Navigating the Linear/CTV Space Through A Currency-Grade Approach
Chief Research Officer of iSpot, Leslie Wood, shares the transformative insights gained from turning big data into a form of measurable currency
Leslie Wood is the Chief Research Officer for the TV measurement company iSpot. A data science leader and advertising expert, Wood leads currency initiatives across linear and streaming and oversees the data science department.
Wood has a PhD in machine learning and decades of experience in media. As the Chief Data Officer for NCSolutions (formerly Nielsen Catalina), she oversaw innovation and developed new approaches to using data for incremental measurement and targeting. Wood started at iSpot in July 2023 after a long-time partnership at Media Trust LLC and Leslie Wood Research, where she spearheaded research into internet reach and frequency, market mix modeling, radio listening patterns, and TV reach optimization, among other crucial industry issues for major advertisers.
The Continuum sat down with Leslie to discuss what it means to develop a currency, the measurement challenges that come with streaming TV, and how numbers always tell a story.
Your background is very different from that of most people in this industry. Can you tell us a little bit about that?
I’m a mathematician. My parents wanted me to be an artist. When I was growing up in New York City, there were a few public schools that made you apply for admission. I was actually one of the first girls to get into Stuyvesant, which is known for being academically rigorous, and I also got into Music & Art, which was the high school in the movie Fame. It should have been a slam dunk, which one I went to, but I knew that if I studied art, I would never keep up with math. That’s how I made my choice.
When I got to college, I chose anthropology and Egyptian mythology as my majors, but I had to work my way through school. Math, science, and statistics were always the easy As for me. So, in the last few weeks of school, I switched my major and graduated with a major in math and a minor in statistics and computer science.
My PhD was on single source data, which is what iSpot has; you get information on purchase and exposure from the same set of data. It was also in what was called expert systems, which is now called machine learning or AI.
You’ve spent most of your career in advertising and media. Tell us about that path and how it led you to iSpot.
I've always enjoyed the creative use of math and statistics. For the first half of my career, I built systems, analytics, and thought leadership for the ad industry. I'd be the white-label product and thought leader behind the scenes for all kinds of agencies and media companies like Nielsen, MRI-Simmons, and Arbitron.
A dozen years ago or so, I went to Nielsen-Catalina Solutions, a joint venture between the two organizations, to understand incremental sales. We looked at both targeting and attribution to determine how much sales increased and how much of that increase could be directly tied to advertising. To understand incremental sales, we used machine learning to build a causal model.
Now I’m at iSpot, a TV measurement company. My role for these first eight months or so has been developing a currency, but ultimately, the goal is to make iSpot the premier company for understanding the whole flywheel between creative, audiences, and outcomes.
We want to understand how advertising can work better and how brands can improve the quality of their advertising. We also want to help brands measure the business impact of both streaming and linear TV.
As a brand, you need the best creative, and you need to deliver that creative to the best audiences for your brand. And then you need to measure the outcome. It’s the full cycle, and it needs to be accurate.
“In the world of media, a currency is the measure that both the buyer and seller agree to use to transact a purchase.”
Can you explain the iSpot model and what you mean by currency?
In the world of media, a currency is the measure that both the buyer and seller agree to use to transact a purchase. Most people transact on audiences, but some look at outcomes or other measures. At one point, the Nielsen People Meter was a currency that both sides agreed on, but they had trouble keeping up with the changes in how consumers use media. Nothing we have now is giving the full picture, which is what our clients are asking us to provide.
To be currency grade, you need to look at impressions and reach for streaming and linear television together. You need to be able to break it down by households and people. You need to be able to look at national and local data so that advertisers with franchisees can build a national plan.
That was our goal, and we're getting there. We’ve been able to do things that no one else has been able to really address thus far because we started with these use cases.
The TV/streaming space has changed so much in the last five years. What is the biggest, most impactful change, in your opinion, and what has it meant for advertisers?
Streaming or CTV has just really exploded. Everybody started watching a lot more TV during the pandemic, and a lot of people have converted to streaming. That brought publishers, brands, and agencies into the sea of change. Now, they want to understand how much reach they get and how the viewing pieces overlap.
That’s why one of our goals is to measure the value of linear and streaming together. Earlier this year, we launched a streaming-centered dashboard that looks at devices to determine where people are watching and what services they’re using. Last week, we announced a new deal with Roku which is really going to help us do this. They bring 80 million television sets and have identity data and other good information about each television. This will help us expand what we can do and improve the accuracy of our data.
Are most people watching streaming with ads or without? How does this change the landscape for advertisers?
That's an interesting question. If you look back two years or so, most streaming was without advertising, and the audiences for commercials dramatically decreased. But the streaming services couldn't make money, so many of them decided to have a second tier. Consumers have a choice: Pay premium prices and have no ads or pay less or nothing, but get ads, like in linear TV. It turns out that many people would rather watch ads than pay for services. In the last year or so, we’ve been seeing that these ad-enabled platforms are being increasingly watched.
Obviously, this has been good for advertisers. It means more inventory is available for media placements, which, for simple supply-and-demand reasons, brings the price down. And advertisers can put a lot more on streaming because there's more content to put it in.
“Brands really want to tackle reach and reduce the waste. They want to map cross-platform delivery to their outcomes so that they have less waste and know they’re delivering the message to the right people.”
What are the challenges that brands are facing in terms of measuring TV audiences or performance?
We recently did a survey where we asked over 500 brands and their clients just this question. Overwhelmingly, they said that fragmentation is the single biggest challenge in today's marketplace. People are watching all over the place which makes it hard to buy a big audience.
Brands really want to tackle reach and reduce the waste. They want to map cross-platform delivery to their outcomes so that they have less waste and know they’re delivering their message to the right people.
Our job is to wrangle this fragmented field into a single measurement that people can track because if you can track it, you can actually improve and optimize against it. That’s one of the things iSpot can do very well because we can match streaming and linear to the IP address of a home’s WiFi which allows us to know in real-time, without modeling, what they’re watching.
Machine learning has made a huge impact on the performance measurement of TV, and everyone is talking about AI these days. Are you working on any projects in machine learning or AI right now that you can share?
Taking big data and turning it into a currency requires lots of machine learning models. The difference between a regular model and machine learning is that a regular model will assume a linear relationship. Think of a graph with a line going up or down, and everything is happening along that same line. But that’s not what real data looks like. Real data is all over the place, and you have to find all of those pieces. In our work, we need to understand the viewing habits and demographics for different kinds of people in different parts of the country on different devices. Machine learning is really good at finding those pieces.
We also have one project right now that relies on AI. We have an app called Tunity, which is very cool. If you’re in an airport or a bar and the TV is on, but you can’t hear it, you hold up the Tunity app to the screen, and it syncs to whatever is on. Now, you can hear it on your phone and use your own earbuds, and as a result, we have a measure of out-of-home viewing. We can add how many people are watching ads in these public places to our currency. The app itself relies on AI, where, within seconds, it looks at a screen, scans everywhere in the country for what’s on right that moment, and syncs to a specific program. That’s a complex AI job.
“Real data is all over the place, and you have to find all of those pieces.”
A lot of creatives these days are afraid that AI will take their jobs. Is this something you worry about?
I’m not worried about AI taking the job of creatives. It can’t make “art.” It makes something, but it’s not really art. What it is good at is polishing; you can take your work and pass it through and say, “Where can I improve it?” Or you can use it to move you forward. If I’m writing a book and come to a scene where I don’t know what to do next, I can tell AI the storyline and ask it for ideas to start the scene. It’s still going to be my book in my voice, but AI can help.
Before we go, what are your favorite shows on TV right now?
I really enjoy making things. I paint, and I did pottery for years and have done lots and lots of sewing. So, I tend to watch programs that are about making things. There’s a British series that called Portrait Artist of the Year, and a Landscape Artist of the Year version. There are a few programs about potters that I like, and there’s a great show about glassblowing.
It sounds like you did become an artist like your parents wanted after all. An artist and a mathematician.
I guess so. I think math can be very artistic. There are columns and rows of numbers, but it’s how you creatively put these things together that makes it into something. I've always been able to look at a set of numbers and see a story. And my job is to figure out how to make other people see the story that's there.
May 7, 2024
© 2024 The Continuum
