A look at how Samba TV collects viewing data from 13.5M smart TVs in the US to make personalized show recommendations and enable ad targeting
The growing concern over online data and user privacy has been focused on tech giants like Facebook and devices like smartphones. Tweets: @sapna , @wolfiechristl , @sapna , @stevebellovin , @natashanyt , and @ennisnyt See also Mediagazer Tweets: Sapna Maheshwari / @sapna : when setting up a TV from one of 14 manufacturers, like Sony and Sharp, people may be clicking yes to a recommendation service that IDs nearly everything that appears on the screen (incl which prez debates they watched) for ads: https://mobile.nytimes.com/... Wolfie Christl / @wolfiechristl : Samba TV, a service for personalized TV recommendations, harvests second-by-second viewing data from 13.5 million #SmartTVs in the US, at individual level, largely without the full knowledge of consumers. Report by @sapna /w @kevinroose: https://www.nytimes.com/... Sapna Maheshwari / @sapna : advertisers can pay the company behind the software, called Samba TV, to then direct ads to the phones, laptops and tablets in a home based on their internet connection. experts say the trade off isn't made clear to consumers. Steven Bellovin / @stevebellovin : In America, TV watches you. In fact, it watches your other devices as well. https://nyti.ms/2lYaFMV Natasha Singer / @natashanyt : The latest device to embrace surveillance-based advertising? TV, writes @sapna, of companies that track your viewing habits, then slice and dice it to sell ads. “But consumers do not typically expect the so-called idiot box to be a savant.” http://nyti.ms/2KMYsWf Connor Ennis / @ennisnyt : On today's front page: another smart @sapna story about how company that seems to be about one thing (in this case, providing TV recommendations) is also hoovering up consumer data to help marketers better target ads https://nyti.ms/2lYaFMV See also Mediagazer
Context & Ripple Effects
Samba TV is not a rogue app but a recommendation service baked into setup flows on TVs from about 14 manufacturers including Sony and Sharp, so the consent moment is a single click most viewers never scrutinize. The lineage runs back to [[a:826260|Samsung's 2015 warning that its SmartTVs could transmit spoken conversations to third parties]] — the living room has been an open question since before Facebook became the default privacy villain.
What Samba adds is individual-level screen recognition: second-by-second viewing across 13.5M US homes, down to which presidential debates a household watched, sold to advertisers who can then reach the same home's phones, laptops and tablets. Later coverage shows where this leads — [[a:959783|advertisers by 2020 were actively seeking TV-based targeting while trying to avoid the online ad industry's privacy mistakes]], and the NYT's own experiment buying ads against 16 profile attributes demonstrated how granular those inferred profiles already were.
First-order effects
- Households that clicked yes during setup on Sony or Sharp sets are tracked at the individual level — nearly everything appearing on screen, including political debate viewership — feeding both show recommendations and ad profiles.
- Advertisers gain a bridge from the TV to the rest of the home: because Samba links devices through the household's internet connection, a viewing signal becomes targeting ammunition on phones, laptops and tablets.
Second-order effects
- TV manufacturers get a new revenue line beyond hardware margins — embedding opt-in data services at setup converts every shipped set into ad inventory, pressuring rivals like Samsung to monetize their own viewing telemetry rather than leave it unmonetized.
- The adjacent-device effect pulls streaming and mobile advertising into the TV data economy, since the same household profile now spans screens, raising the stakes for whoever controls the identity graph.
Third-order effects
- If the pattern holds, the living room becomes what the smartphone already is: a first-party data collection surface whose value justifies subsidizing hardware — a trajectory [[a:840073|Telly made explicit by giving away 500K TVs with a second screen devoted to constant advertising]].
- Privacy scrutiny built around Facebook and smartphones will have to extend to consumer-electronics defaults and setup-screen consent, the same expansion of scope seen when Kinsa was found selling 'illness data' from smart thermometers to Clorox — sensor-bearing consumer products as an unregulated data supply chain.
The trend: Smart TVs are becoming the ad industry's next first-party data layer, with consent compressed into setup clicks and hardware economics increasingly underwritten by viewer data.