John McDermott is content director at Samba, a media intelligence company whose technology tracks viewing habits across 50 million smart TVs and streaming devices.
His team crunched the numbers for every candidate for the next James Bond, measuring each actor’s household reach, age and gender skew, and pull with action fans.
The Oracle spoke with McDermott about what the numbers say about one of Polymarket’s biggest culture markets, and why the front-runner looks overvalued.
This interview has been edited for length. All answers are his own.
Where does your data come from?
Samba is a media intelligence company. We have a proprietary view into how people engage with broadcast, streaming and digital media, which lets us identify trends that are often difficult to see and make predictions about things that are culturally relevant.
What we’re most known for is technology in televisions. With a user’s permission, we can see what apps they’re using, what video comes through HDMI, what games they play on consoles. We look at the video on the screen and match it against a database we’ve built of sports, shows, movies, ads and games. It’s all anonymous. Users opt-in when they set up the TV, and we don’t know your name, address, phone number or email. We have an anonymous device ID, which we can geolocate from its IP address and append household demographics to.
How is that different from, say, Nielsen ratings?
Nielsen puts 40,000 physical meters in homes across the country. Our footprint is about 50 million devices. With 40,000 households, [Nielsen] can only look at a national level. We can break viewing down by state and sometimes by city, look at it internationally, and see niche streaming trends, like anime or Korean drama, much earlier.
One example: There was a question of whether the NFL could broaden its viewership by having Bad Bunny perform at halftime. There was a faction that didn’t like that idea at all, and nobody knew how much was just vitriol on social media versus what would actually affect viewership. We could see second by second how many TVs changed the channel during halftime, and how many came back for the third quarter.
It’s very mixed. Generally, you see peak usage during halftime. This year that was not the case. We saw double-digit attrition during halftime. That all came back in the third quarter, and it ended up being a pretty good game.