Your TV knows you better than you think

Smarter sets could mean more niche programming and fewer annoying ads, says the TV industry. Privacy advocates worry consumers are being kept in the dark.

Your TV knows you better than you think
[Photo: burakkarademir/iStock]

Digital advertisers have it relatively easy. On the web, they know they can send you ads based on the products you’ve browsed, the searches you’ve made online, and the places you’ve been in the real world. And they know they’ll get precise data about which of those ads turn into site visits and purchases.


Traditionally, none of those options have been available to TV advertisers. They’ve bought ad time on shows that seem to appeal to broad target demographics—like women between 18 and 50—but they’ve had no way to only target consumers actually in the market for, say, a new car or mattress. They’ve also had limited ability to measure which ads really motivate people to buy.

“We’ve had Nielsen for years, which is a panel [of survey-takers],” says Jane Clarke, CEO of the Coalition for Innovative Media Measurement, an industry group with members like Disney, Google, and Univision. But Nielsen “is too small to do these kinds of big data matches on at enough of a level where you could really see anything going on.” (Nielsen did not respond to a request for an interview.)

That data is about to get a lot bigger. Thanks to smart TVs, the digitally connected sets that research firm IHS Markit estimates will make up 70% of televisions shipped this year, advertisers and TV networks are rapidly getting new insights into who’s watching what shows and ads for how long and whether they’re making purchases afterward. Marketers are already armed with consumer data like supermarket loyalty card records, mobile phone location data, personality profiles, or any other information held in advertisers’ own databases. Now companies can get data about your TV too, since many smart sets and the apps installed on them can track the shows you watch and the video games you play, often through services that deliver features like show recommendations.

“By virtue of our application that exists in millions of households, we have a really good amount of data on what people watch,” says Ashwin Navin, CEO of Samba TV, a San Francisco startup that makes apps for many popular smart TV brands.

Fewer, more tailored ads—and whatever coffee you like

Privacy advocates aren’t convinced this is a good thing. Lawmakers have expressed concerns too. Last month Democratic senators Edward Markey of Massachusetts and Richard Blumenthal of Connecticut urged the Federal Trade Commission to investigate smart TV privacy in the wake of reporting by The New York Times and others about Samba and the smart TV industry.


“The content consumers watch is private, and it should not be assumed that customers want companies to track and use information on their viewing habits,” they wrote. The senators called for companies to “comprehensively and concisely detail” who will have access to that data, how that data will be used, and what steps will be taken to protect it.

Typically, TV and app makers say they don’t collect your data unless you’ve opted-in to share it, and what data they share isn’t linked to any personally identifiable information, but to an identifier that connects to a wealth of other data about you. In any case, Navin and other TV techies generally like to emphasize all that you get in exchange for turning over your viewing data. Samba’s software, for instance, can recommend shows for you to watch based on what you’ve already seen. That’s a valuable service, says Navin, when a common industry statistic holds that Americans spend about 1.3 years of their lives flipping channels.

People in the smart TV industry also say more data-driven advertising could help niche programs that are typically hard to evaluate with traditional ratings, from special interest shows to non-English broadcasts: better metrics and valuable targeted ads could help them stay on the air. “You would presume it would help to foster small shows with particular types of audiences that maybe before would have a hard time monetizing,” says Clarke.

Data-gathering TVs can potentially also mean fewer minutes of commercials since broadcasters can charge a premium for more effective ads that produce more data. Better targeting could also mean ads more relevant to viewers and fewer repeat ads that get on people’s nerves. That’s especially true if cable companies and digital TV services continue to roll out ways for advertisers to beam specific commercials to specific sets of viewers, what the industry calls addressable TV. Then, two people watching the same show might see different ads just as they would see different ads on the same website.

Lance Neuhauser, CEO of Chicago ad data company 4C Insights, says industry insiders have even floated the idea of dynamic product placements. “You might see the coffee cup look like Dunkin’ Donuts, and I might see the coffee cup look like Starbucks,” he says. Of course, that’s precisely the kind of thing many consumers might find off-putting: a TV show that has different content when you watch it at a friend’s house instead of your own flies in the face of decades of experience with the boob tube.


‘That’s not what TVs did’

For advertisers and marketers, the power of “connected TV” data derives in part from the ability to join it with other data, like web histories or offline purchases. Data Plus Math, a TV analytics company based in Boston, offers a small piece of code, a so-called pixel, that marketers can drop onto their websites to help track when customers drop by and transact business; Data Plus Math can then link that online behavior with viewing data from smart TVs and other sources—like an IP address that’s shared by a TV and a computer—in order to determine the effectiveness of a particular ad.

Addressable TV ad spending in the U.S. is growing quickly but will remain a small portion of total spending for the foreseeable future, according to eMarketer.

“At the highest level, what gets me really excited and gets our customers really excited, it’s really showing that TV works—it’s that simple,” says John Hoctor, the company’s cofounder and CEO.

Legacy measurement firms have also rushed headlong into TV analytics. Nielsen and a subsidiary, Gracenote, said in April that its Automatic Content Recognition software, which can monitor all content on TVs—including any video and images transmitted through an HDMI cable— is installed in more than 40 million smart TVs across eight global electronics brands, including LG.

The data can get quite creepy too: In 2016, analytics giant ComScore partnered with former Trump campaign contractor Cambridge Analytica to merge the disgraced data firm’s “behavioral psychology” data with TV viewership data from 52,000 households. (Representatives for both companies declined to comment on the details of the partnership.)

Facebook has even explored better metrics for TV. Two of the company’s patents describe a technology that can “listen” to learn what TV shows and ads mobile phone users are watching, though the company has said it has no plans to deploy the tools.


While advanced and addressable TV ads have drawn nearly $6.7 billion in spending this year, double last year’s number, that is still a small portion of overall TV ads. But those numbers are likely to spike in the coming years. Last November, the FCC approved a new TV broadcast standard called ATSC 3.0, or Next Gen TV, that will make it easier for even local broadcasters to target ads and other content to consumers. TV makers and broadcasters are expected to begin supporting the standard this year, though there’s no timetable to phase out traditional broadcasts.

However, the rollout—the result of a 3-2 FCC vote—comes despite sharp dissent from the agency’s two Democratic commissioners over concerns related to switchover costs and privacy. FCC chairman Ajit Pai called the critics “naysayers” who would “stubbornly cling to the past.”

Such dismissals are unlikely to quell the privacy concerns. Even if tracking only takes place when users opt in, privacy advocates say many people aren’t even familiar with the idea of a TV that can monitor what you watch, and they’re often not reviewing those agreements in much detail when setting up their devices. (Typically, you can opt-out of a smart TV’s tracking features and still use it, provided you know which settings to click.)

Critics have warned that senior citizens, who are often targeted by scammers and dubious advertisers who see them as gullible and who can struggle to understand new technology, could be especially vulnerable.

“I don’t think the average consumer is very aware, to be honest,” says Jeremy Gillula, tech policy director at the Electronic Frontier Foundation. “People aren’t used to the idea that their TV is basically another computer or like their phone, where it collects data, primarily because, in the past, that’s not what TVs did.”


Last year, the FTC and New Jersey’s Office of the Attorney General entered a settlement with TV maker Vizio, which agreed to pay $2.2 million after the agencies said the company “installed software on its TVs to collect viewing data on 11 million consumer TVs without consumers’ knowledge or consent.” At the time, Vizio said it updated what it disclosed to consumers and emphasized its automated content recognition tools didn’t pair viewing data with personally identifiable information. Instead, it only measured data “in the ‘aggregate’ to create summary reports measuring viewing audiences or behaviors,” Vizio General Counsel Jerry Huang said in a statement. Vizio didn’t respond to an interview request from Fast Company.

Many TV and software makers say they don’t share viewer data directly linked to personally identifiable information. In its privacy policy, Samba says the data it shares includes “device IDs, advertising IDs, IP addresses, hashed/encrypted versions of your email address or other identifier, or pseudonymous IDs.” To join consumer data with viewership information, Data Plus Math relies on a so-called identity graph vendor, which can combine disparate sets of information for advertisers and ad-tech companies without having to reveal personal information where it’s not necessary. “We never touch any PII,” says Hoctor.

Vizio, which signed a deal late last year to provide data to Data Plus Math, says in its privacy policy it doesn’t permit its “data partners” to attempt to identify individual viewers, though they can be allowed to determine things like whether a device that shared an IP with the TV visited a particular website or physical location, or demographics including “gender or age range or income range” associated with a TV’s IP.

Not getting permission is ‘when you’re going to run into problems’

Smart TVs and their data are also potentially vulnerable to hackers who might find ways to use them to get entry to networks, leech bandwidth to launch denial of service attacks or even turn them into spying devices if they’re equipped with cameras or microphones. Last year, WikiLeaks released a trove of CIA documents that claimed the agency could turn TVs into surveillance devices, even if they appeared to be off. (Security experts suggest checking your privacy permissions, keeping your smart TV software updated, and physically covering any built-in webcams if they’re not in use.)

While consumers have arguably gotten used to being tracked as they surf the Web, other media habits have usually been seen as private. After the Washington City Paper reported on 1987 Supreme Court nominee Robert Bork’s video rental history, partly in response to Bork’s skepticism about Constitutional privacy rights, Congress passed a law, the Video Privacy Protection Act, protecting the unauthorized disclosure of those records. And provisions in the Patriot Act that were said to allow law enforcement access to library records drew widespread criticism during the George W. Bush administration.


Still, TV and ad-tech companies are optimistic they can find a way to deliver online-style data to advertisers without running afoul of privacy regulations or creeping out consumers. Generally, they say, that means explicitly getting consumers to give permission for tracking and targeting with a good understanding of the potential pros and cons.

“It’s taking a little bit of time to determine exactly what data points consumers people feel comfortable sharing,” Neuhauser, of 4C, says. “I think when the rules of permission-based marketing are not followed is when you’re going to run into problems.”

About the author

Steven Melendez is an independent journalist living in New Orleans.