CBS's Radha Subramanyam Clarifies Today's Complex Data-Driven Ecosystem

When I first interviewed Radha Subramanyan in 2016, I described her as a "new breed of researcher, easily navigating between the two worlds of 'classical' research and the new discipline of data analytics." At the time, she was president of insights, research, and data analytics at iHeartMedia. Now that she has moved to television as chief research and analytics officer at CBS, she is also one of the leading industry figures in forming data insights.

Creating Content for an Evolving Audience
Subramanyam helms CBS Research at a time when the parameters of television are rapidly changing and expanding. Video transcends platforms and inhabits new devices; audiences are shifting traditional habits and developing new behaviors and purchasing patterns; and data tracks it all, unleashing a torrent of information that is at once illuminating and confusing. How do we best use data to form a full storyline of the viewer and their journey?

According to Subramanyam, TV still reigns supreme as one of the most popular content platforms. "It is front and center in people's lives," she says. "There are more shows being produced today than ever before and people are consuming more hours of quality content and premium video than they ever have."

This programming bounty is resulting in more hours spent on media and accelerating the speed with which viewers conveniently fit more video into their day. "The acceleration of time-shifting continues. Consumers watch what they want, when they want," Subramanyam explains, adding that this is leading to a rise in OTT and other platform options. "Players are coming in and out. It continues to evolve rapidly."

In response to the intense need for quality content across the ecosystem, CBS is producing plenty of premium content for its own properties, while also developing and producing hit shows for other networks and the major streaming service. "CBS is producing 94 shows today, up 20 percent from last year and 120 percent from five years ago," she says.

Evolving Research to Maintain Programming Dominance
When it comes to the primary focal point of any initiative, "it always starts with content," Subramanyam says, "because without great content you have nothing."

But content alone is not enough, she notes. You also need insightful research that will reveal the network's challenges and opportunities, strengths and weaknesses. Under Subramanyam's leadership, CBS is among the first to blend big data, new technology, and traditional research to to complement the development of new shows.

"We have always had a great legacy and investment in insights around content; specifically, testing around shows and other forms of creative. That remains a top priority," she says. "We also use every possible classic research technique," from dial testing to focus groups and neuroscience, "and we are all-in on artificial intelligence and machine learning."

Conducting Marketing Research to Understand Attribution
In addition to being a program creator, CBS is also a marketing organization and Subramanyam's department conducts a range of marketing research projects. "We are doubling down in a way that we have never done before to understand our campaigns across TV, digital, and social," she explains. "And we're generating attribution, not just for our clients and partners, but also for ourselves in our own marketing campaigns." Her work in marketing extends to insights from data-driven ad sales campaigns, as well.

Subramanyam's department measures more than the spots and dots of advertising. Ad creative is also an important consideration. "We know that without great creative and great spots, we won't have great results," she says. "So, we A/B and A/B/C test all of our spots all of the time. We optimize the creative with clients." From there, it is possible to measure the KPIs, such as awareness and intent to view. Subramanyam believes that by tweaking creative, brands can understand the immediate impact of an ad or campaign.

Maximizing Content Success by the Platform
Is content platform-specific? For Subramanyam, some rules apply across all great shows. "You need great characters, strong narrative arcs, and great talent," she asserts. But there are differences between networks, where one show will resonate on one network and not on another, "which may require brand-level tweaking."

That is within the premium video space. Outside of premium video, such as in digital and social, there has to be a different set of attributes according to the platform. "Individual talent needs to speak differently on Twitter versus Facebook versus Instagram," Subramanyam says. "And we use data to suggest what works best for what platform."

Advice for the Next Generation
Subramanyam and I share a similar outlook regarding a career in research. We both highly recommend it to today's college students. "I find it deeply fulfilling as a profession and it is an honor and a privilege to represent the actual customer to a brand and be a voice for that customer," she explains.

What we both love is that research today is creative and scientific, using both sides of the brain. Plus, "you have to be extremely nimble, be willing to change and course-correct as needed, and always be open to learning radically new skills," Subramanyam advises.

Her parting words to today's college students: "If that sounds like you, the world is your oyster. We will need far more analytics and insights professionals in the future than we need today."

This article first appeared in www.MediaVillage.com

TV Measurement in a Digital World. Let’s Discuss

In a highly fragmented media world, not only does the consumer feel at the cutting edge of media consumption, so does media measurement and attribution. How does the industry keep up? This dilemma was deconstructed at the Data Conference, part of TV Week NYC.

What is the Measurement Solution?

In searching for the solution to cross platform media measurement, the panel essentially described the problems in reaching a solution. Paul Lefort, Senior Vice President, Nielsen, noted, “No one solution will work in the long term.” But he believes that comparable metrics matter. “The interesting push in the past few months is the impressions-based approach,” he stated. “Impressions remove friction in process. If we only use ratings then we lose portions of our audience.”

For Radha Subramanyam, Chief Research and Analytics Officer, CBS, a hybrid approach is the best way to go with, “a combination of panel and other data.” But, she noted, there is still the issue of walled gardens. “What do you do with all of that data that sits outside?” she asked. “We need a holistic view because the market doesn’t care about walls or silos of data. Advertisers want the full picture.” The barrier to a holistic solution is a not technical one, she added, “It is will or leadership. We need to think bigger and have more cooperation. We need a fluid ecosystem. The tech easy but getting it done not easy.”

What About the Role of Panels?

Frank Comerford, Chief Revenue Officer and President of Commercial Operations, NBCU Owned Television Stations, listed a range of challenges. “Is a small panel accurately reflecting the audience we have?” he asked. “Is a big dataset reflecting our audience?” He explained that that if people aren’t in the right places in the data set, it will not produce the right information. In addition, “with a big data sample, there are walled gardens that are marking their own homework.”

Panels have both advantages and shortcomings. “Panels provide fundamental source of truth,” noted Subramanyam. “It allow customers and clients understand the households from which it is derived. It is also a source of diversity measurement and ensures an exact or accurate representation of those audiences. I believe in the fundamental need of panel to give us those things. But we also need the scale and stability of big data sets.”

“Panels can’t strive for behavioral but can strive for the demographic,” added Comerford.

Who Leads the Charge?

When it comes to deciding what is the most effective measurement and metric, who has the last word? 

Lefort believes that it is ultimately up to the advertiser, “if the advertiser is happy about what they have spent to get their results.” The good news is, “TV and digital are not as far apart as we pretend it is.”  They are complementary. “It is not TV versus Digital.”

Subramanyam concurred, “We are here to service the marketer and we need a healthy transparent market. There are lots of initiatives at work and a combination of them will lead to a new standard.” But, she added, “Consensus in standardization across media is essential. And people have to get comfortable in getting their data out there.”

Experimentation is Key

Measurement solutions should be a part of tracking attribution and there is a lot of experimentation going on in this area. At CBS, Subramanyam noted that, “We start by experimenting on ourselves and then offer the solutions to our marketers. Attribution, specifically in digital, is still in the early stages. It tends to be versions of last click. But there are lots of experiments.”

For Comerford the local markets tried a range of different attribution experiments with a “small sample on the buy in the local market. We track and see. We can prove that an ad that runs on TV is effective and multiplies the effect of local advertising. The last click did not necessarily get the sale but it might be the execution of sale. We have to see what led to that.”

The overall consensus is that the industry is now working together to find solutions. “The good news is that we are all talking,” concluded Subramanyam. All this results in greater cooperation and, “robust and positive discussion.”

 This article first appeared in www.MediaVillage.com

The 4As Examine Media Measurement Priorities at Advertising Week

There is strength in numbers. And I don’t just mean that in terms of all of the data being gathered and transacted upon in our industry today. I also mean it to suggest that we need to work together - from networks to agencies to the range of other media oriented businesses - to finally solve for cross platform measurement.

The conversation on cross platform measurement has been going on for over a decade through the work of several media organizations. But, frankly, these were often siloed efforts that gathered fleeting attention and struggled for cohesive industry action... until now. The push for an industry standard cross platform measurement is not only gaining momentum, it is also consolidating efforts across cooperating media entities.

As part of Advertising Week, the 4As hosted a panel titled “Media Measurement Priorities” that covered the joint efforts of leading industry entities to facilitate cross platform measurement and to decide, as an industry, what media measurement needs to look like in this new media environment. “What we have now really doesn’t fit the bill,” noted Louis Jones, Executive Vice President, Media and Data, 4As. 

He added that, “We need to have a collaborative point of view,” that also takes into account the needs of agencies. From there, the 4As set out to coordinate the efforts of companies and organizations working on the issue and published a whitepaper titled, Media Measurement Priorities,” as the first salvo.  

The paper set the stage for discussion of the most important priorities from an agency’s perspective. 

Here are the top five:

      1.       Unduplicated Reach

      2.       Currency

      3.       Short term versus long term

      4.       Walled garden and identity graphs

      5.       Attribution

Agency Perspective

Even for these top priorities, there may be flexibility in the solution. Take, for example, Currency. Historically, the TV industry has transacted on a strict set of metrics for currency. For Jonathan Steuer, Chief Research Officer, Omnicom Media Group, “We are in a world that is complicated enough that if everyone had access to the right underlying data, different partners could agree to trade on different metrics and that would be okay.” His point was that agencies seek impressions on specific target audiences and the way these impression are valued may vary across different platforms.

For Ed Gaffney, Managing Partner, Director of Implementation Research and Marketplace Analytics, GroupM, the currency just has to be well understood, transparent and stable. “We can have multiple currencies,” he explained, “We have them now,” with digital and TV and even within TV there are a range of metrics. “As long as everyone knows how they are counted, and can use that data, for both sellers and buyers, it works well.”

For Gaffney, Unduplicated Reach is critical to address waste. But the barrier, according to Steuer, is that the measurement currency for TV “is based on volumetrics and not real humans” and is delivered, “on the aggregate and not the individual. We need a census to tie together and understand device delivery to actual humans.”

Industry Perspective

In addition to agencies, there are businesses and organizations that are deeply involved in the measurement discussion. The MRC has been pivotal in establishing cross media measurement standards. George Ivie, Executive Director Media Ratings Council, explained that the MRC has been involved in a two year effort resulting in a brand new industry standard for video that was just released in early September. Three hundred 300 people and 175 companies participated. “There was a lot of discussion about measurement of exposure and how important it is as a building block to understand who saw your ads and how many times they saw it and the ability to de-duplicate,” he noted.

This standard provides the framework for equalizing the exposures across platforms and de-duplicating it across some general principles: Establishing a  common set of granularity, second to second level starting with counting impressions and then equalizing them as much as possible across the various video outlets, viewability, measurement and requiring invalid traffic and fraud filtering, the ability to measure people – demographics and targets – completes and duration weighted view of impressions so as to measure how long the viewable conditions persisted.

The reaction from the industry was both accepting and guarded. Radha Subramanyam, Chief Research and Analytics Officer, CBS, noted that measuring, “viewability is a good thing. Nobody wants invalid traffic. Duration is important. But the devil is in the details. Implementation versus theory – there is a big gap there.” Brian Smallwood, “Different advertisers are going to want to transact on different measures. This (MRC report) is one way of standardizing it but there are other parts of the ecosystem that want to trade or operate differently.”

If you ask me, an effort that has created the foundation for the trans- corporate cooperation today has been through CIMM. This organization has been working on universal content labeling to help stitch together content on various platforms and devices through Ad-ID and EIDR. Without a UPC-like code, there is no industry wide way to insure that content is accurately being captured wherever it airs. Jane Clarke, CEO and Managing Director, CIMM, noted.  “It is an evolving time in television and we don’t have a granular, nationally representative impressions-based TV measurement system in place right now,” she explained, because the data is siloed, behind walled gardens and not shared.

But, as there is strength in numbers, the first powerful step has now been taken. “The tech environment innovates. Technology improves. The standard is a first step in a long journey,” Ivie concluded.

This article first appeared in www.Mediapost.com

Advancing Cross Media Measurement. Takeaways from the CIMM Cross Platform Video Measurement and Data Summit

 

To those of us in the media world, the discussion of attribution and cross platform measurement has been going on for decades. Waiting for solutions has been painfully slow. But in the most recent CIMM Cross-Platform Video Measurement and Data Summit (now in its 8^th year) the big reveal was that not only has significant progress been made in attribution and cross platform measurement, feverish activity has been taking place behind the scenes for years. 

The reason is that it’s simply not easy to form new metrics and protocols between the need for consensus across all types of companies and the persistent evolution and expansion of technologies.

For Jane Clarke, CEO Managing Director of CIMM, 2018 was a year of significant advancements. At last year’s conference, CIMM announced a Measurement Manifesto to keep the industry focused on cross platform measurement. Her work on content labeling has been pivotal to stitching all pieces of video together across screens. This year Clarke announced the launch of TAXI Complete (Trackable Asset Cross Platform Identification) which is the creation of audio watermarks for content (EIDR) and ads (Ad-ID). It will, according to Clarke, “Advance cross platform video measurement and bring more measurement to TV.”

For me, the major takeaways were:

Ø  While content labeling gains traction, deduplication is pivotal to getting to a reliable and adoptable cross platform measurement currency.

If it can’t be measured it can’t be monetized and if it can’t be measured accurately, monetization will be flawed. Content labeling is continuing apace with more companies adopting the coding and the cost of entry to enroll has been lowered. One next step includes the advancement of deduplicated reach measurement to help accuracy. 

According to Beth Rockwood, Senior Vice President, Ad Sales Research, Turner, “Questions can be answered well with deduplicated reach. It’s an important tool and fits into context of marketing mix models and attribution.”

Ø  Measurement challenges will continue as we try to keep pace with evolving technologies and how consumers use them.

As Krishan Bhatia, Executive Vice President, Business Operations and Strategy, NBCU, noted, “The transformation of the consumption of video on a cross platform basis … has accelerated in the past 18-24 months. Devices are driving it (while) measurement has not caught up or stayed ahead of the curve.“

Evolving technologies such as voice assistants, “are making measurement more difficult,” Jack Smith, Chief Product Officer, Global Investment, GroupM pointed out. “It is happening in apps that are already walled gardens and operating systems are also intermediaries which can control pricing, ordering products and recommendations instead of having a direct relationship with the brand.” These, along with screens in self-driving cars, provide new viewing venues and experiences. As an industry we need to understand how it all works … and morphs over time. 

Ø  Privacy and Transparency need to be addressed. We can no longer kick the can down the road.

“Privacy and data issues will become more important,” stated to Laura Nathanson, Executive Vice President Revenue and Operations, Disney Advertising Sales. “Consumers will demand more privacy and more walls,” she added. But how privacy balances with transparency is still to be discussed. For others there is not enough transparency, at least when it comes to data labeling. 

To that end, part of the content labeling initiative includes  a data transparency label which points to aspects of the data and its collection such as how the data got created and where it came from. The careful balance between privacy and transparency, including the impact of GDPR in Europe on the U.S., require us to continue the discussion and create protocols. 

Ø  As an industry we need to work together to develop and establish adoptable measurements.

“We need to hit the pause button and up level the conversation,” stated Radha Subramanyam, Chief Research and Analytics Officer, CBS. She was referring to the clamor of competing voices around measurement today. “I love the innovation in measurement and the abundance of products and granular data,” she continued, “But are we any closer to making things make sense? My call is for a common sense framework. Stop the noise and see what we are really trying to solve for.”

This article first appeared in www.MediaVillage.com

 

The Coup D’etat in Media. Insights From the TV Data Summit

There has been a coup d’etat in media. Where Content used to be King, it has now been deposed and replaced by what Eric Shenk, Technical Director, Office of the CTO, Google Cloud Media noted at the recent TV Data Summit - “Audience is King.”

Why have audiences taken over the prime consideration in media? Part of the answer is the primacy of data in the consumer journey. With access to first, second and third party datasets, marketers and programmers can more easily and efficiently parse who is motivated, who is engaged and who is ready to transact. But how can we know which datasets are relevant, what might be missing and how to best analyze the results?

Here are some takeaways from the TV Data Summit:

The Impact of Data

The availability of data has led to a seismic philosophical change in the media pipeline. The focus used to be on delivering audiences to advertisers. But now, with audiences more in control of their viewing choices, companies need to consider what triggers their choice and attention.

For Radha Subramanyam, Chief Research and Analytics Officer, CBS, her company’s world class global content, brand championship in a safe environment with a passion for innovation is all tied together with cutting edge analytics supported by data. In looking at the ad marketplace today, she noted that data can be used by targeting demographics to gain mass reach and brand awareness, from data enabled TV to target optimization and in addressable for one to one messaging.

For Shenk, “Data is everything,” touching all aspects of the media business. But, currently, it is highly fragmented, residing in silos within companies and behind walled gardens with frenemy companies.  As best as one can, data needs to be collected, transformed, analyzed, visualized and then the best decisions must be made from the insights.

Consumers Are in Charge

Larry Allen, Vice President Ad Innovation and Programmatic, Turner, noted the shift in focus from the client to the consumer. “TV is under siege,” he warned, and said that while advertising strives to make the experience better for the client, “what if it is alienating the consumer?”  We need to “put them front and center.”

He explained that the current ad model is not working because there is too much frequency, not enough relevancy and too mass. “Really long ads can work depending what you are trying to achieve,” he noted and added, “We have new customer engagement metrics to improve our understanding of what is happening. We want to put ads in front of people that make sense. So they don't avoid them.” To that end, Turner embarked on consumer journey research to find out what builds affinity for content. Turner identified consumer content needs and the path they took to make decisions. “This can make things more contextually relevant in real time,” he stated, “and it can be leveraged. But it is hard to scale today and it is a long term horizon project.”

The Challenge of Measurement

The wide possibilities that data brings to the media business is offset by some of the challenges it creates. “Measurement to business outcomes has been challenging,” admitted Julie DeTraglia, Head of Research, Hulu. Tracking content across screens, calculating co-viewing on shared platforms and establishing an industry standard cross platform measurement are all worthy goals that require a joint industry effort between companies that often compete.

CIMM, now part of the ARF, has been focused on establishing a universal content labeling code to ads and programming. This initiative is beginning to gain traction. Jane Clarke, Managing Director and CEO, CIMM noted that there is, “a lot of support for data but sometimes data produces different results because there is no national dataset. We need to understand what is under the hood.”

Initiatives like OpenAP have enabled frenemy companies to work together to bind datasets together to segment audiences for sales.  For Haile Owusu, Senior Vice President, Analytics, Decisions and Data Sciences at Turner, the focus is on algorithmic output. “We forecast bulk audience flows and the ability to probe intra-show behavior, identifying components of the content for additional churn,” he explained. He is also working on the blending of linear and digital consumption by coupling Iinear TV viewing with viewing on devices and to messaging on the app.

Conclusion

Data and analytics, fueled by machine learning and artificial intelligence, will make the media business more efficient and targeted. If frenemy companies can work together, the opportunities offered by data will multiply and the challenges it presents will be subtracted. The TV Data Summit has put all of the innovative approaches out there. Now is the time for the industry to move forward.

This article first appeared in www.MediaVillage.com