A woman vacuuming hardwood floors in a living room.

Different vacuums are meant for different tasks: You wouldn’t break out the upright to whisk up spilled Cheerios any more than you’d program a robotic to rid a carpet of ground-in dirt.

Consumer Reports designs its vacuum tests accordingly—to assess the inherent capabilities of each type. 

Vacuums fall into one of two broad categories: large and small. Large, or full-sized, vacuums include canisters and uprights, the most powerful type. Small vacuums include stick vacuums, handhelds, and robotic vacuums.

More on Vacuums

“The main difference between uprights and canisters vs. sticks, robotics, and handhelds is in their ability to deep-clean carpet,” says Frank Rizzi, who conducts Consumer Reports’ vacuum tests. “Lighter vacuums don’t have the strong suction and brush roll combination that uprights and canisters do.”

You can compare a canister with an upright using CR’s Overall Score because they undergo the same tough tests for full-sized vacuums. But outside of these two types, the Overall Score is only comparable with other models within a category.

For example, the Eufy RoboVac 11, a robotic vacuum with an Overall Score of 81, is not a better cleaning machine than the Miele Dynamic U1 Cat & Dog, whose Overall Score of 72 makes it a top-rated upright.

Below, we run you through the gauntlet of our extensive vacuum testing. Click on the type to jump to that section:

Canister and upright vacuums. Stick vacuums. Handheld vacuums. Robotic vacuums.



How We Test Canisters and Uprights

Each of these types has its advantages. If you live in a single-story home with lots of carpets, an upright is a good choice. But if you live in a multilevel home, a canister is a better choice for cleaning all those stairs, so you don’t have to drag the full weight of the vacuum up and down.

We purchase two units of every model and rate their performance on each of the following tests:

  • Carpet. We adapt an industry-standard test to lift 10 grams of surface talc and 90 grams of embedded sand from a medium-pile carpet. The soiled carpet and vacuum are weighed to get a baseline measurement. Then, after a specific series of back-and-forth strokes across the test area performed in a climate-controlled chamber, the carpet and vacuum are weighed again to determine precisely how much debris was cleaned up.
  • Bare floors. Test technicians measure 40 grams of sand and sprinkle it evenly over a marked-off area on a linoleum floor. The empty vacuum gets weighed, passed over the debris field once in both directions, and weighed again to determine how much of the sand the vacuum suctioned up.
  • Tool airflow. The hose of each model gets rigged up to a pressure gauge, and the level of suction is measured in three stages: with the vacuum bag or bin empty, with it filled with 100 grams of wood flour, and again with a total of 200 grams of wood flour. Vacuums are rated on how much suction they maintain, and how strong it is relative to other models.
  • Emissions. Technicians fill each vacuum with 50 grams of wood flour, run it with the brush off the floor in a sealed chamber, and use a laser spectrometer to measure the particle concentration—down to 0.1 micrometers—released by the bag or the bin into the room. Then our testers run the same vacuum over a medium-pile carpet scattered with an additional 20 grams of wood flour, and measure the air again for released particles due to brush agitation. 
  • Pet hair. First, technicians take 5 grams of long, feathery fur from Maine coon cats and strew, stomp, and smash it into medium-pile carpet. Then they make multiple passes with the test vacuum to see how much of the hair it picks up. Vacuums that rate an Excellent collect all the fur in the bag or bin. Average vacs leave behind visible patches or fur becomes entangled in the brush.
  • Handling. We assess handling in two ways: first, how easy it is to push, pull, and carry; and second, the convenience of the features. Each vacuum is weighed with its onboard tools. 
  • Noise. Using a decibel meter, we measure sound at ear level with a vacuum set to deep-clean at its maximum power setting.
  • Predicted reliability. To calculate predicted reliability, we survey our members about the products they own, then use that data to make projections about how new models from a given brand will hold up over time. Predicted reliability scores are based on the estimated breakage rates by the end of the 5th year of ownership for vacuums that are not covered by a service contract.
  • Owner satisfaction. We base brand owner satisfaction on the proportion of members who are extremely likely to recommend their vacuum.

Stick Vacuums

Stick vacuums are more powerful than ever. But while some vacuum manufacturers claim that their stick vacs are as good as uprights, they still can’t compete when it comes to carpet.
  • Carpet. This is the most challenging test for stick vacs, even though it’s easier than the one for full-sized vacuums. Technicians spread a calibrated mix of sand (50 grams), rice (50 grams), and Cheerios (25 grams) on a medium-pile carpet and measure how much the vacuum picks up and how quickly it does so, deriving a score from the combination of the two data points.
  • Bare floors. Vacuums are graded on how much debris (the same recipe) they can pick up from a grouted tile floor in 15 seconds.
  • Pet hair. Technicians embed 2.5 grams of Maine coon cat hair in medium-pile carpet—not as much for full-sized vacuums—and give the vacuum up to 5 seconds to pick it up. The precise amount of hair captured is determined down to the tenth of a gram by comparing the weight of the vacuum before and after the test. 
  • Edges. Technicians broadcast sand in a corner along a simulated baseboard of a tile floor and measure how well a vacuum picks up the grains in one pass by comparing the weight of the vacuum before and after the test.
  • Noise. Using a decibel meter, we measure sound at ear level with a vacuum set at its maximum power setting.
  • Predicted reliability. To calculate predicted reliability, we survey our members about the products they own, then use that data to make projections about how new models from a given brand will hold up over time. Predicted reliability scores are based on the estimated breakage rates by the end of the 5th year of ownership for vacuums that are not covered by a service contract.
  • Owner satisfaction. We base brand owner satisfaction on the proportion of members who are extremely likely to recommend their vacuum.

Handheld Vacuums

You’re most likely to use these smaller vacuums in your car or under the cushions of your couch for quick pickups. They aren’t as powerful as larger vacuums or sticks, so the tests are easier—though still tough enough to reveal meaningful differences among models.

  • Carpet. On medium-pile carpet, each hand vac undergoes three single-ingredient trials in which it has to pick up Cheerios (25 grams), sand (50 grams), and rice (50 grams). For the Cheerios, we time how long it takes to get them all—the best finished in 8 seconds; the worst took 35 seconds (it clogged). For the sand and rice, models are graded on how much they can pick up in 15 seconds. (The vacuum is weighed before and after.)
  • Bare floors. The procedure is the same as for carpet, except it’s performed on a tiled floor.
  • Edges. Technicians measure how far a hand vacuum can reach under a piece of furniture and into a crevice.
  • Pet hair. Technicians spread 1 gram of Maine Coon cat fur on medium-pile carpet, make 14 back and forth strokes, then inspect and rate the cleanliness of the rug and the vacuum’s brush roll.
  • Noise. Using a noise meter, we record the hand vac’s sound level for 15 seconds, then note the average decibel level.

Robotic Vacuums

We not only evaluate how well robotic vacuums clean but also test their prowess at navigating a room, as well as how easy they are to set up and use.

  • Carpet. Technicians disperse 0.6 grams of Maine Coon cat fur, 87 Cheerios, and 10 grams of rice across a test area of medium-pile carpet, and measure how much of this debris the robovac can capture during one cleaning sortie.
  • Bare floors. Cheerios and numbered slips of paper are distributed across a grid marked off on the laminate floor of the obstacle course in CR’s labs, which is complete with simulated furniture. After the robot completes one cycle, technicians count how many Cheerios and paper squares are left behind. Any slips of paper that remain in their corresponding square of the grid indicate that the vac has missed that spot—a not-uncommon occurrence.
  • Navigation. Technicians watch each vacuum as it attempts to maneuver around furniture and other obstacles—a task that may take several hours of observation. They note what the vacuum avoids and what it doesn’t, and whether it gets caught on power cords or carpet fringe. Each model is also judged on whether it can find its dock, or if it’s able to transition between different floor types, such as plush carpet to bare floors. We also test the cliff sensor, to assess whether a robovac is likely to tumble down the stairs.
  • Ease of use. This score is a combination of many factors, such as the size of the dustbin, whether it has a handle for carrying it around your home, how easy it is to clean its brush roll, and whether you can schedule the robotic vacuum for routine cleaning.

Inside the Vacuum Test Lab

In the market for a new vacuum cleaner? CR expert, Sue Booth, shows "Consumer 101" TV show host Jack Rico how Consumer Reports puts models through the paces.