Men and Women Watch TV Differently. Here’s how to make money from that.

Most of us don't actually zone out in front of the TV. In fact, we give off all kinds of clues to what we really think about the shows we are watching.

Japanese startup, T-Vision Insights has come up with a way both to measure and to monetize those reactions.

Today we sit down with founder and CEO Yasushi Gunya and we talk about T-Vision's business and the future of advertising in video.

T-Vision Insights already has 100's of customers and is monitoring thousands of households both in Japan and the US and we dive into some of the differences in how different kinds of people watch and react to TV.

I guarantee some of the results will surprise you.

It's a great conversation, and I think you'll enjoy it.

Show Notes

How AI can determine viewer engagement Proof that women watched the super bowl more closely than men How men and women watch TV differently Which TV shows and commercials are most engaging The danger of advertising on the Walking Dead How privacy concerns are addressed Why it's hard to sell genuinely new innovations The most engaging parts of commercials Why starting a startup is not really risky in Japan

Links from the Founder

Everything you ever wanted to know about T-Vision Insights T-Vision Insight's ranking of the most engaging commercials in Japan Friend Yasushi on Facebook

Leave a comment Transcript Welcome to Disrupting Japan, straight talk from Japan’s most successful entrepreneurs.

I’m Tim Romero and thanks for joining me.

This episode is a fun one. You know, I’ve always considered watching TV to be a passive activity. I mean, aside from sleeping, it seems like the most passive thing you could spend your time doing. You zone out while entertainment is poured into your brain, but it turns out, that’s not quite the case. TV watchers are a subtly active bunch and as we watch, we give off all kinds of signals to indicate our opinion of what we are being shown.

Well, Yasushi Gunya, founder and CEO of T-Vision Insights has developed an unobtrusive way to measure viewers’ reactions to TV shows and to TV commercials. It’s already deployed in thousands of homes in Japan and in the United States, and the results are remarkable. T-Vision is already showing global 100 brands how consumers react to their commercials and to the TV shows that they air in, and they provide a data-driven approach to show what content is the most engaging and what kind of response it evokes, but what I think is even more interesting is that T-Vision’s data shows that we all engage with TV differently.

Adults engage differently than children, Americans watch differently than Japanese, and men watch very differently than women do. In fact, there’s a big difference between how men and women watch sports on TV, and I guarantee you, it’s not the difference you think it is.

But Yasushi tells that story much better than I can. So, let’s get right to the interview.

[pro_ad_display_adzone id="1404" info_text="Sponsored by" font_color="grey" ]

[Interview]

Tim: Cheers!

Yasushi: Cheers!

Tim: So, we are sitting here in the We Work office in this incredibly hot Tokyo afternoon with Yasushi Gunya of T-Vision Insights, so thanks for sitting down with us.

Yasushi: Thank you, Tim, and let’s cheer since we have beer here.

Tim: That tastes good on a hot day. So, T-Vision Insights measures the viewer’s reactions to TV shows and the commercials, but why don’t you explain basically how it works and what it is?

Yasushi: Okay, our core technology is AI-backed algorithm and we just inserted to a sensor and set the sensor on the top of TV. As a result, we can understand how people in front of the TV will react to the contents of TV, and actually, we have already said this kind of stuff to 3,000 households in US and Japan.

Tim: Okay, and we say ‘how they react,’ so is this a device sort of like Microsoft Kinect, or what are you measuring or what are you watching?

Yasushi: Yeah, actually, Microsoft Kinect is the sensor we are currently using but it’s not device-oriented. The core technology we developed is AI algorithm inside that and it’s applicable for all sensors.

Tim: Okay, so basically, what you’re trying to do is find a way to objectively measure the quality of the reaction for a TV commercial.

Yasushi: Yeah, different thing, different thing, because if you’re watching the TV, sometimes, you’re in front of TV but sometimes, even if you’re in front of the TV, you are just watching on your mobile phone or you are talking, or you have communication with others, and sometimes, even you go to the restroom.

Tim: Yeah, some people just turn the TV on and leave it on all day as they wander around the house.

Yasushi: Yeah, that’s actually right, that’s a reality, but the problem is, nobody understands how the reality is.

Tim: Okay. So, let’s dig into kind of the technology of this. So, what exactly are you measuring? Are you looking at facial expressions or movements, or…?

Yasushi: All kinds of these, but you can imagine that our core technology is the face recognition and body recognition, and we can see, okay, who in the front of the TV, at panel houses, we just get their pictures so we understand who they are: are they father, mother of the child but not a guest?

Tim: What determines whether someone is engaged with commercials? So, if I’m lying down on the sofa, can I still be engaged, are you looking at my facial features?

Yasushi: Yeah, if we find it by both the movement of face and the movement of body. If you are facing TV but you close your eyes, you cannot say you are engaged in the program, right? We just measure voice, and we have the combination, and we got the score. As a result, we define, okay, this guy is looking at the TV or this guy is not.

Tim: And can you tell what type of emotion you’re getting? So, for example, I might be looking at the TV and very engaged, but can you tell the difference between someone who’s laughing or smiling and someone who’s crying?

Yasushi: Yeah, we do get some emotions like neutral, smile, angry, or surprised, or maybe we don’t call it angry, we call it negative.

Tim: Okay, so those are the four basic categories?

Yasushi: Four basic categories, but once we get the data, it seems like okay, the audience just do not smile at a lot in front of the TV.

Tim: People have done research on this and a lot of times, when you’re watching TV, your brain state actually changes a bit. You’re in this zone where you’re not – I don’t know, you are not thinking or emotionally different.

Yasushi: Yeah, even there are some people who do this kind of research in neuroscience, and I think one big advantage of our data is, we are taking the natural reaction in the natural environment.

Tim: This year, you did an analysis of the viewers of the Super Bowl in the US.

Yasushi: Yeah, we do, we do.

Tim: And you had a really interesting finding about how men and women watched the game differently.

Yasushi: Yeah, that should say our difference. Not only in US but Japan, we do find much bigger difference between men and women than that between different ages.

Tim: So, in this case, the women were actually more engaged with the Super Bowl than the men were, right?

Yasushi: Yeah, that’s what we are surprised.

Tim: Yeah, me too. Are women just more engaged with television in general or was it the Super Bowl in particular?

Yasushi: Not in general, not in general. So, that’s what surprised us a lot. It’s true, for Super Bowl commercials, sometimes you are surprised, okay, the Super Bowl is very – I’m not generist – but yeah, people over here are saying, okay –

Tim: I don’t think it’s sexist to assume that men would be more engaged with the Super Bowl than women.

Yasushi: But sometimes, you can find the TV commercials or halftime shows will focused more on women.

Tim: I mean, but the game itself, were women more engaged or just …

Yasushi: Yeah, itself, itself.

Tim: Does that work across culture too? For example, Japanese soccer games.

Yasushi: Oh, that is maybe interesting. We haven’t, but in Japanese soccer game, right? I don’t want to say an inaccurate saying, so we haven’t done such thing in Japanese soccer game, but I imagine, we can find a lot of interesting things.

Tim: And I think that would be data that advertisers would be fascinated to learn, and I assume you can also tell the difference between adults and children and sort of different ages? What are the different demographics you can distinguish?

Yasushi: We have one year by one year, so actually, we have this kind of demo data from all of our panels, so from four-year-old to I think the oldest is 87.

Tim: But I mean, you’re AI can’t tell a difference between a four and a five-year-old, right?

Yasushi: That’s why we do gather the pictures of our panelists.

Tim: Oh, I see. So, when someone signs up to be part of this program, they register, these are the people in my family and Taro is five and Rie is 10.

Yasushi: Yeah, so that’s what we can give this kind of data – men and women, and each age, and even, we got a lot of information from them. So, we know their salaries, their family conditions.

Tim: Do you analyze the interaction between the people who are watching? For example, if I have a bunch of my friends over to watch a sporting event and have a party, can you tell that we are engaged with each other and the TV?

[pro_ad_display_adzone id="1653" info_text="Sponsored by" font_color="grey” ]

Yasushi: In our lab, actually, we are doing this kind of research, but not in general yet, but that’s a very interesting thing because sports always is big event for TV as well.

Tim: I imagine that sports would be one of the most engaging types of TV content.

Yasushi: Yeah. It is, it is,