Honored to have David Kenny of Nielsen joining us today. Thank you, David, and welcome. And we're going to do something special this afternoon and ask one of our longtime supporters and board members to introduce David. I'm referring to our good friend Adonis Hoffman, who also serves on the Institute's First Amendment Advisory Council. During his career of more than 30 years, Adonis served as counsel of the House Foreign Affairs Committee, held senior positions at the Federal Communications Commission, and taught at Georgetown University. He now has his own consulting firm, The Advisory Council LLC, where he provides advice to corporate CEOs and boards, financial institutions, and international organizations. Ladies and gentlemen, please join me in a warm welcome for our very own Adonis Hoffman. [Applause]
Thank you, Rick. It's great to be here with you today. As you mentioned, this is the last luncheon of the year. It's been a wonderful year, and I think it's only appropriate for us to take a few minutes to extend our appreciation to Rick Kapler, to Susanna CTO, and to our chairman Jack Wy for the wonderful job. [Applause] And before I introduce our speaker, I just want to acknowledge a couple of special friends who are here today, some distinguished guests. And if I do not mention your name, it's not that you're not distinguished, it's just that I don't have you on my list. But I want to start by welcoming our first-time visitor to the Media Institute, the distinguished chairwoman of the US Export-Import Bank, Ambassador R Lewis. [Applause] I was going to send a shout out to Mike O'Reilly, but he's not here, so Mike, wherever you are, this is for you, man. Very few people in the media can see the big picture. Many of us, in fact most of us, view media from a narrow perspective. Consumers want choice and convenience at a good price. Content providers want to reach as many consumers as possible. Advertisers mostly see costs and returns in the bottom line. And even regulators tend to focus somewhat narrowly on just the rules, often overlooking broader market realities. But the media ecosystem is vast, complex, diverse, dynamic, and driven by data. And few of us have the wherewithal, indeed the mandate and responsibility, to see it in its entirety, let alone to make sense of it all and translate that into something meaningful for all the key components within that ecosystem. Our guest speaker today is one of the very few individuals who does just that. David Kenny is the executive chairman of Nielsen Media. In his early career, he was a consultant at Bain and Company. In 1997, he became the CEO of Digitas, leading the company through an initial public offering in 2000 and its subsequent acquisition by Publicis Group in 2007. He was the president of Akamai Technologies, and in 2012 he became CEO of the Weather Company and served as its chairman of the board. When the Weather Channel was acquired by IBM in 2016, David became the senior vice president of IBM's Watson and their Cloud platform. He joined Nielsen as CEO in December 2018 at a pivotal time in the company's history. He saw the company public and took the company private, making a boatload of money, I'm sure, in the process. For someone who has been steeped in data his entire professional life, David has a remarkably broad worldview. He has a knowledge of the financial markets, having taken companies public and private, perhaps befitting his MBA from Harvard, and he has a keen awareness of the social impact not only of media but of large corporate actors on a global scale. If there's anybody at the epicenter of the media universe, it is our guest speaker today. So please join me in warmly welcoming Mr. David Kenny to the Media Institute. [Applause]
That was a completely unnecessary and long introduction, but it'll help you kind of understand where I might be coming from. I have these AI glasses, so they talk to you, but you don't need a teleprompter anymore, you can just wear your lesson. But if you heard from my career, all the time I was in advertising, I would complain about Nielsen ratings that the numbers were too high and they were giving credit for audience that wasn't there and they weren't counting when people were in the bathroom. And then I went and ran the Weather Channel and the Nielsen was terrible and its numbers were too low because we weren't counting everybody who was watching. And then I took this crazy job to see if I could help make it better, and it's been a great privilege to do that. And I would say the first thing you learned when you read Nielsen is this year, I think this is true, 83 of the top 100 programs in the last 12 months were National Football League games and the rest were the Olympics, and so it's the thing that works. And what you realize is that people, you know, we don't have division there, we don't have taste, people love sports. Part of what they love is that they know the rules, they can actually tell that the first down was a touchdown, and all of that works because the rules committee sets the rules and there are referees and judges and empires in all sports. And what I realized when you come to Nielsen is that's what you get to be: the referee. And no one ever loves the referee; they complain when they get it wrong and they take it for granted when they get it right. So for me, it's just a privilege for once in a very long time to have people who wanted to see me, so thank you, because that's what referees are. When I came in in 2018, part of what I was brought to the board was I had this history in machine learning and AI, trying to understand new technology. And so we did a reset that is just finishing now for a fundamental change in ratings. Because ratings are fundamentally market share. Arthur Nielsen, you know, 100 years ago, was the first person to use the phrase 'market share'. A.C. Nielsen was built on the belief that every market has scarcity and the way you compete and you know whether you're winning or losing, the way you make investment decisions, the way you can trust industry is to have a fundamental ground truth of market share. Which is why when there's a lot of different ways people can measure, they still come back to the ratings because it's consistent. And market share matters. But it was my belief that market share couldn't... originally market share was for broadcast, that was pretty easy to do at the time. There were three on television, there's a handful of networks, and you could do a number of choices against 48 day parts. And the big debate was much to do with cable, but we figured that out. We figured out DVRs. But streaming was going to change things. And so we did make the decision that we had to actually measure time, and we had to measure all time in video and all time in audio regardless of how it had come. It is our view that that's the right way to judge the market because that's the way the consumer views it. So we measured all screen time and we measured all listening time. What I would say is for all the billions of dollars in content and all the technology and all the new platforms that are out there, it's still a fixed market because there still is about 300 million people, that population grows 2% a year, there are still 24 hours a day, and neither Google nor Meta have added any to that. And so there's a little bit of double counting because people are doing two things at once now, including some of you right now, I see you. But that was the way to get at it. And so we shifted to measuring time, and it's a big change because outside of sports and live events, most of it's not viewed on a schedule when it's streaming, so day parts are kind of less relevant. They largely watch in the same time period, people largely watch after dinner, but we moved to a share of day model as well, which I think has been helpful. And we've learned a lot in that process. So when we started this and the first one came out in 2021, in 2020 it was like May of 2021, broadcast was 25% of the total, cable was 39%, streaming was 27%, and the rest was 9%. And that was a big shape because it meant a little over a quarter of everything was now on streaming. Just two years later, if I take the last month, streaming is now 41%, mostly the expensive cable which has fallen to 26%, broadcast is still 23%, and the rest is about 10% so social media game. So it's moving. But there's more to come because then we cut the data by age group. I'm going to just do under 35 and over 35. Over 35, a lot of folks in the room, it looks kind of similar: broadcast is 27%, cable is 31%, streaming is 34%, and there's about 7% on others. Under 35, streaming is 65%, broadcast is 7%, cable is 10%, and the rest is 20%. Those people under 35 are going to be over 35 every year, this is what's coming. So you can just follow the demographics. There's no evidence that it will change back as they age. I think what happens is people over 35 are getting more of the habits as it becomes easier. So I think it's pretty clear this is inevitable. And I think any of us who are trying to build policy or write rules or we think about ratings, it's impossible for me to understand how you approach the media and say 'I'd love to see more streaming people in this room' when under 35 is two-thirds of the time and growing. That's what's changed. Media is fundamentally a streaming on-demand service and the others are finding their way through, but it's shifting and it's shifting quickly. And I think that is a change that's complicated and people have business models and they're making decisions about their portfolio and all that needs to get sorted. But I also think it's really important our role in this is to do our best to keep giving people one truth set so they can make decisions as to where the consumer is spending their time, because ultimately all of us are trying to serve an audience. And having our best sense of it and our best sense of the competitive environment, so we know when they're not watching us what they are watching, if they're not getting news from us where are they getting it from. And we think this could be super helpful. We could do Q&A. A lot of people were asking me about how to do that. I would say it's why has this taken so long? It was a fundamental shift in the way we do it. So when you had 150 channels and 48 day parts, it was kind of a fixed system and you could get there with statistical analysis with a sample. So we would run around and get people in every county to represent that county and put meters in homes. And that worked. But now they're watching on phones and tablets and TVs and in lots of other locations than their home. The amount of people watching together is very small outside of sports; people are largely watching individually. So we had to get to individual meters that people actually put in their pockets that they carry around, and more importantly we had to bring as much data in as we could because every TV that's connected to the internet, every car that's connected to the internet for audio gives us data. So we're bringing in a massive amount of data from all the players, cleaning it, correcting it, using the panel to apply demographics and take out bots and make sure you get it right. But it's been a fundamental shift in the way that works, and it's been a worthwhile investment because I think we all needed to know that. So my hope for all of you is using ratings or something else, making sure you define the market that broad way, I think it's the only way to make the right decisions. Then what's next? And I'm here now so I've got my CEO. I think where we have to figure out next is AI. And you mentioned I've been around machine learning for a long time. AI is so much easier to use when everything is stream, so we will just find so much more that is, whether it be news or entertainment content, that is going to be produced with AI. We're all excited about the models, we're all excited about the breakthrough of GPT. But here's the thing about models: those are kind of the engine, but data is the fuel. And data, you can't just put crude oil in an engine, does it work? It's really important that the data, like fuel, be refined, be consistent, be something that knows how to use. It's also really important that it be audited so that we can make sure that it actually came from a true source. A lot of the problem I have with AI right now is that they ran out of data to train OpenAI, so now they have synthetic data, so the AI makes its own data. It's a little bit off, and then the next time it's a little bit more off, and it just keeps drifting. And unfortunately it drifts in the bias of the program. We saw this in weather that when we did weather models to put them on a phone, the first ones were really good and then it kind of drifted until you had empirical evidence. So you still needed meters, you still needed numbers, barometers, thermometers. And in this case, I still think you need empirical evidence of the data about consumers that you're all going to use to feature models in each other's models. And I don't know that this is the work of Nielsen, but if I were an institute committed to people trusting media, I would spend time on data principles. Because I know we're in Washington, but I generally think the government responds late to technology industry, and self-regulation tends to respond faster. I think the urgent need in my ask for you to consider is what should be generally accepted data principles in the media business. And the analogy I'll give you is the reason the financial markets work, the reason companies buy and sell at the right price, is at the foundation we have generally accepted accounting principles. And we all kind of agree what's revenue, what's cost, what even done. Yes, there can be non-GAAP measures and people try to slip those through, but GAAP measures are still overused and you can trust it. Now that data is in, particularly customer data is the fuel, what are generally accepted data principles? And as AI is going to be useful in producing content, it's going to be useful in producing news and information, what are generally accepted data principles to know that the data has a truth set and can be audited the way financial statements can be audited? Because I think that the data assets in all of your companies are as valuable to you as the financial assets. And certainly what we're trying to figure out is our role in making that. But I think it's you know we can do ratings because we've got a Media Ratings Council and there's an agreement on how to measure, but we need this for data. We need it urgently because AI is going to move at an exponential rate into virtually every business that we're in. So those were the couple of messages I thought would be useful to get the conversation started. And I know people have lots of opinions, so I'll take opinions or I'll take opinions as questions or...
Yes sir, so when you calculate streaming, does that include like TV or some of these over-the-top services, or how do you break those out into your categories that you gave?
So they, that we do, we do measure all of them. So this is the thing that I think was a bit of a chunk. So the number one distributor or network is actually YouTube, more than Netflix. And so it's about 11% of all time consumed, and if I go to those under 35, it's over 20%, over a fifth of the time. So we started with time, so it was over-the-top, FAST channels, streaming platforms on the video side, and similar on the other side. That's time.
Yes ma'am, my name is Dan Work at a research service for Congress. Some of also cover competition issues and use Nielsen. Yes, well, yes, getting congressed by your service is really hard. She was a champion. Speaking of what denominator regulators use when they're looking at local market of broadcast TV stations or other, I guess say broadcast household market time periods demographic, so how do you reconcile that?
No, it's a good question. My belief is you should start with a measurement of time. DMA is useful because it's where broadcast reaches and they have to reach us in a certain way. But within that geography, I think if you're not looking at the total picture, you really risk making rules around the 7% that's still broadcast. And if your goal in competition is to serve the market, then I think you want to know everything you know that they're receiving from everywhere. So I would do that. There you probably have to look at both levels, but to me, I think total time of total people. I also think you know Nielsen was built on households. Households have a limitation in that households were established at the time when most people watch TV as a household. That is not true at all anymore. The cable companies connect routers, those connect to Wi-Fi, and that means everybody can have their own TV. And so this is why even our meters are moving to a portable personal meter because the people level is just super important to get right. So it's a shift, and you're going to want to compare before and after, but I think I'm glad you asked the question. I think getting about that is going to be pretty important because otherwise you're defining share of buggy whips. Yeah, exactly.
Okay, yes sir. How are you preparing for the generation beyond? Gen Z people are growing up. You just sort of the idea of multitasking. I wonder they're doing it and doing it. And qualitative things like you know CBA with partial attention. How do you get to audiences that aren't or doing something else?
It's why the value of sports keeps going up because there's less multitasking in that. I think you have to for an advertiser you definitely have to try to understand that. We're just measuring it. The issue is how do we attribute when they're doing two or three things at once? But we are trying to spend more time on that by what do they recall? There are ways to look at where the eyes are, that often what screen they're actually looking at tends to be the one they're most focused on. But it's a really important area to hear, and I think it's going to be even more true. This was reading my emails to me, so even as that becomes more pervasive, everything is going to have... pardon? How are's? We will do qualitative research to put factors on top of it, but probably in a you know as a currency is what we've always done with outcome, which is there's a primary measure which is absolute time, and then there's a secondary measure which is more qualitative with a range around it because you're never going to be as precise on that one. And I think these things, that's how we're going to enable it. I think the other thing we're working on at Nielsen is to make sure that our data is more fluid and able to be used with other systems. Now the whole infrastructure of buying and selling in media is got to be overhauled to be more modern, but ultimately we need a system which makes it easier to adjust for those things in a way that both sides can trust.
Yes sir, you mentioned data principles. Can you give us an idea of some areas where data principles come into play, an example of what a data principle might look like?
