of a recurring suction warning), and if that doesn’t resolve it either, treating it as a hardware fault worth a support ticket rather than a third attempt at the same clean. The households that spend the most time frustrated with a robot vacuum are usually the ones repeating a cleaning step that already didn’t work the first two times, not the ones who stopped and asked what else it could be.

Last updated: September 2026

Two people can test the exact same category of robot vacuum, spend real money doing it, and walk away with opposite advice about whether paying more helps with dog hair specifically. That disagreement is the most useful thing anyone can tell you before you shop, because it means the price tag is not the variable to optimize. Three things are: whether the brush is actually built to resist wrapping, whether the bin math works for how much a dog produces (not a cat), and whether the unit can survive whatever is already sitting on your floor — leash, chew toy, charging cord — without getting stuck.

Two reviewers, two opposite lessons about price

One hands-on tester who bought several mid-to-premium robot vacuum-and-mop combos out of pocket concluded that the pricier units simply packed in more unfinished technology, and that his mid-range pick beat two more expensive competitors outright on real-world results. A separate hands-on comparison, testing across a wider price spread specifically for hair pickup, reached a different practical conclusion: a sub-$300 vacuum-only model with dual rubber brush rollers picked up hair effectively and resisted tangling about as well as anything in the test, while a roughly $2,200 flagship also earned genuine praise for functioning premium features like a mop that reaches baseboards and washes itself with hot water.

Put those two side by side and the contradiction isn’t really about price — it’s about what price is actually buying. Neither test found that a bigger number on the box reliably bought better hair pickup by itself. The flagship’s extra money bought mopping sophistication and obstacle handling, not fundamentally better hair removal; the budget pick’s hair performance came specifically from its brush geometry, not from being cheap. If dog hair is your primary problem and mopping isn’t, that argues for spending on brush design first and everything else second, regardless of which price tier that brush happens to sit in.

The bin math changes with a dog in the house

Dust-bin capacity gets marketed in absolute terms — 400ml, 800ml, whatever the largest number a brand can claim — but the number that actually matters is how many days it buys you, and that depends heavily on how much hair a specific animal sheds, not on a generic “pet owner” assumption. A budget entry-level robot vacuum marketed specifically around living with a heavy-shedding double-coated breed leans on an oversized 800ml bin as its main selling point, precisely because a smaller bin fills mid-cycle during shedding season and the robot just stops working until someone empties it.

There’s a second bin number that gets conflated with the first, and it matters more if you’re shopping today: whether the robot has an auto-emptying dock at all changes which capacity figure to even look at. Without a dock, the robot’s own onboard bin is the whole story, and a small onboard bin has shown up as a real complaint in hands-on reviews of heavy-shedding, multi-dog households — that’s a pattern from testing, not a published threshold, so treat it as a starting point to check against a specific model rather than a rule. With a dock, the robot’s tiny onboard bin barely matters — what matters is the dock’s own bag or bin size, which is usually measured in liters, not milliliters, and is the number that actually determines how many weeks pass before a human has to intervene. Comparing a bagless dock’s multi-liter bin against a robot-only model’s few-hundred-milliliter onboard bin, as if they were the same kind of number, is comparing the wrong thing.

What’s already on your floor matters more than what’s in the spec sheet

One budget pick praised elsewhere in this same testing for its hair pickup was singled out, specifically and by name, for “loving” to wrap itself around loose cables left on the floor — not hair, cables. That’s not a hypothetical for a dog household: a leash left by the door, a chew toy with a cord-like shape, a charging cable for an outdoor camera are all exactly the kind of thin, floor-level clutter that trips up navigation systems regardless of price. None of it is hair, and none of it shows up in a hair-pickup test, but it’s what actually determines whether the robot completes a cleaning run instead of calling for help from across the room.

It’s also worth being honest about where most hands-on testing happens: the reviewer behind the price comparison above explicitly noted his own test home had no kids, no pets, and almost nothing on the floor at all — about as easy a layout as a robot vacuum will ever see. That’s not a knock on the testing; it’s a reason to read obstacle-handling scores as a floor, not a guarantee, for a home that actually has a dog living in it day to day.

What a realistic result looks like

One detailed hands-on account put a number on the gap between marketing and reality: after extensive testing, the estimate was that a capable robot vacuum does roughly 70% of what a human cleaner would do, and takes about twice as long to do it. That’s one person’s assessment of the units they tested, not an audited industry figure, but it’s a useful anchor for expectations: a robot vacuum in a dog household is a maintenance tool that reduces manual vacuuming, not a replacement for it, and it’s worth setting that expectation before buying rather than after.

Brush design is the one place a modest robot can outperform a more expensive one on this exact job. A vacuum-only model with dual rubber brush rollers — a mechanism built specifically to resist hair wrapping rather than just spin faster — picked up hair as well as units with far higher suction ratings, at a fraction of the price, in the same comparison referenced above. Suction moves air; it’s the brush that has to physically release the hair it collects instead of holding onto it, and that’s a mechanical design choice a manufacturer makes once, not something that scales up simply because a model costs more.

Buy for your dog, not for the category

Small or short-haired dog, mostly hard floors. A basic vacuum-only model with an anti-tangle brush roller is enough. You don’t need auto-empty or a large bin to get through a week, and spending more here mostly buys navigation polish you may not need.

Double-coated or heavy-shedding breed. Bin math is the priority, not suction rating. If there’s no auto-empty dock, look for an onboard bin on the larger side for its category — hands-on reviews treat small onboard bins as a real limitation for heavy shedders, though there’s no published industry cutoff for what counts as “large enough.” If there is a dock, check the dock’s own bag or bin size in liters, since that’s the number governing how many weeks you go between manually intervening.

Multiple dogs, or a household that’s genuinely cluttered. Treat obstacle handling as more important than hair-pickup scores, because a robot that gets stuck on a leash or a toy never gets to the hair-pickup part of its job. Look specifically for testing that mentions cables, loose objects, or multi-pet homes — not just a clean demo floor.

You’re deciding whether to spend up for a flagship. Do it for mopping sophistication, obstacle avoidance, or dock convenience — not on the assumption that a higher price alone buys meaningfully better hair pickup. Check the brush design specifically; that’s the spec that predicts hair performance, at any price.

You want the fuller picture on suction, sensors, and what the spec sheet leaves out generally. Robot Vacuum for Pet Hair: What the Spec Sheet Leaves Out covers the broader buying question this page assumes as background — how suction claims are measured, what sensors are now standard, and the cost of ownership that shows up after the return window closes.

Suction rating is the number every listing leads with, and it’s also the number that predicts the least about whether a robot will actually keep up with a shedding dog or cat. Three things matter more, and none of them show up in a bulleted feature list: how the manufacturer measured the number you’re reading, whether the company behind the robot is still standing in three years, and what the machine will cost you in parts and patience after the return window closes. Get those three right first, and the suction spec mostly takes care of itself.

The Pascal number on the box isn’t the number the robot runs at

Every listing prints a maximum suction figure — 5,500 Pa, 8,000 Pa, 10,000 Pa, sometimes higher. That figure is real, but it’s a ceiling, not an operating setting: it’s typically only reached on the strongest manual mode, not the one a robot defaults to on its own. This isn’t published anywhere as a formal spec. It’s a pattern that shows up consistently across the hands-on testing available on this exact behavior, not a documented industry standard, and it hasn’t been verified against every brand: on its default auto mode, a robot generally runs well under its rated maximum, and only ramps up when a sensor detects carpet or a raised threshold. Treat “roughly half” as a rule of thumb worth checking against your specific model, not a number to quote back as fact.

That’s worth knowing because the gap between models tightens once you compare like for like. Ecovacs’ Deebot N30 Pro Omni lists a 10,000 Pa maximum. Eufy’s X10 Pro Omni lists 8,000 Pa. Roborock’s Q5 Max+ — a machine that skips mopping entirely and costs a fraction of either — lists 5,500 Pa paired with a dual rubber brush roll designed specifically to resist hair wrap. None of those three numbers tell you which one will leave less fur on your rug on a Tuesday, because none of them tell you what mode the robot actually runs in day to day, and none of them account for brush design, which does more of the real work than the motor does.

Battery-life claims have the same problem from a different angle: the runtime printed on the box is usually measured at the lowest suction setting, the one you’d rarely choose if you actually have a shedding animal. Push the machine into a stronger mode for carpet or a heavy-shedding week, and the real runtime drops well below the advertised figure. If a listing won’t tell you which mode its runtime number was measured at, assume it wasn’t the strong one.

Sensors that used to be a selling point are now the floor, not the ceiling

A few years ago, anti-collision and anti-drop sensors were marketed as premium features. They’re standard now, on nearly every model at nearly every price point, so a listing that highlights them as a differentiator is telling you less than it sounds like. The feature gap has moved up a level, into how a robot understands what it’s looking at rather than whether it notices an object at all.

LiDAR-based mapping is the current baseline — accurate, works in total darkness, and by 2026 it’s what you should expect even in mid-range machines. Camera-only navigation, without any laser mapping underneath it, is the one configuration worth being cautious about: it struggles more in low light, and it has a harder time telling a shoe apart from a cable apart from a pet accident. iRobot’s Roomba lineup has long relied on this kind of camera-based visual navigation rather than LiDAR, even as competitors moved past it. That track record is part of a larger financial story: on December 14, 2025, iRobot filed for Chapter 11 bankruptcy reorganization in a Delaware court, and on January 23, 2026 it completed a court-supervised transaction handing full ownership to Shenzhen Picea Robotics, the contract manufacturer and lender iRobot already owed a substantial, unpaid manufacturing debt. This is a reorganization, not a shutdown: iRobot’s products, app, and customer support have continued operating throughout. But it’s a useful reminder that the brand behind a robot is a spec too, and one that doesn’t show up on the box.

The blind spot that shows up more than any sensor spec is furniture with thin, framed legs — dining chairs, metal-frame stools, anything a robot can’t distinguish from open floor until it’s already wedged underneath. This trips up cheap and expensive models alike, on more than one navigation system, and no amount of camera-plus-LiDAR sophistication has fully solved it yet. Clear that kind of furniture out of the cleaning path regardless of what you spend.

Two testers, two roller-mop verdicts — and that’s the actual finding

If you read enough hands-on testing, you’ll run into a genuine disagreement about mop design, and it’s worth sitting with rather than smoothing over. One line of testing describes continuous roller mops — the kind that behave like a wetted conveyor belt — leaving thin streaks of dirty water at the edge of each pass, worse on turns, on the reasoning that a spinning roller pushes water sideways the same way a windshield wiper does rather than lifting it clear. Separate testing on a different roller-equipped machine reports the opposite: dried, stubborn spills lifted with almost no residue left behind. Both are describing roller mops. Both are describing real machines. They don’t agree.

The likely explanation is that “roller mop” isn’t one design — how tightly the roller is scraped and rewetted between passes varies enough between brands and generations that the category-level generalization breaks down. The practical takeaway isn’t “roller mops are bad” or “roller mops are good.” It’s that mop mechanism matters less than mop maintenance engineering, and that’s exactly the kind of detail a spec sheet won’t surface. If wet-cleaning performance matters as much to you as pet hair pickup, look for a specific, recent test of your exact model rather than trusting the mop-type category alone.

That disagreement points at a bigger reliability problem: nearly every published “best robot vacuum” list, including the review sites that show up first when you search this exact topic, discloses that it earns a commission on the products it ranks. That’s not automatically dishonest — plenty of these lists still test the hardware they recommend — but it means a list’s ranking and a list’s incentive are pointed in the same direction, and a review site quietly stops recommending a brand for reasons that have nothing to do with performance. Reading two or three independent tests of the same specific model, not just one “top 5” roundup, catches more of this than checking a star rating ever will.

The costs that show up after the return window closes

A robot vacuum’s real cost of ownership starts after the first month, and almost none of it appears on a product page.

Batteries in this category will measurably degrade over time — not a defect, just lithium-ion chemistry doing what it does, and worth factoring in if you’re comparing a $300 machine against a $900 one on a per-year basis. Models without an auto-empty dock need a replacement dust bag or filter on a recurring schedule, and that recurring cost rarely shows up next to the sticker price. Brushes and rollers wear out on a schedule tied to how much hair they’re actually processing, which for a shedding-pet household is faster than the manual suggests — budget for replacing them, not just for the one-time purchase.

iRobot’s own reorganization is proof that a bankruptcy filing doesn’t automatically kill app support or spare-parts availability — its products and customer service kept running through the entire process. But it’s still a real risk to the multi-year relationship a robot vacuum assumes you’re entering into: ownership can change hands, priorities can shift, and a machine that needs a working app, a cloud account, and a steady supply of replacement filters is a subscription to a company’s continued existence as much as it’s an appliance purchase.

Before you shortlist anything, run it through this

There’s a one-page checklist below that turns everything above into yes/no questions to hold up against any specific listing — what suction mode a runtime claim was measured at, what a replacement filter or dust bag actually costs on a recurring basis, whether the brand behind the listing is financially stable, and how to tell a load-bearing test from a sponsored one. It’s a plain download, not an email-gated PDF: Download the Robot Vacuum Spec-Check List (CSV). Open it in a spreadsheet, paste in the two or three models you’re actually comparing, and fill in the columns before you buy anything.

Who should buy what

You want the least possible daily upkeep. Pay for auto-empty and auto mop-wash together, not one without the other — hand-emptying a bin or hand-washing a mop pad defeats most of the point once shedding season hits. That combination generally starts appearing well above entry-level pricing; treat it as the feature that defines the tier, not suction.

You have mostly hard floors and a fixed budget. A vacuum-only model with a dual rubber brush roll and no mopping hardware at all is a legitimate, cheaper answer — mop maintenance and mop-related carpet issues simply don’t apply to you, and hair pickup on hard floors is well solved even at low suction ratings.

You’re choosing between a well-known brand and a newer one at a similar price. Spend five minutes checking who currently owns and services the brand, not just who founded it. A newer company with a healthy balance sheet can be a safer multi-year bet than an established name going through financial restructuring, and neither fact is on the spec sheet.

You live somewhere cluttered — kids, multiple pets, cables that never quite get put away. This is where paying for a camera-plus-LiDAR combo earns its keep, but go in knowing it reduces the problem rather than solving it; framed furniture legs will still catch it out occasionally, at any price.

Your pet hair problem is specifically dog hair. Robot Vacuum for Dog Hair: The Price Doesn’t Predict the Performance gets into the bin-capacity math for heavy-shedding breeds and a genuine disagreement between two testers about whether a higher price actually buys better hair pickup.

You already know your floors are mostly carpet and pet hair is the whole problem. That’s a narrower question than this page answers, and it deserves its own answer: The Brush-Height Test walks through why carpet performance and suction rating decouple, and what to check specifically for hair embedded in carpet pile.

Forget the suction number on the box. The single spec that predicts whether a robot vacuum will actually clear pet hair off your carpet is something almost no listing mentions: how high its brush roller sits once the mop deck is lowered.

Every hybrid vacuum-and-mop robot has to lift its mop pads off carpet automatically, so they don’t drag dirty water into the fibers. To make room for that lift mechanism, the main brush roller on most hybrids sits slightly higher off the floor than it would on a vacuum-only model. Higher brush, less contact, less hair pulled out of the pile. That’s the whole mechanism, and it explains a pattern that shows up once you compare real carpet-pickup numbers instead of marketing copy: expensive combo robots with roller mops and 25,000+ Pa suction ratings sometimes lose to $300 vacuum-only models on the one job pet owners actually care about.

Roborock’s Qrevo MaxV is a clear example. It ships with a genuinely capable all-rubber anti-tangle brush and 7,000Pa suction, and Roborock’s own spec sheet confirms it raises its dual spinning mop pads by 10mm on carpet to keep them dry — exactly the clearance mechanism that costs it pickup depth. Owners have flagged the same thing independently: more than one has run their own side-by-side comparison against a plain vacuum-only model and reported the same shortfall, visibly more hair left behind on carpet than the spec sheet implies.

The same pattern turns up on the budget end, just in reverse. The eufy L60, a roughly $250–400 hybrid rated at only 5,000Pa suction, has been marketed and tested by some as a strong pet-hair performer, while at least one independent lab test lists pet hair pickup specifically as one of its weak points. Two carefully run tests, same exact model, opposite conclusions. The real lesson isn’t which test is right — it’s that a Pascal rating and a price tag tell you almost nothing about carpet pet-hair performance on their own, and even people who test the same unit carefully can disagree.

Roller mops carry a second, separate problem that has nothing to do with suction: streaking. A roller mop works like a continuously wetted conveyor belt dragged across the floor, and the same way a windshield wiper pushes water sideways instead of lifting it away, a roller mop can leave thin lines of dirty water at the edges of its pass — worse on turns, where the robot pivots and the roller keeps spinning against one spot. This isn’t universal, and the better implementations scrape and rewet the roller constantly to fight it, but it’s a real, repeatable failure mode that a spinning dual-pad mop design doesn’t share in the same way. If your floors are hard surface throughout and mopping matters as much as vacuuming, that’s worth weighing against a roller mop’s other advantages, like more even water coverage per pass.

Then there’s a failure mode that a simple pickup-percentage test can miss entirely: hair that gets lifted off the floor but never makes it to the bin. Some robots trap a large share of what they collect directly in the brush housing and axle instead of depositing it where it’s supposed to go — which still counts as “picked up” on paper, but does nothing for your floor and just moves the mess to a part you now have to clean by hand. A robot with a merely average pickup score but a brush that actually releases what it collects can outperform, in daily use, one that scores higher in a lab test.

None of this shows up on a spec sheet, so here’s what to actually check before buying:

• Look for a documented carpet-pickup percentage, not just a Pascal rating. Suction above roughly 6,000–8,000 Pa shows diminishing returns on hard floors; carpet is where the real gap between models opens up.
• If you have a multi-dog household or a heavy shedder, look specifically for an adjustable auto-empty frequency setting — the ability to have the robot return to its dock and empty its bin more than once per cleaning run, not just a bigger bin. Most major apps bury this under dock or vacuum settings, and it’s worth turning up to high before the first heavy-shedding season, not after.
• Plan on pulling hair out of the brush and axle by hand roughly once a week, regardless of what the anti-tangle marketing promises. Even the better designs slow this down; none eliminate it.
• If the model you’re considering is a vacuum-and-mop hybrid, check whether the brush height changes on carpet, or whether the mop lifts and detaches completely before a vacuum pass. A mop that comes off entirely avoids the clearance problem altogether.
• Don’t assume price tracks carpet pickup specifically. It doesn’t, consistently enough that it’s worth checking independent carpet-specific test data for the exact model, not just its brand’s general reputation.

So who should actually buy what?

Carpet-heavy, heavy shedder. Skip hybrids that keep a permanently attached mop and only lift it a few millimeters over carpet — that lift is precisely the tradeoff that costs pickup depth. Favor a vacuum-only model, or a hybrid whose mop detaches completely before a vacuum pass.

Mostly hard floor, light rugs. The brush-height problem barely applies here, since nearly every current model clears hard-floor pet hair well. Prioritize mopping quality instead, and pick a spinning dual-pad mop over a roller mop if streak-free floors matter more to you than raw suction.

Multiple dogs, or one heavy shedder. Treat adjustable auto-empty frequency as a hard requirement, not a nice-to-have. Bin size matters far less than how often the robot can empty itself mid-run.

Puppy, senior pet, or accident-prone animal. Spend up for real obstacle avoidance — that’s where premium pricing genuinely earns its keep. Just don’t assume the same premium price buys better carpet pickup, because it often doesn’t.

High-pile carpet, long-haired heavy shedder. Be honest about the limits: no robot in this category reliably matches a human running an upright on hair embedded that deep. Treat it as a maintenance tool between real vacuum sessions, not a replacement for one.