Skip to Content

Video doorbell firm Ring says its devices slash crime—but the evidence looks flimsy

Amazon paid $1 billion for the security company. Our data analysis questions the claims that purchase was based on.
October 19, 2018
Ring | Ms. Tech

When Amazon bought the video doorbell company Ring earlier this year for a reported $1 billion, it was buying more than a hardware startup. Ring doesn’t want to be just another consumer gadget maker: it has the ambitious mission of reducing crime in neighborhoods around the world.

The company’s motion-sensing doorbells have been a hit since they were launched in 2013. An accompanying smartphone app means homeowners don’t have to be home to see and respond to visitors. The company went on to sell security cameras and entire home security systems, first catching the eye of Amazon’s investment fund in 2016. At an event at its Seattle headquarters in September, Amazon devices chief Dave Limp said of Ring’s crime-fighting objective: “I can think of no nobler mission.”

The company seemed to have some evidence that it was succeeding in that aim. Announcing the acquisition, Amazon claimed  a Los Angeles Police Department pilot program in 2015 had found that Ring’s doorbells had reduced burglaries in neighborhoods “by as much as 55%.”

Not only Amazon was convinced. Following that initial pilot, Ring told cities across Los Angeles County that the drop in burglaries occurred after doorbells were installed on just 10% of homes. Apparently, even just a few devices could dramatically reduce break-ins across a community, providing a kind of herd immunity against crime.

“It was this amazing learning that we could be even more effective than we thought,” said Ring founder and CEO Jamie Siminoff at the GeekWire Summit earlier this month. “It taught us that Ring actually protects the homes around it, as much as it protects the home.”

The company went on to make communities an attractive offer: Ring would give their residents a $100 discount on its doorbells, which were selling for about $180 at the time, provided the cities kicked in half the cash. Public records show that over a dozen small cities leaped at the offer, committing between $10,000 and $100,000 in the hope of reducing burglaries. By September 2018, Ring had negotiated contracts worth at least $420,000 for such neighborhood rebate programs.

But a review of public crime data and a previously unreported study show that the evidence the doorbells slash crime is far shakier than the company would have cities and consumers believe. In fact, the only study carried out independently of Ring found that neighborhoods without Ring doorbells were actually less likely to suffer break-ins than those with them.

Fear and break-ins in Los Angeles

In early 2015, Siminoff approached the nonprofit Wilshire Park Association in Los Angeles with a deal. Ring would give hundreds of neighborhood residents a free Ring doorbell, plus a subscription allowing them to record, download, and play back recordings from the device.

Siminoff encountered skepticism. “We were so new that no one would take these things,” he recalled at GeekWire. “We wanted to have 80 to 90% coverage. We only got 10%.”

Ring pushed ahead anyway, installing around 40 devices beginning that May, and by March 2016, it announced the results: a 55% reduction in burglaries in six months, compared with the surrounding neighborhoods. After a joint press conference with the LAPD, the company received glowing media coverage and launched its cash-back campaign in nearby cities.

Ring would not supply MIT Technology Review with any data, methodology, or analysis from the pilot. However, the company later wrote that the 55% reduction took place over seven months.

In April 2016, the LAPD carried out further analysis of the Wilshire Park project “to reflect more recent activity.” That study covered a 10-month period from July 2015 and found a 21% decline in burglaries compared with the year before. Because the surrounding area had experienced an increase in burglaries over the same time, the LAPD researchers “normalized” the Wilshire Park figures to yield a 42% reduction compared with other neighborhoods—closer to but still lower than Ring’s claimed figure. 

Although the LAPD would not share its full methodology or data, the city does make raw crime figures publicly available, classified by neighborhood district. The Wilshire Park Association told MIT Technology Review where the Rings were distributed. Our analysis of the public data shows that, taken together, these three districts had year-on-year increases in burglaries during both the seven-month period detailed by Ring and the 10-month period reported by the LAPD. Without knowing the exact location of the test and which surrounding areas Ring used for comparison, however, it is impossible to check the company’s claims against the public data.

Even if the doorbells had a positive effect, it seemed not to last. In 2017, Wilshire Park suffered more burglaries than in any of the previous seven years.

“I don’t see the decrease in crime [Ring claims],” says Maria Cuellar, a statistician and assistant professor of criminology at the University of Pennsylvania, referring to the public district-level data. She says the sample size is too small, too: “It’s not enough to say whether the effect is something you see in the data, or just some random variation.”

A scientifically rigorous study would ideally involve dozens of geographically separated districts, many more doorbells, and a longer time period to allow any effects to play out, Cuellar says.

A rigorous study would also assign doorbells to homes at random, she adds. “The problem with letting people choose is that you don’t know whether any effect is because of the doorbell or because the people who chose the doorbells are different.” For example, the 10% of residents who opted for the devices might have been more fearful of crime than average, perhaps with reason, and they might already have had taken other security measures. 

Without seeing Ring’s data set, it is also impossible to know how selective the company has been in reporting its findings. The homeowners’ association in the neighboring community of Country Club Park told MIT Technology Review that Ring carried out a parallel study there with 200 devices. The company never reported the results of any such test. Public crime data show that Country Club Park experienced steady increases in domestic burglaries in 2015, 2016, and 2017.

Rival results

Not every community took Ring at its word. John Rock is a management analyst for West Valley City , the  second-largest city in Utah. In August 2017, he suggested that the city test the effectiveness of Ring’s video doorbells for itself. Following Ring’s lead, it would distribute doorbells to about 10% of households in a target area of 764 single-family homes. A neighboring parcel of 754 homes would act as a control group.

The pilot and control areas had experienced similar crime levels in previous years and shared a Neighborhood Watch group. It would be an “apples to apples” comparison, Rock told the city council. His hope was that the program would quantify the police hours (and associated dollars) saved by each doorbell, possibly justifying a subsidy program like the ones in California.

West Valley’s program ran from September 2017 to August 2018. Its results were fairly clear. In one of the two neighborhoods, burglaries fell by 50% over the year, compared with a 41% drop next door. The first neighborhood also saw all property crime (including auto thefts and burglaries) fall by 32%, while its twin had just a 25% dip. And during the last month of the test, the charmed area experienced no property crime whatsoever—an “outlier month that’s hard to account for,” says Rock.

The only problem is, the safer neighborhood was the control group. In other words, homes without Ring doorbells were less likely to suffer a break-in, or property crime of any kind.

Although larger and longer than the LA study, the West Valley program is probably also too small to draw firm conclusions from, and it could likewise suffer from having adjacent target and control areas. Ring also offered a discount to people in both areas to buy a doorbell, “ which I had no way to track,” Rock said.  

Ring would make a second effort to prove its doorbells cut crime before Amazon bought it. Early in 2018, it donated more than 500 doorbells and security cameras to houses in two neighborhoods in Newark, New Jersey. In September, the company announced that home burglaries there were reduced by more than 50% from April through July, compared with the same period in 2017.

Newark makes some crime data available online, although not in as much detail as Los Angeles. In the precincts containing the two pilot neighborhoods, burglaries did fall between April and July—but other parts of the city saw even greater drops, of up to 65%. The latest data show burglaries down by nearly 30% citywide this year, and the biggest improvements were outside the Ring trial precincts.

Newark’s police department told MIT Technology Review that it could not confirm the numbers Ring reported, and once again, Ring declined to share data or results. The only comment the company would make was: “Ring continues to execute on our mission to reduce crime in neighborhoods. We are excited to work with numerous law enforcement agencies … as we support neighbors in pursuit of our mission.”

“I would like to see an honest analysis comparing [Newark’s trial zones] to the surrounding areas,” says Cuellar. “Otherwise, if one only looks at before and after in one precinct, it could be that something else is causing the rise or drop of crime.”

None of which is to say that Ring doorbells are ineffective at deterring criminals or preventing burglaries, of course. Their popularity, especially when heavily subsidized by cities, is testament to the fact that many people feel insecure in their homes and enjoy the reassurance of a security device.

But, says Cuellar, the evidence so far is insufficient to prove any reduction in domestic burglaries, let alone the hefty declines or herd immunity the company claims. “It could be a very interesting study, if someone did it properly,” she says. “But I’m just afraid that it was not really done properly.”

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

The problem with plug-in hybrids? Their drivers.

Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.

Google DeepMind’s new generative model makes Super Mario–like games from scratch

Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.