After running more than 90 cleaning cycles with 14 of the most promising robot vacuums in 2017, we think the EcoVacs Deebot N79 is the best choice for most people who want a basic, affordable bot that can actually keep their floors tidy.
The Deebot N79 is nimble enough to navigate through most homes without getting stuck very often—and that’s what really makes most bot-owners happy. Its battery life is the longest we’ve seen, it runs the quietest, and it has one of the better control schemes we’ve seen for the price—including Wi-Fi and a smartphone app. It works best on short carpet and bare floors, in areas smaller than 1,200 square feet, with modest amounts of pet hair—limitations that stronger, smarter, pricier bots don’t have. But the Deebot N79 is the best of the affordable robot vacuums, and if you run it at least a few times per week, it can keep your floors free of obvious debris with very little effort on your part.
If the price of the EcoVacs Deebot N79 jumps, or if the robot goes out of stock, check out the Eufy Robovac 11 instead. They are almost exactly the same, but the Eufy doesn’t have Wi-Fi and usually costs extra. It was our top recommendation in this guide for the first half of 2017, and we’ve heard mostly very positive things from our readers about it.
Compared with our main pick, the iRobot Roomba 690 has a shorter battery life and a steeper price, and it’s a little more likely to get stuck mid-cycle. But we think it’ll last longer than those models because it’s based on the same design as the Roomba 650, our main pick from 2013 until early 2017, which has a years-long track record for reliability and owner satisfaction. It’s meant to be repaired over time, and the brand has always done an excellent job keeping replacement parts available. On balance, the Roomba 690 cleans and navigates about as well as the EcoVacs Deebot N79 and Eufy Robovac 11. It can also connect to Wi-Fi, so you can control it with your phone or with Alexa voice commands.
For tougher jobs, like cleaning a large home or digging a lot of pet hair out of carpets, we really like the iRobot Roomba 960. It has much more cleaning power than our main picks, and a sophisticated navigation system lets it clean an entire level of your home, room by room, without missing any patches. It also works with a smartphone app and Alexa voice commands. Compared to other high-end, full-featured robot vacuums, the Roomba 960 is more likely to finish a cycle without getting stuck or quitting (though slightly less likely than our main pick or runner-up) and hits a more right balance of price and performance.
I’ve covered robot vacuums for The Sweethome since 2013, logging hundreds of hours of research and testing in that time. Altogether I’ve tried out 25 bots from 10 brands. I also write about other types of vacuums for The Sweethome, including cordless, handheld, and traditional plug-in styles.
This version covers all robot vacuums available in the US as of late July 2017. In 2017 alone, I’ve put in about 40 hours of research into the latest robots, dedicated 35 hours to comparative side-by-side testing, and spent dozens of hours just letting the best few bots do their thing cleaning my condo.
Over the years, we’ve also spoken with several experts, including:
Many informal conversations with other editors, enthusiasts, engineers, brand representatives, and regular robot owners influenced me. I’ve also made a point to listen to as many of our readers as I can, through comments on our guides, emails, and tweets. I’ve spent some time on message boards, too, particularly Robot Reviews.
I like to read vacuum reviews from a variety of other sources. I have scanned through a few thousand user reviews and read dozens of product-specific reviews from testing houses including CNET, Consumer Reports (subscription required), Reviewed.com, and Good Housekeeping.
On top of all that, I’ve run about 200 at-home cleaning cycles over the years with our previous picks. It would be more, but I have to test a whole mess of other vacuums, too.
A robot vacuum can’t fully replace a regular vacuum. However, it’s a hell of a lot more convenient to let the robot do the work most of the time. A bot that runs 90 minutes a day 3 times per week will keep your home much tidier than 10 minutes of half-assed, human-driven vacuuming a few times per month.
“[Robot vacuums] are best at what I’d call maintenance cleaning,” said Sal Cangeloso, former editor in chief at Geek.com. He reviewed a bunch of different bots during his tenure and has owned a few iRobot Roomba models. “The human does the big clean, say, once a month, and then you have the robot clean a few times a week. This’ll keep your place clean and make it so that a few missed corners and stuck-on grime aren’t a big deal.”
Over several years of research and testing, we’ve found that a robot vacuum can work well for most people in most homes on most kinds of flooring. Any decent bot can pick up obvious, surface-level debris like pet hair, crumbs, road grit, or anything else you can see from eye level or feel stuck to your feet.
Most people who own a robot vacuum also own a stronger, human-driven vacuum for tougher jobs like deep-cleaning fine dust and ground-in hair from carpets. Long carpets in general are a no-no, because the fibers can jam the bot’s wheels and brushes. Robots also won’t work at all on stairs, couch cushions, drapes, a car interior—anywhere except a floor. (If you need advice, we have recommendations for plug-in vacuums, cordless vacuums, and handheld vacuums.)
Even the best bots, working in the right environment, won’t work perfectly all the time. They’ll all get stuck or tangled occasionally, and in some homes, they may get stuck fairly often. And they often navigate in ways that don’t make sense to human observers.
Most owners learn to accept these quirks, make small changes to adapt to them, or simply don’t notice them at all. But some people, in the end, can’t get comfortable with the limitations. Try to buy from a retailer with a return policy of at least a few weeks in case the bot just isn’t working out for you.
We started by making a list of all the cordless vacuums we could find. Since 2013, we’ve tracked 121 models (though many are now discontinued). Then, we prioritized some baseline specs:
Beyond that, we tried to keep an open mind. Experience has taught us that better specs and advanced features often don’t add up to a better bot for the money, so we stayed impartial to all navigation styles, battery run times, brush designs, suction strengths, and control schemes.
The least you can pay for a decent bot (at the time of writing, at least) is about $160. Most bots below that price don’t meet our baseline specs and tend to have low user ratings due to dumb navigation and weak cleaning.
Up to $300, the best bots are good enough and cheap enough to make most people happy, we think. They’re not as smart or strong as the pricier bots, but they’ll keep your floors tidy if you run them a few times per week. They have semi-random navigation patterns and modest cleaning power, but the good models don’t get stuck and have enough battery life to offset their limitations. They’re at their best in smaller spaces (comfortably up to 800 square feet, stretching it up to 1,200 square feet) with mostly bare floors and some short rugs or carpets. We never used to consider bots in this price range, but they’re much better than they used to be. Our main pick and runner-up fall into this category.
Between $300 and $600, some advanced features start to appear, including orderly room-to-room navigation, stronger cleaning power, or at least better brand support. None of them have all of those features, though. Some people will find their ideal bot in this price range (we’ll talk about a few of the best models later), but we don’t think any one model here is the best choice for most people.
At $600 and above, full-featured bots start to appear. These high-end models come with all of the useful advanced features found smattered across the midrange bots. That is, they can clean an entire floor of your home in an orderly fashion, they can be controlled with a smartphone, and they have more cleaning power than the cheaper models. For people in larger homes, with some extra money, and who don’t want to mess around with the limitations of cheaper bots, this is where you should look. Our upgrade pick falls into this category.
For this update, we narrowed down to the following models and then tested them.1
High-end ($601 and up)
|iLife A4siLife A4||iRobot Roomba 690||iRobot Roomba 960|
|Eufy RoboVac 11||LG Hom-Bot Turbo+|
|EcoVacs Deebot N79||Samsung Powerbot R7070|
|Monoprice Smartvac 2.0|
The most important trait we looked for was continuous navigation— no getting stuck or otherwise quitting mid-cleaning. Over many years of testing, we have found that as long as a robot keeps moving, it will do a good job keeping your floors tidy. In most homes, this means the robot will need to be able to successfully navigate around or through the most common bot traps including:
We also looked for bots that could cover the most ground on a single charge. A longer battery life helps here, but so does quickness—some bots move faster and back out of obstacles swiftly, while others move more deliberately. For pricier bots, we also expected orderly room-to-room or straight-line navigation and the ability to consistently return to their dock to recharge at the end of a cycle.
Cleaning ability was another focus point. Realistically, a bot should be able to suck up all the obvious, surface-level bits on your floors and short carpets, including along edges and in corners, without missing big patches of open ground. (Higher-priced bots should be able to pick up more fine dust, and work better overall on higher-pile rugs.) There are a few paths to that goal. Some bots to clean everything in one pass with strong suction and aggressive brushes; others make multiple passes with weaker cleaning. We think that either approach is valid, and results are what matter.
We also favored quieter bots, simpler control schemes, Wi-Fi connectivity and smartphone-app control, availability of replacement parts, expected reliability based on the brand’s reputation and any customer reviews we can find, and the quality of customer service.
To measure all this, we ran each finalist through at least three whole-house cleaning sessions, like a regular owner would. We run the cycles in my current condo, which has roughly 1,000 square feet of robot-accessible floor space. It has no permanent carpet, but it does have 10 different area rugs, ranging from lightweight doormats to rubber-backed, short-plush rugs that take up half a room. The home was built in the 1920s and still has an old-school layout, with several small rooms rather than a few large spaces. So the bots had lots of edges, thresholds, and tight spaces to navigate. This is one of the trickier environments for a bot to navigate smoothly.
For the first couple of sessions, we closely monitored each bot, noting its nav style and obstacle-detection and hazard-escaping habits. (To save time, we ran some bots concurrently in this stage—yes, it’s fun to watch robots bash into each other.) We also looked at how full the bin was at the end of each cycle and made a note of the predominant kind of debris. Some bots reliably came back with mostly full bins, even when the floor seemed clean. Others started to come back mostly empty after a couple day of intense testing. If the bots did well in the first few sessions, we ran then up to a dozen more times, experimenting with different dock positions, cleaning modes, and the number and severity of hazards we left on the floor.
We also ran a more-controlled stress test where we enclosed each robot vacuum into an area cluttered with several chairs, stray USB cables, a sock, a medium-lightweight area rug with tassels, and a tall threshold, plus ¼ cup of all-purpose flour spread across the floor and rug, including along a baseboard. (We used to use ground coffee in this test as well, but we didn’t learn much from that because they all do well with it, so we cut that part.) Watching the bots deal with these obstacles gave us a clearer idea of how each one handled certain obstacles that are known to cause trouble for some robots, and it also gave us an obvious visual of how well each handled difficult debris.
The EcoVacs Deebot N79 is the most basic, affordable robot vacuum that actually works well. In our testing and research, the Deebot N79 was the most likely to complete a cleaning cycle on its own without getting stuck and waiting for a human to rescue it. That’s the most important part of a robot vacuum’s job, and the N79 does it better than almost any other model we’ve seen, even those that cost hundreds more. It’s also one of the most affordable robots that you can control from a smartphone, which gives you the flexibility to start a cycle while you’re away from home. And it’s tied for the quietest robot we’ve tested. It’s not a particularly strong cleaner, and its semi-random navigation system may miss patches of floor in larger homes. But the two-hour battery life (among the longest we’ve seen) helps to offset those limitations through sheer persistence.
If you live in a home with a sprawling floor plan or you have lots of plush carpeting and a few long-haired pets, you’ll want something smarter or stronger than this. But for most homes, the Deebot N79 is good enough to keep the floors tidy if run a few times per week.
The navigation system is the feature that makes the Deebot N79 great. Of any vacuum we’ve tested at any price, it’s the least likely to get stuck mid-session on common hazards and obstacles. And after years of experience, we’ve found that the most important trait a bot can have is to just keep driving. Look at it this way: If you schedule your bot to clean while you’re at work and it gets stuck on carpet fringe 10 minutes into the cleaning cycle, your floors will still be dirty when you get home. Defeats the purpose of having an automatic cleaner, no?
We’ve run the Deebot N79 and functionally identical Eufy Robovac 11 more than 40 times, and in total they’ve only gotten stuck mid-cycle a few times. It’s always been on a stray USB cable or sock, which are easy to pick up, rather than the thresholds and rug fringes that trip up lots of other bots and are harder to move. Whenever the bot does sense that it’s caught up in one of those typical hazards, it has a routine for escaping, like turning off its brush and rolling backwards from cables or bunched up fabric, or rocking back and forth to get off of a tall threshold. Some competing bots will try to power through obstacles and get even more stuck, while others give up and cry for help as soon as they sense trouble.
We’re not exactly sure why the Deebot N79 is more flexible and less prone to getting stuck than others. As far as we can tell, it doesn’t have any groundbreaking technology that gives it an edge. It’s just a smarter implementation of the same basic designs used in most sub-$600 robots—a semi-random, bump-and-run navigation style that relies on near-range sensors to feel its way around an environment.
The Deebot N79 also gets into some areas that some competitors can’t or won’t. It’s shorter and narrower than most of its rivals, so it gets under more furniture and toe kicks, and it’s more likely to glide between tighter chair legs. It’s not shy about driving into crowded areas, like the space under a dining room table—more-advanced bots may purposely avoid those types of hazards.
In our testing, we found the Deebot N79 to be an effective cleaner. It regularly sucked up just about all of the cat hair, crumbs, and grit in our 1,000-square-foot test home, even along edges. It doesn’t have as much raw cleaning power as midrange bots do, but it’s strong enough on bare floors and short carpets to pick up the kinds of debris you can see from eye level or feel stuck to the bottom of your feet. It also just covers more ground than most other bots, thanks to the long battery life and nimble, wide-ranging nav system. (EcoVacs claims that the battery life is around 100 minutes, though we’ve seen it run as long as 150 minutes. It just switches to a lower-suction mode and starts searching for its dock after the 100-minute threshold.) Since it’s out on the floor for such a long time, making multiple passes over most areas, it picks up an impressive amount of debris for a robot at this price.
The Deebot N79 is one of the most affordable robots that works with a smartphone app. You can connect the bot to your home Wi-Fi network in about three minutes and then use the app to start, stop, or steer it from anywhere—from work, the store, another room, whatever. (Most other robots use a basic remote control for these functions.) Since the N79 is a lower-end bot, with no on-board camera for the nav system, the app has fewer of the gimmicky features than high-end bots offer. But the important functionality is there. The app should support multiple EcoVacs bots if you want it to, though we didn’t test this feature.
The N79 is also the quietest bot we’ve tested, tied with the Eufy Robovac 11, iLife A4s, and other similar models from these brands. It probably won’t disturb you if you’re home while it’s running. From 6 feet away, we measured the volume at a comfortable 60 dBC, with no irritating whooshing or whining frequencies. From two rooms away, we measured it at 44 dBC, which is barely audible. It’s a little easier on the ears than the Roomba 690, which runs at about the same volume but is a bit more grating in the midrange frequencies. The N79 is also noticeably quieter than any high-end models, which have stronger motors. If you’re out of the house when your bot runs, as most people are, noise is a non-issue anyhow. But the N79 gives you the option of running it while you watch TV without forcing you to crank the volume. The inoffensive whirring and seemingly nonsensical cleaning patterns almost make it feel like a pet.
The control scheme on the Deebot N79 is simple if you want it to be. All you really ever need to do is to press the auto-mode start button on the bot itself, the remote, or the smartphone app. Based on our testing, we think you’ll only ever need to use auto mode. Some of the extra modes could prove useful, like the manual-steering controls or the edge-cleaning mode, but we’re hard-pressed to think of when that might be.
EcoVacs covers the Deebot N79 with a one-year warranty, which is typical for robot vacuums. It’s hard to accurately gauge the reliability and customer service from a relatively new and small brand like EcoVacs (at least their US operation). They’ve been on the fringes of the US robot vacuum category for a few years, and so far we can’t recall hearing about any egregious problems with their products or customer service. Their shipping is also ‘Fulfilled by Amazon’ so you get the protections of Amazon’s return and exchange policy. We’re comfortable spending our own money on this robot, but if you want to buy from a more-recognizable brand, check out our other picks.
Maintaining the Deebot N79 is like most other bots. Most of the moving parts pop out for cleaning, no tools needed. You’ll need a small Phillips-head screwdriver bit to remove the side brushes, but any decent toolkit will have one. It also comes with some spare side brushes and filters and a brush-cleaning tool (as most bots do). It’s as easy to clean as any of the other budget bots we’ve come across, at least in the short- and medium-term. We’re not sure yet how well this bit will hold up over time, or whether replacement parts will be available in a few years.
EcoVacs does not currently sell replacement parts for the Deebot N79, which is a problem. You’ll need new filters and brushes after a few months, and maybe a battery after a couple of years. An EcoVacs representative told us that these third-party replacement brushes from BettaWell are safe to use. They also said that they planned to start selling parts through their own site in July, but that didn’t happen by the time of this writing. We don’t know if those parts will actually materialize, or, more importantly, whether more substantial parts like new batteries, rollers, or wheels will be available. This is one of the main reasons you might want to consider buying a bot from a brand with a reputation for supporting their products over time, like iRobot.
The Deebot N79 does not have as much raw cleaning power as most of the other bots we’ve tested. If you have lots of thick carpet and long-haired pets, you’ll probably need to upgrade to a stronger bot than the N79. Compared with most others, its suction seems weaker, and its single 5.5-inch brush roll is smaller. It just does not create as much airflow or agitation as most others.
The N79 makes up for the weaker cleaning power by running longer cycles and making more passes. On bare floors, it works just about as well as most bots. But on rugs and carpets, its shortcomings are more obvious. Even with multiple passes, we found that it often left behind some flour and cat hair that other bots could get out of our rugs.
Over a couple of weeks of consistent, multiple-times-per-day cleaning sessions with several bots, we found that the functionally identical Eufy Robovac 11 finished cycles with a mostly empty bin, while our upgrade pick consistently collected noticeable amounts of hair and dust. (When we scaled back the frequency of cleaning sessions, the Robovac 11 started coming back with full bins again.)
All that said, the Deebot N79 is effective enough in the real world to make your floors look and feel tidy, just not keep your carpets perfectly debris-free. We’ve heard from many of our readers over the years that clean enough is good enough, and the N79 hits that mark as well as any other low-end bot. If you deep-clean your carpets with a strong human-powered vacuum anyway (as most bot manufacturers suggest that you do), then the N79 can keep the crumbs and hair off the floor between the big sessions.
Like all other bots in this price range, the N79 navigates semi-randomly. It scoots forward until it gets right near an obstacle, then turns at an angle, and it repeats until the battery runs low (with a few exceptions, like edge-cleaning runs and trap-escaping routines). It doesn’t remember where it’s been in your home and doesn’t really plan where it’s going either.
Some people just hate the “aggravating randomness” of the bump-and-run style, as former Wirecutter editor John Neff put it. He has owned both an older bump-and-run Roomba and a straight-line Neato, and he greatly prefers the straight-line style. “I guess I’m a perfectionist-type, logical person in that regard,” he said.
The other side-effect of the bump-and-run style is that the robot can miss patches of your floor on any given cycle. The best thing to do is run the robot every day, or at least a few times per week. Whatever it misses today, it will usually pick up next time. It’s not the flashiest technology, but as one Amazon customer reviewer wrote about the functionally identical Eufy Robovac 11, “in its own dumb way it will get the job done more often than not.” We’ve found time and time again that this semi-random system works as well as orderly, straight-line navigation for keeping floors tidy over time, as long as the area isn’t too large.
We estimate that the N79 can handle up to 1,200 square feet fairly reliably. Beyond that, you’ll have to resort to tactics like shutting the doors to certain rooms to focus the cleaning area (it does not come with virtual barriers, unlike most midrange or high-end bots), or starting it in a different room every day. That’s still more space than the Roomba 690 and its shorter battery life can comfortably handle.
Like all other robot vacuums, the N79 can struggle with certain hazards. It can’t deal with very high thresholds, roughly anything taller than 15 millimeters. It can get caught on stray cords or curtains and will sometimes get jammed if it drives into laundry that’s been left on the floor. And like other robots with ledge sensors, it does not always work on dark, non-reflective surfaces like matte black wood floors or very dark rugs. The sensor interprets the absence of a reflection as a drop-off, so it stops in place to prevent itself from faceplanting down the stairs. All of these issues are typical of most other robot vacuums, and we still think that the N79 will navigate the real-world with fewer hiccups than other bots.
At the time of writing, not many other reviews of the N79 are available. None of the other major editorial outlets have published reviews yet. And while there are more than 150 Amazon customer ratings, Fakespot gives the reviews a grade of D, which means that they think around half of the ratings are low-quality or inauthentic. This doesn’t mean that the N79 is a bad product, just that somebody could be trying to artificially inflate the rating to try to get more people to buy the product—a shady tactic that’s unfortunately pretty common across the electronics industry. We’ll keep an eye out for more reviews over the next few months.
If the EcoVacs Deebot N79 goes out of stock or the price jumps, check out the Eufy Robovac 11 instead. They are almost exactly the same robots, but the Eufy doesn’t have Wi-Fi or smartphone-app control, and it usually costs extra. It was our top recommendation in this guide for the first half of 2017, and we’ve heard mostly very positive things from our readers about it.
The obvious downside to the missing Wi-Fi is that you can’t start the Eufy Robovac 11 from your phone while you’re away from home. Another downside is that it’s trickier to pre-schedule cleanings. If you want the Eufy to automatically start at a certain time, you’ll need to set it the time on the remote, and make sure that the remote is pointing toward the bot at that time. You can’t set a 7-day schedule, either.
Apart from the connectivity, the same pros and cons to our main pick also apply to this Eufy. We’re more familiar with Eufy as a brand than with EcoVacs. Eufy is an offshoot of Anker, one of the more reliable and value-minded mobile-accessories brands of the past few years. Their customer service representatives are present in the Q&A sections of Amazon product listings, and we can’t recall hearing any horror stories about poor customer service. The Robovac 11 has an average Amazon customer rating of 4.5 stars based on more than 1,200 reviews. That’s a great score, even though Fakespot found a few reviews that may not be authentic or reliable, giving the product listing a B.
While Eufy promised us earlier this year that they would stock authentic replacement parts for the Robovac 11, they only stayed in stock for a few weeks. A rep from Eufy told us that they had “underestimated the market demands” for those parts, and “are expecting to restock all replacements in August and keep them to be continuously available on Amazon.” They also plan to start selling replacement batteries sometime in 2017. We’ll see. In the meantime, those third-party BettaWell brushes and filters should work fine, since the Robovac 11 has the same hardware as the EcoVacs N79, and EcoVacs said that those parts are fine to use. But a representative from Eufy told us that owners should wait for their first-party parts to come out—do whatever you’re comfortable doing.
The iRobot Roomba 690 costs a bunch more than our main pick and performs about the same, but you might save money in the long run. While it’s a new bot with slick features like Wi-Fi connectivity and integration with Alexa voice-commands, it’s based on a tried-and-true design that we know will hold up over many years of use, and its brand has a reputation for supporting their products indefinitely.
The Roomba 690 is pretty much the same bot (plus Wi-Fi) as the old Roomba 650, which was the top pick in this guide from 2013 through early 2017, and the runner-up pick through early May 2017, when Roomba finally discontinued it. The Roomba 650 was as durable as any bot we’ve tested or even read about. We never found any indications of long-term reliability problems in customer reviews, outside of the typical wear and tear that iRobot tells owners to expect. We tested the same Roomba 650 unit on and off for more than three years (about 120 cleaning cycles, we’re estimating), and apart from requiring us to replace the brushes and filters, it never gave us any trouble. The average Amazon customer rating for the Roomba 650 is 4.4 stars out of five, across more than 5,200 reviews. That’s an enormous number of ratings over many years, and it suggests that buyers tend to be happy with what they paid for.
iRobot has also made it easy to continually repair Roomba bots over time (they still sell parts for the very first bots that they released back in 2002!) and we’re confident that the Roomba 690 will get the same treatment. Beyond the basic parts, such as the filters and brushes, iRobot also sells replacement wheels, bins, batteries, and cleaning heads (the brush-roll assembly, basically). The components are easy enough to replace with simple tools, so you should be able to keep the Roomba 690 going for a long time. As we mentioned earlier, Eufy and EcoVacs do not yet sell authentic replacement parts for the RoboVac 11 through its retail channels—not even filters, let alone wheel assemblies. We’re not sure whether continually repairing the Roomba 690 will cost less over time than, say, buying a new N79 or Robovac 11 (or whatever is available) every couple of years, but it might be.
The Roomba 690, on balance, should work about as well as our main pick and runner-up in most homes. It uses a similar bump-and-run navigation system. We found that it’s more prone to getting stuck on cords and cables, as well as light area rugs. It also has a shorter battery life (90 minutes, down from 150 minutes). But it has more raw cleaning power, with stronger suction and two brush rolls, each an inch wider than our main pick’s single roller, so it sucks up more debris with each pass. In most cycles, we’d expect them all to collect similar amounts of debris overall. That said, we think that in larger spaces, the Roomba 690 will miss more patches because it doesn’t run as long. Its effective cleaning area is probably around 1,000 square feet, down from 1,200. The Roomba 690 does come with a virtual wall (basically a battery-powered beacon that tells the bot, “don’t cross this line”), so if you’re in a larger space, you can block off certain areas to keep the bot more focused.
We found that the Roomba 690’s smart features worked smoothly. Adding the bot to our home Wi-Fi network was painless. We’ve tested out six different bot-vac apps and we think that iRobot’s is the best. It has the least lag time, and the most useful features like maintenance reminders and links to replacement parts. We also found it very easy to add the iRobot Alexa skill, and while the commands are a little bit of a mouthful, the bot always responded to them promptly. That said, the Roomba 690 has a handful of negative customer ratings citing the difficulties that some owners had with setting up the Wi-Fi or Alexa integration. Your mileage may vary, but we can confirm that we got all the “smart” features to work for us.
The main downside to the Roomba 690 is the price. The official list price is $375, and since it’s a brand-new model, we expect that it’ll stay there for a while. The Roomba 650 regularly dropped down to $325 or even lower toward the end of its lifespan, but you may have to wait a bit before that happens with the Roomba 690. This model is much more expensive than our main pick, even though they have similar cleaning capabilities.
The Roomba 650 also bonked into walls and furniture harder than our main pick (or any robot vacuum we’ve tested, actually), and we expect the Roomba 690 to be similar. For most people, this won’t matter, though it could occasionally leave smudges along baseboards and on light-colored furniture. We’ve been long-term testing the 650 for more than three years and can count on one hand the number of Roomba smudges we’ve seen, and they’ve been easy enough to clean with a little dish detergent and water. But most other bots are gentler than this one.
Like the Eufy bot, the Roomba 690 navigates with a semi-random bump-and-run style, which can bother some people. As Amazon customer reviewer Jack V. Briner put it in his take on the Roomba 650, “If you are a micro manager, you will have problems with how Roomba gets the job done. If you are [a] results driven manager, you will find that the Roomba does a good job.” Life is chaotic, so if a robot that drives in straight lines will bring you a small sense of order or peace or whatever, then step up to our upgrade pick, or look into one of the midrange Neato models.
If you’re concerned about the kind of user data that iRobot can collect through its robots, know that the Roomba 690, with its basic navigation system, is not capable of collecting detailed info about your floor plan.
Plenty of Roomba 650 stock is still available at the time of writing. Most days, it’s the same price as the newer, better Roomba 690, so we think it makes more sense to get the new model. But if you find the Roomba 650 on sale and don’t think you’ll use the smart features, go for it.
If you want the best of what robot vacuums have to offer, check out the iRobot Roomba 960. Given enough time, it can methodically clean an entire level of a house, no matter how large or small, without missing any patches of flooring. Thanks to its tangle-resistant brush rolls and agile trap-escaping tricks, it’s less likely to get stuck mid-cycle than most other high-end bots that we tested. It’s a strong cleaner, even on carpet and with long pet hair. The control scheme is simple, and it works with a smartphone app and Alexa voice commands. Plenty of other high-end bots have similar features, but the Roomba 960 runs more reliably in more homes with less fuss than just about all of them, and actually costs hundreds of dollars less than many.
The feature that sets the Roomba 960 apart from almost all of its competitors (at any price) is the navigation system. It’s built on the same base as the lower-priced Roomba models, using loads of short-range sensors and clever software to work its way around obstacles and wiggle out of hazards, so that it rarely gets stuck before the cleaning cycle is done. As we said earlier, as long as a bot can just keep moving, it’s going to do a pretty good job keeping floors tidy. This basic Roomba nav system is one of the very best at doing that.
But the Roomba 960 adds an extra layer to the nav system that cheaper Roomba models don’t have, allowing it to clean an entire level of your home. It uses a camera on the top of the body and an optical sensor on the bottom (like a computer mouse) to map your floor plan and track its location within that map, so that it cleans your entire floor in a logical, orderly fashion without missing spots. If the battery runs out before the cleaning is complete, it can return to the dock on its own, recharge for a while, then pick up where it left off. This means that it can work effectively even in homes with sprawling floor plans.
Plenty of other bots can work in large homes or clean reliably without getting stuck even in a cluttered floor plan. But the Roomba 960 does both, and it’s the one that we’re most confident will work well in almost any home. Most others come with significant caveats. Other bots that can clean an entire level of a large home, like Neato, Samsung, and Dyson models, tend to get stuck or otherwise stop mid-cycle much more often than the Roomba 960. Cheaper bots that don’t get stuck, like our main pick or runner-up, struggle to clean larger spaces (greater than 1,200 square feet, roughly) thoroughly in a single session.
The Roomba 960 also has a set of tangle-resistant brush rolls, unique to higher-end Roomba models. iRobot actually calls the rolls “extractors,” since they don’t really have any brushes, just nubbed ridges. The extractors don’t get all wrapped up with hair over time like other styles of brush rolls do, so you won’t need to put as much effort into cleaning them out every few weeks.
The extractors are also less likely than other brush roll styles to get caught on your power cords and charging cables. The Roomba 960 didn’t get caught on any dangling USB cables or power cords during our stress testing or around-the-house testing. None of the user reviews we’d seen at the time of publication mentioned it as a problem either. Almost every other model we’ve tested in the past few years got caught on a cable at least once while we used it (except for other Roomba models that use the extractors).
Like most other high-end robot vacuums now, the Roomba 960 works with a smartphone app (for iOS and Android) that can start or stop a cycle, notify you of any errors, help you schedule daily cleanings, and track the bot’s maintenance schedule. Setup is pretty simple; the bot didn’t lose its connection to our home Wi-Fi network during testing, and the app works fine. It also works with Alexa voice commands, and we found the feature to work smoothly. We’ve also found the network reliability to be better and the app to be both easier to use and more robust than other Wi-Fi-connected bots. The control scheme in general (including on-board controls) is wonderfully simple. It also comes with a virtual wall, to block off certain rooms or to set up a perimeter around certain objects, like pet-food bowls.
The Roomba 960 is a strong cleaner by robot vacuum standards and notably better than our main pick at digging hair (and other debris) out of carpet. In our semi-controlled testing, we found that it picked up just about all of the hair and crumbs (coffee grounds) from a wood floor, area rug, and along a baseboard, and managed to get most of the finer dust (flour) we laid out as well—something most other bots struggled with. During our around-the-house testing, it always finished a cleaning cycle with notable amounts of hair and small debris in the bin, even when we’d been running lots of cycles with other models and thought the floors were pretty clean. Part of the extra cleaning power comes from the extractors, which are wider and more aggressive than the bristle and blade brushes that many cheaper bots use. Some of it also comes from the motor, which iRobot claims has “5x the air power” of the Roomba 650, our runner-up pick. (However, it’s not the absolute strongest robot vacuum out there—more on that soon.)
Reviews for the Roomba 960 are strong. The average customer rating at Amazon is 4.5 out of five stars based on 413 reviews at the time of writing. That’s a strong rating for a robot vacuum and consistent with other Roomba models. None of the major editorial testing houses have posted a hands-on review yet, though Consumer Reports (subscription required) gives it a score of 87 and a Recommended status based on its review of the Roomba 880. (We don’t think this is a good comparison, because the Roomba 880 had a much different navigation system, much more like the bump-and-run approach of a cheaper bot.)
And like all Roomba models, replacement parts are easy to find, and we expect iRobot to keep them available for years.
On the downside, the Roomba 960 may be reasonably priced compared with similar models, but it’s a very expensive gadget. It costs almost as much as a full-featured dishwasher or washing machine, and is unlikely to last as long without several rounds of replacement batteries and brushes (which all cost extra). You can get a decent robot vacuum for much less money than this! But if you want the best of what the robot vacuum world has to offer, this is the most sensible way to get it.
A few other models have more raw cleaning power than the Roomba 960, as best we can tell. The Roomba 980 has a motor that can pull twice as much “air power.” CNET wrote that the flagship Neato Botvac Connected is the strongest cleaner they’ve tested. The Dyson 360 Eye is built around a motor that was used in a real vacuum (the cordless DC44 from a few years back), and we do not doubt the company’s claim that it has “twice the suction of any robot vacuum.” Samsung’s latest 7000-series Powerbot models are also very strong. But the Roomba 960 still gets great results.
Like most robot vacuums, the Roomba 960 doesn’t really work on matte-black flooring, very dark carpets, or other dark, non-reflective surfaces.
It can also still occasionally get stuck on hazards like a stray sock, though we believe that it will happen less often than with other bots.
It may also struggle in low light. We’ve run into this issue in our own testing every handful of cycles, and it happened in our Roomba 980 tests, too. Some user reviews mention these problems as well. We’re not even talking about dark rooms at night—this can happen during the daytime with the blinds closed. iRobot told us that “the robot is able to go under dark beds and into a dark room. However, all vision-based systems need at least some light and the 980 will have a limited range in very low light. The [low-light] Error 17 is more likely to happen in a crowded area, like if someone runs it in a dining room in the dark or near dark—because the error in the other sensors becomes too great. iRobot’s customer feedback/studies has found that the vast majority of people run their Roomba during the day.” We still think the Roomba 960 will navigate more successfully more of the time in more homes than the other high-end bots that are out there, but the low-light errors will absolutely be a problem for some people.
Some owner reviews mention more significant navigation errors with the Roomba 960. Several mentioned that it has trouble transitioning between floors and area rugs, and a few more said that it seemed to struggle to navigate around objects and back to its dock. We’re not sure why that seems to happen to some owners; it did not happen with our test unit, and plenty of other owner reviews say the opposite: that it transitions and navigates very smoothly. A representative from iRobot told us that when a Roomba can’t return to its dock, it’s “almost always caused by misplaced home bases, or virtual walls pointing in the wrong direction.” They also pointed out that the Roomba needs “a clear runway on the sides, in front, and above” the dock.
We also noticed that the “full bin” indicator is prone to going on before the bin is actually full, which stops the cycle if a certain setting is enabled. A handful of user reviews pointed this out as well. An iRobot rep told us that this can happen when the sensors get covered in dust, and can be fixed by just wiping the sensors clean. But if it’s a common occurrence during most of your cleaning cycles, just make sure the setting that ends the cycle stays turned off. This way, if the bin isn’t actually full, it’ll keep cleaning. If the bin is full, there’s really no harm—the debris will just stay on the ground, like it would if the bot wasn’t cleaning at all.
Robots from the Roomba 900 Series (including the Roomba 960) create maps as they work. The map is what lets them pause mid-cleaning to recharge their battery and pick up where they left off, and also to get very close to full floor coverage during every cycle. If you connect the bot to Wi-Fi and agree to the terms of service in the iRobot app, then (by default) the bot sends data (not a direct photo or video feed) about your floor plan to iRobot’s servers at the end of a cleaning cycle.
Many other robots from many other brands behave similarly. Bots that have both mapping capabilities and Wi-Fi connectivity include the Neato Botvac Connected series, Samsung Powerbot Series, the Dyson 360 Eye, and certain high-end models from the LG Hom-Bot and EcoVacs Deebot series.
In July 2017, based on an interview with iRobot CEO Colin Angle, Reuters reported that iRobot “could reach a deal to sell its maps,” collected from its robots, to one or more of the “Big Three” smart-home players, Google, Apple, and Amazon. This cause a stir among customers and privacy experts, ranging from concerns about the data being used for ultra-targeted advertising, to more general outrage that a $700 robot would then be used to spy on you for profit.
However, iRobot insisted that Reuters had misinterpreted Angle’s comments, and that they would never sell users’ data. “Colin never said that iRobot would sell customer data to other companies,” a representative from iRobot told us. Reuters later corrected their article to say that iRobot “could reach a deal to share its maps for free with customer consent.” According to iRobot, the idea that Angle was trying to get across in his Reuters interview was that in the future, the detailed maps that their robots can collect could be used to improve the services of other smart-home devices, if the owner chooses to use such a feature.
We don’t think that you need to worry about iRobot, or any other robot brand, selling data about your home’s floor plan. The customer backlash against such a policy could be swift and brutal, if this uproar is any indication, so it doesn’t seem like part of a good business plan. If you’re still nervous, this interview with ZDNet provides some more background on how the mapping features work, so you can decide if you’re comfortable with the technical aspects. (The important detail is that the raw camera feed can’t be served over Wi-Fi, so it seems highly unlikely that anyone could peek into your home that way.) You can also choose to opt out of sending your map data to iRobot’s servers but still use the bot’s other smart features. You can also choose never to connect your Roomba to Wi-Fi at all. It will perform the same, you just won’t be able to control it from the app or with voice commands. If you decide you don’t like the idea of a robot creating a map of your home, you can always buy a lower-end bump-and-run robot without mapping capabilities (like the other picks in this guide).
In August 2017, Eufy announced the RoboVac 11c—its limited-edition upgrade to our runner-up pick. Unlike the RoboVac 11, this upgrade model comes equipped with self-docking technology, as well as Wi-Fi and Alexa voice commands. It’s also $80 more than our runner-up pick. Although we don’t plan to test this version of the RoboVac—nor do we plan to make it a replacement pick—we think it may be a solid option if smart tech integration and a RoboVac that can follow a seven-day cleaning schedule are important features.
The iLife A4s looks similar to the EcoVacs Deebot N79 and Eufy RoboVac 11 and tends to cost a bit less. But it’s different enough that we can’t recommend it ahead of those models. It sits lower to the ground, which causes it to get caught on carpet fringe more easily. It also sometimes tries to power through traps instead of backing out of them, so it ends up getting stuck mid-cycle a little more often.
We also tested the Monoprice Strata Smartvac 2.0. While it’s one of the cheapest bots we could find with all the necessary specs for good performance (plus a mopping attachment), it doesn’t work all that well. The cleaning power is too weak for rugs or fine debris, and it tends to get stuck on rug transitions, tassels, and cords. It also shies away from any kind of thresholds—doesn’t even try to cross them if they’re taller than a few millimeters. it won’t be very effective in most homes.
We’ve also tested out the flagship Roomba 980. It’s a lot like the Roomba 960 (our upgrade pick), but with stronger suction and a longer battery life. It probably does pick up more debris from carpets, but we don’t have any hard data to show how much more debris it really gets. We’re guessing it’s a pretty marginal amount, since the motor itself isn’t any different (the power boost comes from a software tweak). And even the Roomba 980 is nowhere near as powerful as a plug-in or even a good cordless vacuum anyway. Some people might be comfortable spending the extra $200 for these upgrades, but we think most people will have almost the same experience with the Roomba 960.
Tons of discontinued or retailer-specific Roomba models are still floating around from the 600, 700, and 800 series. We’re not going to go over a pecking order of which models have which features and why they might be better deals. Basically, they’re all good robots, so if you find any of these for a good deal, you could consider getting one. But if you want the modern smart features—which we think are worth it—you need to get one of the current bots (690, 890, 960, 980).
iLife and EcoVacs both make tons of other robots. Popular models include the V3s (no brush roll), A6 (higher price, no obvious advantage), and Deebot M80 (higher price, mop attachment). But we think our finalists represent the best values in their lineups.
We’ve tested five Neato robot vacuums from every price point over the past four years in four different homes. The trouble with Neato bots is that they get stuck in silly places much more often than the best bots. We’ve seen them freak out and shut down in crowded areas, like under a dining room table. They’re more prone to wedging themselves between and under furniture. When they get caught on a cable or cord, they’re not gentle, and we’ve even seen them strip the housing off a USB cable. They’re more likely to get stuck on carpet fringe, too, and can’t deal with lightweight area rugs (anything without rubber backing, basically). They don’t really know how to navigate around objects with chrome finish, or objects that curve on a vertical plane, like inclined furniture legs. Their “virtual walls” are just ugly strips of magnetic tape that you need to lay on the floor. As cool as their LIDAR nav-systems are from a geeky technological standpoint, our experience has been that their advanced 2D mapping struggles in real 3D homes.
All that said, a Neato bot can be a good choice if your home has an open, uncluttered floor plan. We think the best model in their lineup right now is the Botvac D5 Connected. It costs less than most other bots with such strong cleaning power and a mapping nav system that can clean an entire level of your home. It also works with a smartphone and has a long battery life. The Botvac D80 is also a good choice that costs a bit less and comes with specialty brushes, but doesn’t connect to Wi-Fi, and has a shorter battery life. The flagship Botvac Connected has a longer battery life and somewhat stronger suction than the D5, but it costs as much as the Roomba 960 (a better bot, if you’re willing to spend that much). The Botvac D3 Connected is the most affordable Neato, but it has no side brush, so it left tons of debris along baseboards in our testing. That’s a real and frustrating weakness.
The LG Hom-Bot Turbo+ did great in our navigation tests. It never got stuck and handled every obstacle easily. It’s actually a bit nimbler than the Roomba 960. It’s quiet, too. However, we found it to be the weakest cleaner of all the high-end bots. In our tests, it left behind more dust and hair than competing models. The control scheme was also pretty confusing, with too many cleaning modes to pick from and unresponsive on-board buttons. The companion app was the least polished of the bunch. It’s also one of the most expensive bots available. We don’t think this will be the best choice for many people.
The Samsung Powerbot R7070 is a very strong cleaner, able to suck up more dust than most of the other bots we tested. It has some cool features, like a side brush that only extends when the bot detects a wall, and a self-cleaning main brush. But like the previous-generation Samsung bots we tested, the R7070 struggles with common hazards more than most other bots. It’s more likely to get itself stuck on rug fringe or stray cords, and is also less able to remove itself from those traps. The interface is cluttered and unintuitive, and even its beeps and bloops are more annoying than other bots’ sounds.
The Dyson 360 Eye is a deeply flawed robot vacuum. Everything we’ve seen suggests that it’s a poor navigator, struggling with even basic floor plans. In this Today segment, it needed human attention five times before it finished cleaning one living room. It didn’t even meet our criteria for testing: It has no side brush, and an awful Amazon customer rating (2.8 stars at the time of writing). Dyson sent us one anyway, and our time with it only confirmed the problems we’d heard about. It’s absolutely the strongest cleaner of any robot vacuum we’ve tested, able to suck up dust from a rug in a single pass, like a good plug-in or cordless vacuum. But it has the least-reliable navigation system of any bot, the shortest battery life, the least-intuitive control scheme, and one of the highest price tags.
We also tested the Hoover Quest 1000, which was one of the more affordable room-to-room navigating bots at the time of its release but is now nothing special. It gets confused and gives up cleaning so easily (sort of like the Dyson and Samsung models) that we don’t think it’s worth considering. The Quest 600, 700 and 800 models are simple bump-and-run bots that cost more than their competitors for no discernible advantage.
Readers keep asking us about the Xiaomi Mi, a LIDAR-navigating bot that costs less than most Neato models (the other brand that uses LIDAR nav). This product is not distributed in the US—it’s only available through importers. We’re not comfortable recommending a product that might be difficult to exchange or repair. If somebody starts to distribute it in the US, we’ll consider looking at it, but for now we have no opinion on this bot.
We also researched dozens of other robot vacuums from Bissell, Black+Decker, Bobi, Infinuvo, Yujin iClebo, and more than a dozen other lesser-known brands that sell through Amazon. They tend to be cheaper bump-and-run models, prone to getting stuck where better models do not. Some are from brands with very little presence in the US, so customer service can be difficult to find if you need it. At this time, we don’t recommend any models from these brands.
There’s some pretty compelling circumstantial evidence that Shark (the brand behind our favorite plug-in upright vacuum) will be releasing a robot vacuum in fall 2017 and it will be manufactured by EcoVacs.
Eufy is very likely to release some new models by the end of the year.
EcoVacs recently released a few higher-end bots, including the LIDAR-mapping Deebot R95. We’ll keep an eye on them for our next update.
Miele will be unveiling the Scout RX2 this fall. We’ll keep an eye out for updates, though if it’s like their last bot, it may just be a rebranded model.
Depending on your bot, you may need to do a little tidying up before you start a cleaning session. Here’s my experience: I prep my apartment by picking up any laundry and charging cables off the floor. I used to have to move my cat’s water bowl, but I’ve found that’s no longer necessary with our latest picks, because they either avoid collisions (EcoVacs and Eufy) or they come with a virtual wall that can set up a “do not cross” perimeter around it (Roomba). I also used to have to pick up lightweight area rugs so that they don’t get bunched up under the bot’s wheels; with the Roomba 690, I still do, but all of our other picks handle them fine.
Expect a few hiccups during the first few sessions. But after that, you’ll figure out your bot’s pain points and learn to make quick adjustments to your home to help it run more smoothly.
All bots need a little maintenance. In most homes, we think a bot will stay in good shape with an hour of work per month, maybe a little more if the bot has a heavy workload.
Filters and side brushes should be replaced a few times per year, the brush roll about once a year, and the battery as needed—probably every second year, though that depends on how often you use the bot. And if your bot suffers a mechanical malfunction outside of the warranty period, you may be able to repair it, particularly if it’s a Roomba or Neato model. Don’t chalk up the bot as a total loss until you check to see if the broken part is available as a replacement.
(Photos by Liam McCabe.)
The alarm code is 1-2-3-4-5.