Jan 18, 2018

Advanced advertising challenges the traditional limits of ads on linear television, and these are the professionals driving the change.

When Tim Leonard was getting his doctorate in neuroscience, he spent his days researching a cure for Alzheimer’s. He used a webcam to track the daily differences in patients’ eye movements after watching a TV clip, and monitored electrodes connected to the exposed brains of epileptic patients. Now, as a senior data scientist at Viacom, he continues to contribute to the collective understanding of human behavior and memory, applying his lab skills to uncover the content that viewers respond to and remember.

Leonard is part of a growing number of scientists finding new ways to apply science to business. They’re using their unique mix of skills—statistics, analysis, modeling, and computer programming—to capitalize on big data and influence business decisions. They are tapped by recruiters for hedge funds, healthcare conglomerates, and Silicon Valley start-ups.

At Viacom, they are a central part of the Advanced Advertising data team that has over 100 staffers, more than half of who have backgrounds in applied science, including physicists, neuroscientists, economists …even a geophysicist. These professionals are helping to transform the TV advertising industry so marketers can predict intent, target audiences, and measure results. They’re also applying neuroscience to understand how audiences respond to content, from uncovering how personality relates to content choices to using fMRI machines to maps brain activity and track emotional responses.

“The pitch I make to prospective employees is ‘Come teach us,’” says Bryson Gordon, Viacom’s executive vice president of advanced advertising. “You know things that nobody in this organization knows … that nobody in this industry knows. Imagine what you can impact.”

 

Tim Leonard, Ph.D., neuroscience

Senior Data Scientist

From: Toronto, Canada.

Watching: Working through Veep.

House rules: “My wife always says, ‘Don’t touch it.’ She knows if I take something apart it’ll be in pieces for a few days.”

Data scientists are valuable at Viacom and in the business sector overall in part because they have a unique way of looking at problems. Another reason: they are more in demand than ever before. Since the term “data scientist” was coined a decade ago, the number of companies hiring them has increased exponentially. Based on data from Indeed.com, the number of job postings for data scientists increased by 300% between 2013 and 2017. The trend is slated to continue. This year the United States could face a shortage of 140,000 to 190,000 people with deep analytical skills as well as 1.5 million managers and analysts with the know-how to use big data to make effective decisions, according to the McKinsey Global Institute, the research arm of the consulting firm.

Their data sifting skills are also incredibly valuable in ensuring the $178 billion global TV ad industry remains competitive amid shifting consumer behaviors and digital advertising growth (last year, for the first time, digital ad spend surpassed TV). Their talents make it possible to use big data to understand and predict consumer behavior and develop increasingly sophisticated audience targeting—with scale and brand safety that digital advertising can’t match.

“The common thread is the desire to understand patterns,” says Geneviève Smith, the director of product for Insight, a Silicon Valley-based fellowship program that teaches scientists how to transition from academia to business. “They want to know what causes those patterns and how you can use data from a number of different places to answer questions. It’s like being a detective.”

“I see many parallels of my work at Viacom, studying the interaction of fans from our brands and our advertising partners, with my training in physics,” says Matt Moocarme, a senior data scientist. “In fact, the way we model the strength of interactions between the social network of our different brands and our partner advertisers is based on the principles and methodologies built by statistical physicists.”

 

Matthew Moocarme, Ph.D., physics

Senior Data Scientist

From: Bournemouth, England.

Watching: Forged in Fire on The History Channel—a competition show in which master bladesmiths vie for a $10,000 prize.

Extracurriculars: “I love to combine musicology with deep learning and have trained neural nets to recognize patterns in music tablature. I trained a network on the guitar parts of Metallica’s The Black Album, so that I could generate music in that style.”

Using data to find a receptive audience

On Viacom’s advanced advertising team, these data scientists aggregate demographic, social, and viewing data to find trends about what attracts consumers to brands. They use that information to help clients create ads based on viewers’ interests and target them on television and social media.

“We make models internally of what kind of products and brands they would be most attracted to,”  says Sashi Marella, a senior data scientist, who holds a Ph.D. in theoretical neuroscience. “When clients come in for an exploratory meeting they can see for themselves in our apps what kind of brand or what kind of audience is attracted to their particular product. For example, everybody needs shoes, but not everybody wants Nike shoes. The people who do want Nikes are looking for a certain ethos in the products they buy. We match that to audiences on certain networks and distribution channels.”

It’s proven effective. A footwear retailer, for example, used Vantage, Viacom’s data-driven platform, to find and deliver ads to their audience at 162% better concentration than the rest of its TV average according to internal data. An auto brand experienced 20% higher brand favorability, 27% higher purchase intent, and 33% higher likelihood to recommend its vehicles to friends and family than non-segmented audiences. These metrics are the result of the types of end-to-end programs—from campaign ideation to audience segmentation to online and out-of-home sales measurement—that aren’t usually associated with TV.

Adeyemi Arogunmati applied his deep knowledge of algorithms—a result of his Ph.D. in geophysics—to prove BET is the most influential cable network on black Twitter (a digital community that engages in conversations around topics of importance to African Americans). BET recognized the opportunity to “go to an advertiser and say, ‘If you partner with us we can bring your product to black Twitter organically,’” says Arogunmati, a senior data scientist.

From there, Arogunmati created a tool that predicts hashtags that are about to explode in popularity on black Twitter. The goal is for marketers to connect hashtags with product placement seamlessly. “And because BET is influential within black Twitter they can guarantee the eyeballs,” he says.

Analyzing and predicting viewer reactions

As a kid in India, Preeti Vaidya enjoyed taking household items apart and putting them back together, eventually building her first computer at age eight. After stints at a bank and a market research firm, she came to Viacom because of the impact its programs made on her as a teen, when she watched the Viacom 18-operated Colors network. Its shows “spoke about so many cultural issues that were taboo in the Indian society,” says Vaidya, who holds a masters in computer science with a focus in machine learning.

 

Preeti Vaidya, M.S., computer science with a focus on machine learning

Data Engineer

From: India

Watching: The Daily Show with Trevor Noah

Singular pursuit:  “There’s no ambiguity in math. It’s a universal language. If you can prove this, than what you get out of it is a logical conclusion. And computer science is just like that. Just as we were building that computer, this fits here this fits here then oh, this last piece fits here and now we have a CPU. That’s what got me into this—finding the way things logically fit together.”

She’s part of an advertising science team in the Advanced Advertising department that’s building a platform to aggregate linear, social, and digital performance data across all networks and all advertisers. She’s developing the sentiment analysis tool, which analyzes comments and actions taken on social platforms. Once the platform is complete, users will be able to get a full view of an advertiser’s campaign metrics across platforms in a single dashboard.

Understanding the social response enables marketers to tell advertisers whether people liked the spot and most importantly, the actions users took after viewing it. Another use is to predict the number of impressions more accurately and clicks a campaign will get on social and digital when selling and pricing those campaigns.

Democratizing data science

With the push to integrate social and linear success metrics for marketers, the need to access performance information is greater than ever. For Fabio Luzzi’s data science and advanced analytics team, that meant creating a tool that allows anyone at Viacom to access data about viewers and users of their social platforms. The result of the multi-year project is Science Central. It’s a web platform that provides self-service data analytics. The key was making it as user-friendly as a smartphone app, so everyone from the CEO to a marketing manager can get data on-demand. The platform includes native apps created by Luzzi’s team and partner apps from branded collaborations.

Internal media planners use it to find associations among viewers for scheduling and promo placement. For example, they might learn that millennial fans of unscripted shows also like high-quality scripted programs such as Billions. If they’ve got a new reality show premiering, they know not to schedule it during a show that appeals to the same audience. They also use those associations to determine the best time slot to promote upcoming shows.

One of the apps on the platform, created by Nickelodeon, identifies social influencers, publishers, and talent that are associated with the brand. Internal marketers can use that information to identify opportunities for social promotion, casting, and ever public relations.

Before Science Central, it took days, sometimes weeks to access customized audience data, frequently provided by outside sources. Now it’s done in-house in seconds. Another advantage is its ability to track viewers’ habits over an extended time. External measurement companies typically change participants every few months, but Science Central’s machine learning algorithm allows users to track viewers over long periods of time.

“Our main goal is to make access to advanced analytics easy,” says Luzzi, Viacom’s vice president of data science, machine learning, and artificial intelligence. “It’s the democratization of data science.”

 

Diana Saafi, M.B.A. with a specialization in business analytics;

M.S. in data science (in-progress)

Senior Data Scientist

From: Virginia. She also spent part of her childhood in Tonga, a small island nation in the South Pacific.

Watching:  The Daily Show with Trevor Noah and the Korean dramas Shopping King Louie and The Heirs

Past life: I vetted potential reality series cast members, putting them through medical evaluations and background checks. Reality shows attract a lot of very colorful characters, so you’re not looking for some average Joe who has nothing to hide.

Connecting physics to fans

The proliferation of data from social media has created more ways to glean consumer insights. For instance, Leonard is analyzing Facebook posts across all Viacom properties for an entire year—about 5 million comments. He’s assigning a value to the emojis, words, and phrases. From there, he’s able to categorize comments, tweets, and GIFs based on emotions, helping marketers understand how viewers feel about content.

“It adds more color to the success story we can share with our advertisers, hoping they’ll continue to partner with us,” says Jennifer Perucki-Strapp, vice president of advertising science.

But Leonard is most excited about the potential eye trackers have in measuring brand awareness. He used trackers mounted on desktop computers extensively in his laboratory work to measure memory. At Viacom, he hopes to use them to measure brand awareness, which is traditionally done through surveys and focus groups.

His idea puts a webcam on a computer to track and map eye movement while watching an ad or integrated marketing clip. Then, show the same clip the next day. The differences in the two reveals what viewers remember. “You can correlate eye movement with when memories are occurring,” he says.

Viacom is in a strong position to do this, he says, because its content reaches 700 million subscribers all over the world. “This is something I would never have been able to do in a lab,” says Leonard.