Day 1 – Sept 20 | Day 2 – Sept 21 | Day 3 – Sept 22
Tuesday, September 21
WELCOME
Scott McDonald, Ph.D. — CEO and President, ARF
KEYNOTE FIRESIDE CHAT
Leading the Next Frontier of Media Disruption
Scott McDonald, Ph.D. — CEO and President, ARF
Summary:
Scott introduced Marc Pritchard to the ARF stage with a discussion about the most important marketing and advertising measurement challenges in the industry and Procter & Gamble’s priorities in approaching solutions. While the perpetual fragmentation of the media ecosystem is complicating accurate measurement across platforms, advertisers still have the same simple questions, according to Marc. The consumer is the center of it all – how many people are advertisers reaching, how often are they reaching them, and how effective and efficient is the advertising itself?
In a call to action for all marketers, Marc urged the industry to innovate a consumer-centric measurement that focuses on the most important outcome for advertisers: the superior consumer experience and the communication of that experience. While still trying to reach as many people as possible with greater precision, the P&G exec believes measurement of that experience has to answer three crucial questions:
- Is it an objective measurement with an independent, third-party, accredited verification of that reach, frequency, effectiveness & efficiency?
- Is it transparent , with a clear understanding of shared rules?
- Is it complete for the consumer’s experience across multiple platforms, factoring for different experiences and context?
The following are edited highlights from the conversation:
- It is incumbent on marketers to lead the way in terms of measurement – with work from within, inclusive of publishers and platforms to broadcasters, and led by the trade associations – to achieve an objective view. This requires innovation from both individual companies and measurement entities to achieve an understanding of consumer behavior and experience.
- It all comes back to the consumer when advertisers’ needs for identity solutions might contradict ensuring consumer privacy and content choices. Marc recommended two key approaches:
- Focus on 1st party data. Give consumers the privacy they want and the value that they need so they are willing to exchange their data with permission. In an example from Pampers, Marc cited how the brand’s 1 to 1 consumer engagement delivered a value exchange that provided P&G with a deep understanding of those consumers and what they needed.
- There is not one solution, but there are some good solutions that allow mass reach with greater precision. Not on a 1 to 1 basis, not generic A18-49, but more of a precise cohort approach that enables the ability to access relevant data.
- Measurement should help advertisers recognize that every medium has a different context so they can place ads knowing their effectiveness in that context with the ability to remove excessive frequency.
- Consumer experience can be a poor experience: Seven out of ten consumers are annoyed by ad frequency and 53% of consumers feel it hurts brand equity if they see the same ad over and over again. A good experience for consumers is a valuable experience.
AGENCY LEADER INTERVIEW
Getting the Most Out of Data: An Agency Perspective
Learn how one agency leader and her team looks at consumers’ entire spectrum across multiple touch points, products and services and threads those journeys together. How does that understanding leads to correct measurement? To guaranteed outcomes for clients’ businesses?
Lisa Giacosa — President, Global Head of Data, Technology, Analytics & Insights, Spark Foundry
Jane Clarke — CEO, Managing Director, CIMM
Summary:
Jane Clarke, CEO, Managing Director, CIMM, interviewed Lisa Giacosa, President, Global Head of Data, Technology, Analytics & Insights, Spark Foundry. Below are edited highlights from the conversation:
- Understanding human beings, creating the right brand experience for them, and building their emotional connection to brands is essential for a business outcome, Lisa noted. Start with the audience, the insights, and feelings of the people. How do they engage with the brand, what is their receptivity, through to measuring impacts? Was the campaign effective? Understand the triggers that drove the outcomes, such as brand engagement, trust, or perception.
- Focus on the business impacts you want. There is a plethora of data, too much data. Focus on the drivers of business outcomes.
- The big challenge is the disparate data sources and their high cost. “Juice is not worth the squeeze.” Lisa suggested harnessing the power of your data and strategic partners, although stitching together the data and platforms of multiple partners is challenging.
- Fascinating data and strategic modelling drive the full picture to understand and engage with the consumer. However, transparency, privacy, compliance, and regulations are important considerations.
- Humanizing data means understanding consumer behavior. It is about how consumers actually behave, rather than what they say. This builds a picture of the consumer and connects the dots leading to an understanding of what stimulates the behavior. To achieve an emotional connection with brands, more than flat data and exposure details are needed. Data is only useful if you distill it to create a story.
- New approaches to understanding human behavior and media planning were developed at Kimberly-Clark during Lisa’s time as a marketer at the company. A lot of behavioral science research was done, and she has carried those learnings with her.
- Look at behavior, how did the consumer react? Body language vs. verbal responses have different meanings. These insights lead to more creative relevancy, and receptivity from the media perspective.
- Jane and Lisa agreed that human understanding still has a critical role to play. Combine best of human insights + best of data to understand human behavior. This represents art and science at its best.
- Brand equity moves much more quickly now. Social media and influencers have been impacting brand equity for the last 5-10 years. Word-of-mouth travels rapidly and impacts brand equity. More recently, the impact of COVID needs to be factored in. Spark Foundry has developed a tool to measure brand equity in real time.
- We need to come together as an industry to provide a fuller measurement picture. A one size fits all approach will not work. There’s lots of work for the industry to do over the next couple of years.
COO INTERVIEW
Perspectives on the Most Important Consumer and Video Marketplace Changes
Karthik Rao — COO, Nielsen
Interviewer: Scott McDonald, Ph.D. — President & CEO, ARF
Summary:
The major challenge to measurement is the vast amount of content to measure and the ability to deduplicate across the many screens. Inclusiveness is the second major challenge – showing the face of America. But this is an opportunity also. The third is chasing identity and privacy at the same time.
Karthik felt that interoperability will have to be a core piece of the identity future. Nielsen built its identity graph in 2020 – it is necessary for coverage of the populations. There can be multiple ID systems dedicated to different purpose – targeting systems are for scale, measurement for coverage.
What will the role of panels be in the future with it being more difficult to recruit them and service them during a pandemic? Nielsen suggested it has been using big data for a long time and foresees a role for both in future measurement. Karthik announced that in January, Nielsen will be showing impact data combining ACR, STB, and panel data.
The census is one data set that serves as a truth set that is used to calibrate ratings, but Nielsen will be looking at other datasets to use for calibration. This may be buy data for outcome based analysis. Nielsen is a big believer in the WFA initiative. It also believes it can create cross-platform working with multiple walled gardens.
Key takeaways:
- Challenges are the vast content viewed on multiple portable devices, inclusiveness of the population, and the balance between identity and privacy.
- Panels and big date each have their own strong points: panels are directly observable – you simply know much more about the home while big data may be less representation, but its size affords stability.
- Nielsen will be showing impact data based on ACR, STB, and panels in January.
- Nielsen supports the WFA initiative and interoperability of measures moving forward.
SPEED PAPERS
Unlocking GenZ on Social Video
Kate Ginsburg — Head of Content & Product Marketing, Tubular Labs
Summary:
There are 500 hours of content uploaded to YouTube every minute. TikTok is the most downloaded social app. According to the Global Video Measurement Alliance, there was a 55% growth in total time spent on consumption of social video. Gen Z spends more time with social video than any other demographic. Kate Ginsburg, Tubular Labs, explored this coveted demo’s social video viewership, e-commerce consumer behaviors, and the opportunities for brands. Social video needs to a part of the marketing strategy. It is where brands and advertisers must be to engage with Gen Z.
Key takeaways:
- Social video is the best reflection of what is happening in culture, according to Kate. Advertisers, brands, and creators should identify opportunities to influence culture and increase their relevance to Gen Z on social video.
- Gen Z spent 82% of their time on social video in July 2021 with influencer-created content, 10% more than any other demographic. BuzzFeed has been successful engaging with Gen Z audiences.
- The keys needed to unlock social video are categories, audiences, and partners.
- The shorter the video content, the better. Content under one minute resonates best, and is consumed most, by Gen Z.
- Gen Z values authenticity and being at the pulse of culture. Viral TikTok, MrBeast, and Cam Casey have large viewership and high levels of engagement.
- Genres heavily watch by Gen Z audiences are gaming, entertainment, music & dance, and people & blogs.
- Brands should recognize creators and genres for optimal partnerships, including gaming brands and influencers.
- Associating social view behaviors and preferences with e-commerce purchases can determine the right kind of partnerships for advertisers, publishers, and creators looking to expand their audiences.
United TV and Digital Measurement: Growing your Brand’s ROI With Today’s Integrated Reporting
Alexander Knudsen — VP, Solutions, Amobee
Summary:
Amobee’s Alexander Knudsen covered the benefits of cross-screen measurement in a fragmented media landscape and the insights it provides for planning and buying teams with a client case study from Toyota and H&L Partners. Addressing marketers’ key issues with a converged planning approach, Alexander detailed how cross-screen analysis worked with Amobee’s 4Screen Experience measurement product as well as how Toyota used the resulting insights to optimize in real-time.
Key takeaways:
- Cross-screen reporting not only provided new insights and analyses, but also brought traditional investment and digital teams closer together.
- TV-focused channel overlap shows traditional planning teams how digital is supporting their campaign and how optimal reach and frequency are being driven.
CONCURRENT SESSIONS – WINNING PAPERS & SOLUTIONS
Television Disrupted
Diversity (or lack thereof) in TV
Rebecca Fine — Associate Director, Research & Media Insights, Samba TV
Kristen Hsu — Senior Research Manager, Samba TV
Summary:
Samba’s research study on the lack of diversity in TV sought to understand the state of representation in popular TV by asking who was watching content with diverse casts, whether the show actors were representative of the U.S. population, and if people were more likely to watch TV shows where they see themselves represented.
Although the results proving TV shows’ lack of diversity were not surprising, the value of diverse audiences is growing in tandem with these populations, putting advertisers on notice to improve the status quo.
Samba devised a study of the top 50 viewed shows from Feb 2021 which combined cast data from IMDB and Rotten Tomatoes with Samba’s research panel (3M HH) and ACR data. The researchers were able to analyze the relationship between diversity in shows and the viewers that watched them against U.S. Census data.
The study found positive correlations between higher levels in diversity programming and diversity of viewers to conclude that there is a relationship between diverse viewers watching shows with higher levels of diversity.
- Correlations by Ethnicity: Black viewers had the stronger relationship to diverse shows, Hispanics followed, with no correlation for Asian viewers as there was very little representation, and a negative correlation for white viewers.
- Cast Diversity Findings: 76% of lead actors were white, where 24% were black (17%), Hispanic (3%), Asian (3%) or another ethnicity. Compared head to head with the U.S. Census, the significant gap between the percentages illustrated how far off the Top TV cast compositions were in relation to the U.S. population, where 61% were white and 6% were Asian.
- Top Shows Cast Breakdown: 97% White leads; 53% Black leads; 14% Hispanic leads; 11% Asian leads
Key takeaways:
- By 2065, over 50% of the U.S. will identify as non-White – “minorities” will soon become the majority.
- Diversity on TV is not representative of the U.S. population.
- Diverse households are most likely to watch shows with diverse casts, with a positive correlation among diverse households for programs that feature characters that look like them.
- Don’t assume you are reaching and engaging a diverse target audience through popular shows. Non-white audiences are more likely to watch diverse programming.
- Shows with diverse casting represent valuable airtime with engaged audiences: Include representative shows on your media plan. Relevancy of content and ads is important.
Assuring Research Integrity in Data-Driven TV
Peter Doe — Chief Research Officer, Xandr
Summary:
Xandr’s Peter Doe reinforced the omnipresence of bias in TV measurement as he outlined four key areas of bias in assessing DirecTV’s (DTV) set-top box (STB) data for its national data-driven linear TV advertising. Noting DTV’s relatively low sampling size (7M STB homes) has a high level of bias when measuring for national TV viewing, Peter provided a top-line overview of Xandr’s viewership data methodology relevant to advertisers and marketers working with big datasets.
Four key areas of potential bias with correction approach:
- TV On/Off Model Validation – 40 % if uncorrected
- Matched DTV HH to TVision HH panel data and quantified using model efficiency concept.
- Network availability – 50% if uncorrected
- Compared DTV to Tru TV ratings with SNL Kagan data and addressed network coverage and audience change across multiple networks.
- Projection (DirecTV customers to U.S.) – 10% if uncorrected
- Controlled for initial projection categories and more detailed projections that over-represented audience demos.
- Validating Experian Demo Classifications – 20% if uncorrected
- Corrected HH size against DTV, TVision, and Experian matching homes.
Key takeaways:
- Bias is ever-present in research.
- Panels may be selected to be representative, but they are not immune from bias.
- Large sample sizes reduce sampling error, but they do not guarantee less bias.
Neuroscience Got to the Heart of Esports Audiences
Jonathan Stringfield — VP and Global Head of Business Marketing, Measurement and Insights, Activision Blizzard
Devora Rogers — Chief Strategy Officer, Alter
Summary:
This presentation, based on research conducted jointly by Activision Blizzard and Alter, looks at the power of gaming with emphasis on ROI. Jonathan provided the background of esports, which he defined as professional gaming organized across teams or leagues with a range of game types. Professional e-gaming has grown rapidly to 436 million global viewers as of 2020 and is a highly fragmented marketplace.
Understanding esports fans is no more intuitive than understanding fans of traditional sports. Because esports is relatively new, deep insights across the fanbase are few; nevertheless, there is a continuous need to be able to tell a story about fans and their mindsets, not only in terms of their breadth across audiences, but with depth into the individual fan.
Devora described the insights challenge as one that is complex and evolving and therefore cannot be solved with standard research approaches. They developed a robust multi-modal process known as immersion neuroscience (i.e., second by second response to the content users are engaging with), quantitative ad effectiveness studies, and qualitative interview sessions to get a holistic view of esports and traditional sports viewers.
Key takeaways:
- Esports fans are more engaged (i.e., receptive to both content and ads) at levels that exceed fans of traditional sports, especially at higher levels of fandom.
- Accordingly, esports sustain attention. During sponsorship ads, immersion actually increases for esports fans (it decreases for traditional sports viewers).
- Esports drives higher brand affinity for advertisers. Both brand perceptions and favorability are higher for brands advertising on or sponsoring esports content.
Cross-Platform: Measurement & Identity
Wearables: The Future of Media Measurement
Erin Wittkowski — Principal Data Scientist, Nielsen
Summary:
Nielsen uses the Portable People Meter (PPM) to passively measure audiences for radio and to enhance its national and local TV ratings. It is now testing a new wearable device to replace the PPM to measure audio and to replace button-pushing to measure person-level television audiences. The new wearable is intended to enable cross-platform measurement of audio and video. Two other pieces of equipment accompany the device – a beacon to allow Nielsen to distinguish between in-home and out-of-home exposure and a data hub to transmit the data back to Nielsen. Another component of the wearable system is a mobile app which will enable engagement and communication with Nielsen and can also be used to transmit data back to Nielsen.
The device has been tested in a lab setting, in qualitative research, and among friends and family of employees. Nielsen followed these initial tests with a “dual-carry” test among 500 former PPM panelists who were asked to carry both the PPM and the new wearable. Then, Nielsen carried out a “wearable-only” test with 400 former PPM households. On the whole, the tests were successful. Nielsen therefore launched a “live” test with 3,000 current PPM panelists.
Key takeaways:
- The PPM is being phased out because it relies on no-longer-supported 2G/3G technology.
- The new wearable is smaller than the PPM and can be worn as a belt clip, a wristband, or a pendant.
- In the dual-carry test, the PPM and the wearable yielded nearly identical average quarter-hour audio listening and video viewing levels overall.
- The average amounts of time that the devices were undocked and carried were higher for the wearable than for the PPM. Audio in-tab rates were the same for the wearable as the PPM. TV in-tab rates were higher overall for the wearable than for the PPM, except among those 18 to 34. In a post-test survey, participants expressed a preference for the wearable over the PPM.
- In the wearable-only test, 53% of the smartphone owners installed the app, but an app was installed by at least one person in 67% of the test households.
- The wearables are smaller and so need to be charged more frequently and are easier to lose. So, Nielsen is testing various noises that can be emitted to help find them.
How Truthscores Can Significantly Improve the Accuracy of Addressable Marketing
Joel Rubinson — President, Rubinson Partners, Inc.
Aaron Fetters — Head, Client Development, Truthset
Kathryn Barnitt — Statistician, Truthset
Summary:
Truthset was founded 18 months ago to address the concern that attributes in Big Datasets used to define target audiences were not always accurate. Aaron reported that a female audience segment from a data provider that Truthset had examined turned out to be only 58% female. Truthset receives data on 2.6 billion IDs from 19 data providers each quarter and provides them with “truth scores” for the demographic attributes of each ID. The truth scores, which range from 0 to 1, represent Truthset’s estimate of the probability that an assigned attribute is correct.
Truth Scores can be used by marketers to target IDs with varying levels of probable accuracy, enabling them to achieve both accuracy and scale. Aaron cited a project for a CPG client. The client had been conducting a campaign among Hispanics. Truthset discovered that 29% of their target was actually Hispanic. Truthset helped them pick segments and evaluate how Hispanic they actually were. Kathryn discussed Truthset’s methodology for creating the Truth Scores, and Joel reported on an independent review he had done of that methodology.
Key takeaways:
- Truthset’s methodology is based on three hypotheses:
- Average accuracy levels for large segments vary across data providers.
- Consumer IDs in a large segment are not all equally accurate.
- It is possible to estimate the underlying probability of accuracy with a Bayesian wisdom-of-crowds method.
- To carry out its analysis, Truthset compares the demographic attributes in the Big Datasets to self-declared data in a number of small, independent validation sets.
- Kathryn provided an example with two Hashed Email addresses (HEMs) in two different providers’ databases. If their attributes agree, then Truthset gives them a high Truth Score. If they do not agree, Truthset assigns a slightly higher Truth Score to the database whose accuracy rate has been higher in comparisons to the validation data sets.
- Joel interviewed the Truthset technical team and conducted a series of statistical tests. He compared the actual incidence of gender, age, and ethnicity characteristics in a set of IDs with lower truth scores to a set of IDs with higher Truth Scores. The accuracy levels of those attributes in the low-scoring set tended to be in the single digits, while those in the set with higher scores were 62% to 84%.
MediaCell — a passive approach to future-ready cross media measurement
Jim Ford — Global Development Director, Ipsos
Summary:
MediaCell is an app that transforms a smartphone, tablet, or laptop into a passive portable audiometer. It also captures on-device consumption of audio and video. It is currently deployed in panels in the U.K. and Australia and will soon by deployed in South Africa. Ipsos is working with Kantar to use MediaCell in the Netherlands to produce the radio currency there. Its audio measurement capabilities are based on ACR technology. It is easy to on-board panelists; Jim noted that there were no breaks in the data in their U.K. panel due to the pandemic.
The U.K. panel of 3,000 was built for the BBC. It captures live and time-shifted listening and viewing for over 250 channels and over 1,000 properties. The overall TV viewing levels by day within this panel were similar to those recorded by the “gold standard” TV currency study in the U.K., which is recruited face-to-face and measures TV tuning with household set meters. Jim said that MediaCell does not replace set meters but enables a researcher to look at audiences for programs on all devices in all locations.
Key takeaways:
- MediaCell is a low-impact passive measurement technology that leads to greater panel participation and compliance.
- It can track engagement with a TV program on social media during broadcasts of the program.
- In the U.K., there was increase in streaming video and audio audiences in 2020 during the first lockdown. among those 55+, according to their panel.
- Video-on-Demand (VOD) viewing surged during the 2020 lockdown. In 2021, VOD viewing dropped below pre-pandemic levels in the U.K., especially among younger consumers.
- There are great differences in VOD viewing by age. Among those 16-24, there is far more time spent viewing VOD than viewing of linear TV; among those 25 to 34, the average amounts of time devoted to VOD and linear TV are roughly equal; among those 55+, linear TV viewing not only dominates over VOD viewing, but it is at far greater levels than TV viewing among those in any younger age group.
Data Deprecation & Rising Privacy Concerns
Cookieless Audience Targeting and Attribution: A Pharma Case Study
Joy Joseph — Chief Data & Analytics Officer (US), Initiative Media
Summary:
While attribution has been around two decades, a great breakthrough for digital, the deprecation of third party cookies is likely to have a significant impact. Options to MTA include: walled gardens, focused on one channel only, and data clean rooms such as those operated by the identity companies. Cleanrooms are data intensive but much better than single channel walled gardens.
However, pharma is a unique category because of its privacy restrictions. A third model is an extension of the clean room approach. Google’s Ads Data Hub lets you integrate your data with segments of 50 persons so there is no identifiable information. The data are analyzed as groups of people and so it is similar to MMM but with more granular data, as MMM operates on data such as GRPs and ratings.
Joy showed an example of their analysis of some 45 campaigns: where their lifts were, they found significant differences in the lifts by publisher. Optimizing on these findings, they were able to drive a 1.7x increase in ROAS. Real time optimizations were hence possible despite the deprecation of third party cookies. Next steps for this approach will be predictive audience activation to score audiences for remarketing and propensity based targeting.
Key takeaways:
- Clean rooms are an option to walled gardens in the face of third party cookie deprecation.
- Initiative has had much success using Google’s Ads Data Hub to conduct a much more granular version of MMM.
- Importing much of your data and integrating with segments within the hub yielded real opportunities to optimize, even real time, resulting in a 1.7 lift in ROAS.
- Initiative will be further developing this model to permit predictive audience activation.
Building Trust Through Transparency
Reg Baker — North American Regional Ambassador, ESOMAR
Martin Oxley — Founder & Managing Director, buzzback
Summary:
In a 2019 Pew survey, 79% of Americans say that they know big companies track their online behavior while 59% say they don’t know how the data is used and have little control over how the data is used. In a 2020 survey, Pew found that most Americans want the right to permanently delete their health related data.
Buzzback surveyed several respondents across several global markets. In this short period of time, the incidences of bad online experiences have increased. These include spam, scans and phishing. Online experiences are growing more negative.
How does this impact data sharing? 60% of global respondents are highly concerned about data sharing (called their Privacy Calculus). The U.S. decline was even more precipitous. Confidence in laws and regulation’s ability to protect respondents has also declined dramatically. Willingness to share is declining globally even if it is capable of saving lives.
Key takeaways:
- Americans know their online behavior is being tracked but don’t know how their data is used.
- Globally, people’s online experiences are becoming more negative especially since COVID.
- The declines in online confidence were greater in the U.S. than in the other 10 measured countries.
- Confidence in law and regulation to protect the public has gone down over the past few years.
- Data collectors must earn trust through transparency.
The Context Effect: How Context is Critical & Transforming Digital Advertising
Tony Marlow — CMO, Integral Ad Science
Summary:
As consumer concerns about privacy are growing and legislators are responding by introducing legal restrictions, marketers are focusing on contextual targeting (instead of cookies) to make their advertising more effective. Research by Integral Ad Science (IAS) has shown that the growth of contextual targeting is not only being embraced by marketers, but also welcomed by consumers: they prefer ads that match the content they are viewing – which makes them more relevant to them.
While many studies show that context can increase ad impact and that matching ads to the context is a great way to take advantage of context effect, IAS commissioned a neuroscience study to gain new insights. The research explored different ways of matching ads to context to help advertisers optimize their strategies. The study compared how memory and emotional involvement with the ad were affected depending on how ad and context were matched.
Key takeaways:
- The resurgence of contextual targeting not only benefits advertisers but also consumers since they find contextually relevant ads more appealing.
- The new research presented here confirms that context-ad matching is effective. Further, it suggests that informational ads are more likely to perform well when matched with the message of the context, whereas emotive ads are likely to excel when they are matched thematically with the context and its mood.
Attention Measures: What Counts & How Much Does it Cost?
Unlocking the New Currency of Attention
Joanne Leong — VP, Director, Global Media Partnerships, Dentsu
Heather O’Shea — Marketing Science Lead, Snap, Inc.
Summary:
This presentation focused on the implications of attention for advertisers, and how we can start applying these insights from a planning and buying perspective. Although viewability has become a proxy for attention, just because an ad is considered viewable does not mean eyes on ads. Viewability isn’t necessarily a true indication of attention – it can be quite arbitrary, and we need to go beyond that to the better and more accurate representation of those natural attention behaviors.
Key takeaways:
- What is the relationship between attention advertising and desired outcomes (recall and choice)? Across both, there is a pretty strong relationship beyond eyes on dwell time and recall of choice uplift.
- Overall, creative is the main driver of attention’s impact on outcomes. Three to five seconds on one platform can be different from three to five seconds on another platform.
- Shorter video ads can be more “attentionally efficient” in delivering outcomes, compared to some of the longer ads. For instance, if you see five seconds of a 10-second ad you probably saw most of what you were meant to see in terms of what the brand wanted to land in messaging and branding. In contrast, if you saw five seconds of a 30-second ad, you probably saw very little of what you were meant to see.
- Combining all of these things together, along the cost of each of the formats, we can start to develop a CPM for an effective attentive second.
- Snap research shows that the key ingredients for effective attention on the platform are: SNAP 100% screen coverage, sound on, non-skippable.
Attentive Reach: The Case for Human Attention in Brand Advertising
Max Kalehoff — VP, Marketing, Growth, Realeyes
Summary:
Realeyes uses software solutions for measuring human attentional and emotional response to advertising. Max Kalehoff explained the importance of attention measurement as a gauge of brand performance. He urged companies, particularly those with latent sales, to incorporate it into their business models.
Max’s case for prioritizing attention measures starts with the premise that lowest-cost reach is a broken strategy. The COVID-19 pandemic-driven surge in direct-to-consumer (DTC) marketing in CPG is influencing how we think about brand advertising, but the focus has been too oriented on performance. Forty percent of sales growth in CPG in 2021 is in DTC, but a huge share of CPG sales (nearly 94%) is not in e-commerce. So, there’s still a big share of sales that needs to be addressed that brand-lift studies haven’t been able to explain. For many advertising teams, what matters is reaching their lowest-cost target and demographic. This leads to a lowest-cost mentality, which prioritizes cost over quality, which sacrifices consumer experiences and welcomes fraud.
Next, Max argued that attention is a measurable business outcome. Attentive reach—the audiences you’re trying to reach that are actually paying attention—can unlock performance for brands. But, attention is a prerequisite of emotion, and emotion drives memory and behavior; it unlocks other measurable outcomes: awareness, considerations, sales and lifetime value. The three major drivers of attentive reach are creative (how well it stands out, captures and maintains attention); audience (who is your audience, is it receptive to the ad and how well does the ad/brand fit their interests and psychographics); and context (media context and channels and devices).
Key takeaways:
- Lowest-cost reach is a broken strategy.
- Attention is a measurable biz outcome.
- Creative quality is key in attention measurement.
- Brands, particularly those with latent sales, will benefit from prioritizing attention measures but this must be led at the CEO level, with creative, media and data teams working collaboratively.
- Mobile apps can have up to 80% more attention than an ad on linear TV.
- With more ad exposures per day, the more likely you are to reach users when they are attentive.
Prior Attentive Ad Exposures Increase Ad Attention
Tristan Webster — SVP, Products & Operation, TVision
Kenneth C. Wilbur — Professor, UCSD
Summary:
Tristan Webster and Kenneth Wilbur showcased their most recent collaborative work examining attention and frequency in advertising: the impact of multiple exposures on people’s attention to TV ads. They applied CTV data which TVision has collected natively in the field to provide insight into the long-examined question, “Is there an optimal frequency for TV ads?”, but more granularly: “What is happening in the media environment while viewers see ads, and how does that affect their attention?”
With a panel of 5,000 homes across U.S. who have agreed to install TVision’s proprietary technology, the researchers can measure what and when people are watching, who’s watching and whether or not they’re paying attention. The data is collected passively in the living room from real viewing sessions, at scale.
The research also studies wearout, specifically measuring the point at which the consumer stops showing interest in (or paying attention to) the message. It’s clear from the data that the greater the repetition within this range of exposures analyzed (between 6 and 10), the more attention occurred, and the more familiar and appealing the ads were, the greater the viewer interest in them.
The researchers also study recency’s effect on attention. Compared to linear TV, where you might not see the same ad for an hour or even days, in CTV if an ad is delivered within a couple minutes within same ad being delivered to a household, there is a massive decrease in attention being paid. It does level off, so it seems people are okay with an ad being delivered multiple times as long as there’s space in between. If it’s the same ad that the viewer sees sequentially, this has a material negative effect on attention and has important implications for the way ad content is trafficked in CTV space. Marketers, programmers and schedulers need to be aware that understanding what drives whether people to engage with content is a complex dynamic. This is key to creating a more meaningful and higher-attention overall experience, which is what we’re all looking for whether it’s the audience, marketer, or programmer.
Key takeaways:
- Attention to TV ads rises uniformly across the first 10 attentive ad exposures.
- Frequency’s positive impact on attention is consistent across age demographics.
- Co-viewers are less likely to increase attention over multiple exposures (in situations with +3 people in the room, the incremental impact of frequency lessens after the third exposure).
- Viewers pay less attention when CTV ads are quickly repeated.
- Viewers are more tolerant of short ads. Recency doesn’t seem to impact attention to 15-second ads, but viewers of 30- and 60-second ads are less likely to pay attention when the same ad is quickly repeated.
MODERATED TRACK DISCUSSIONS
Television Disrupted
Session Chair: Andrea Zapata — Head of Research, Data & Insights, WarnerMedia
Cross-Platform: Measurement & Identity
Session Chair: Jorge Ruiz — Head of Measurement, TikTok
Summary:
Moderator Jorge (TikTok) asked the panelists three key questions:
- The pros and cons of using mobile phones as meters in a time of such strong privacy concerns;
- The panelists’ views on measurement of advertising effectiveness; and
- The most unexpected feedback they received.
Key takeaways:
- Erin (Nielsen) said that Nielsen’s wearable helps to “future-proof” their technology against changes in “the privacy space” and changes in mobile behavior.
- Jim (Ipsos) said that MediaCell fits naturally into people’s lives, requiring them to do little more than they already do. All panelists must go through a “triple opt-in” procedure to participate. Ipsos does need to maintain strong relationships with Apple and Google to ensure that their meter continues to be available in their app stores.
- Jim said that measurement is a “team sport;” panels, such as MediaCell panels, provide a source of truth “in a census world.” In the Netherlands, the individual media are breaking down silos to work together in a collaborative way.
- Aaron (Truthset) sees Truthset as providing panel-like accuracy to census-level data assets. By moving away from demographic attributes as discrete binary properties and considering them as Bayesian probabilities, Truthset can give clients both flexibility and control. They can set the accuracy threshold – the minimum likelihood that their consumers fall into the pre-defined target – anywhere they like.
- For Jim, what was most unexpected was clients’ willingness to embrace innovation. The most unexpected feedback for Erin was the positive feedback from research participants about the wearables, which will help make it easier for Nielsen to recruit panelists and will improve compliance in their panels. For Aaron, the most unexpected feedback he received was that clients who had invested heavily in multicultural audiences and were shocked to find that the vast majority of their advertising spend was reaching out-of-target audiences. He was also surprised, he said, in the wide variation in accuracy between the datasets Truthset has examined.
Data Deprecation & Rising Privacy Concerns
Session Chair: Andy Fisher — Head of Merkury Advanced TV, Merkle
Summary:
Three core factors driving/limiting targeting are privacy, devices (IDFA), and at browser level (third party cookies). There are two broad trends in marketers’ response: 1) recreate identities or improve contextual targeting. Initiative suggests their attempt to recreate identity is balanced by their respect for privacy. Initiative is trying to maintain journey based engagement through technology.
Contextual however, gets the right message to the right person at the right time. Many have said this is important to rebuild brands.
Predictions for the future: privacy as a service; Apple and AI become a privacy intermediate, keeps track of terms of service; and finally, marketing by mindset (context gives the right mindset).
Key takeaways:
- Companies are using privacy as a selling proposition.
- New technology such as telehealth and self-driving cars will make consumers even more concerned.
- Contextual advertising captures people when they are in the right frame of mind for the ad.
Attention Measures: What Counts & How Much Does it Cost?
Session Chair: Earl L. Taylor — Chief Knowledge Officer, MSI
Summary:
An impressive body of work is building in attention measurement. The three winning-papers sessions preceding this panel revealed a work in progress with shared goals as well as differences in approaches. Moderator Earl Taylor of the ARF’s MSI division asked the speakers about their views on barriers to the process, and opportunities for further improving attention measures.
Max Kalehoff of Realeyes criticized the tendency among many brands to follow low-cost reach strategies toward sales outcomes, at the expense of creative quality. His model provided evidence that a focus on creative quality is imperative for developing attention as a measurable business outcome.
Joanne Leong of Dentsu and Heather O’Shea of Snap, Inc. saw a strong relationship between creative and media experience — meaning that creative could either positively or negatively impact viewer’s reactions to an ad. “Creative alone (as a measurement focus) is not guaranteed if you want to maximize your performance,” Heather said, “so it’s important to look at a combination of media and creative” — which media channels you are using and whether the combination aligns with your campaign goals.
Creative measures also are affected by the environmental factors which drive attentional behaviors, TVision’s Tristan Webster and U.C. San Diego’s Ken Wilbur, added. The way people engage with creative can change over time. Tristan pointed to factors such as “time of day, whether or not other viewers are in the room, the context in which an ad runs, and the start of a pod or whether the ad is tucked away in the middle. These factors are very predictive over time and are helpful in determining whether a given ad will drive attention.”
Key takeaways:
- Attention is a desirable measure by itself, and the industry needs to establish it as a common currency.
- Attention may be the way to understand how the consumer processes the ad.
- Creative has a key impact on attention, but researchers differ in the role it plays.
- Media context and channel selection can affect attention differentially.
- Brands need to prioritize resources for attention measures, and this will require reinforcement at the CEO level.
PRESENTATION & PANEL
An Economic Analysis of TV Advertising Profitability
Presenter: Anna Tuchman, Ph.D. — Associate Professor of Marketing, Kellogg School of Management, Northwestern University
Moderator: Carl F. Mela, Ph.D. — T. Austin Finch Foundation Professor of Marketing, Duke University
Panelists: Anna Tuchman, Ph.D. — Associate Professor of Marketing, Kellogg School of Management, Northwestern University
Radha Subramanyam, Ph.D. — Chief Research and Analytics Officer, CBS Corporation; President, CBS Vision
Marc Vermut — VP, The Knowledge Lab, Neustar
Brian Wieser, CFA — Global CEO, Business Intelligence, GroupM
Summary:
Professor Anna Tuchman of Northwestern’s Kellogg School of Management presented an economic analysis of the effectiveness of a major portion of the U.S. linear television advertising spend, worth $66B in 2019. The study, conducted with Bradley Shapiro and Günter Hitsch of the University of Chicago’s Booth School of Business, investigated the elasticity of actual sales for 288 distinct CPG brands on the basis of brand-specific ad spend as captured by Nielsen store and household panels along with Nielsen Ad Intel data. They used log-log regression and flexible machine learning algorithms. Their goal is to help firms evaluate ad campaign effectiveness with the hope their measurement will help make better ad spend decisions.
The key finding of their research is that linear TV advertising spend bears little correlation with actual sales. The effect of linear TV advertising on sales is smaller (5 to 20 times smaller) than previous estimates in the literature, producing an ad elasticity coefficient of 0.014. For a majority of CPG brands studied, linear TV advertising yields negative ROIs. The paper’s conclusion: Ad budgets (at least in CPG) are heavily over-weighted to linear TV advertising. As a potential “silver lining,” at lower absolute levels of linear TV ad spend, buying additional ad placements might be profitable, driving larger marginal benefits over marginal costs.
The presentation was followed by a lively panel discussion with Radha Subramanyam, CBS Corporation, President, CBS Vision; Marc Vermut, VP, The Knowledge Lab, Neustar; and Brian Wieser, CFA, Global President, Business Intelligence, GroupM. The discussion was moderated by Professor Carl F. Mela, T., Austin Finch Foundation Professor of Marketing, Duke University.
The panelists provided a variety of critiques of the research:
- The focus on short-term sales may be myopic; outcomes should be broadened to incorporate long-term value creation for brands. In addition, the focus on store sales to the exclusion of e-commerce is problematic for generalizability of findings. It was suggested that profit and brand lift might be included in future, as well as search volume, web visits, and survey data on brand awareness, equity, and health.
- This research did not include assessment of creative in advertising, which a panelist stated represents two-thirds of ad effectiveness. One panelist suggested to “always remember the humanity of the consumer” when thinking about ad effectiveness.
- The focus on CPG brands only is limiting (the panel called to include FinServ, telco, pharma, auto, etc.).
- The research was limited to linear TV advertising and didn’t incorporate other channels of advertising, particularly addressable media, which have been growing in importance. As a point of contrast, it was suggested that the current focus in the advertising industry is the optimization of ads in premium video and online search, as well as the interactions between online and offline media and behavior, which are mutually reinforcing.
Professor Tuchman pointed out that research must start somewhere as it is impossible to address all of these limitations in a single study; indeed, she indicated that other categories will be explored next. Regarding outcome measures, she responded that although these intermediate outcomes may be important for marketers, shareholders only care about those to the extent that they ultimately result in sales outcomes. Regarding the role of creative, she indicated that the average response to observed creatives was used in the data analysis, and she welcomed exploring this further in the future.
Repeated calls were made to “not lose the forest for the trees” in this kind of work. There was agreement on the need for more collaboration between academics and industry practitioners. MSI provides the right auspices for just this kind of work to develop most profitably, since MSI provides a unique platform for industry and academics to collaboratively address such “weighty issues” in marketing.
THE LAST WORD
Ameneh Atai — GM, Digital and Advanced TV Commercial, Nielsen
Stephen DiMarco — Chief Strategy Officer, Tubular Labs
Jed Meyer — NA, Media Domain Leader, Kantar
Mark Truss — Chief Research Officer, Wunderman Thompson Data
Summary:
Stephen DiMarco (Tubular Labs) moderated this discussion on the second day of the conference with the anchor commentators who shared their perspectives on what they heard and what it means. Below are edited highlights from the conversation.
- A common thread of Tuesday’s interviews and presentations was the complexity of attention, no matter whether you are measuring, people, media, or content. It is hard to get accurate agreement on what attention is, according to Stephen. There is the additional challenge of persuading consumers when you cannot get their attention, noted Jed (Kantar). It is crucial to cut through clutter to speak to customers.
- Marc Pritchard’s call-to-action for constructive disruption by marketers to help pave the path to cross-media measurement was inspiring to Stephen and Ameneh (Nielsen). Marc’s keynote fireside chat with Scott was complimented by conference sessions on why this is challenging, as well as sessions suggesting that it may be possible to achieve these goals.
- There are tectonic shifts and disruptions impacting consumer experiences, privacy, next gen media, etc. during a global pandemic, noted Ameneh. Transparency, objectivity, and thoroughness in measurement will move us in the right direction. Additionally, consumers are annoyed with brands due to overexposure to ads. Just because frequency has been mastered, there is a need to layer on other metrics. Efficiency is needed. This will help brands achieve their objectives. However, there is too much focus on self-serving agendas. Ameneh stated that we need advertisers to refocus us on the customer experience.
- There is so little agreement on how our industry works. People are digging in and taking hard positions within the industry like never before, observed Mark (Wunderman Thompson Data). He added that he was struck by how our industry mirrors our country. Can’t we agree on anything? There are absolute truths and untruths. No agreement on whether TV ads work, on what drives attention, on the role of creativity, etc.
- Jed agreed with Mark that people are divided and only see their own truths. But he believes there are multiple paths forward. This divisiveness speaks for the need for the ARF as an independent body. Rigor and science are needed. However, stepping back is hard to do remotely. Jed also suggested that post-COVID, there is an ongoing need for agility by brands, publishers, marketers, and vendors. Test and learn is important.
- In terms of the learnings the commentators want to bring back from this conference, Mark spoke about agility, Ameneh felt that always keeping the customer in mind is essential, and Jed stated that this is a great time to be alive in media and consumer research.
- Let’s make the industry better, make it more objective, don’t just use your own facts. How do we move it forward? Everyone agreed with these remarks by Jed.
CLOSING REMARKS
Scott McDonald, Ph.D. — CEO and President, ARF
Wednesday, September 22
WELCOME
Scott McDonald, Ph.D. — CEO and President, ARF
KEYNOTE
Collaborating on Research and Solutions for Private Measurement
Dennis Buchheim — VP, Advertising Ecosystem, Facebook
Summary:
Dennis Buchheim has a unique perspective from “two sides” of the ad industry. He leads Facebook’s Ads Ecosystem team after serving as CEO of IAB Tech Lab. Dennis shared a view on how the industry is grappling with the shifting regulatory, platform, and technology landscape. Only together can the industry understand these changes and create paths forward. As opportunities are evaluated to evolve how data is used, research will be critical in refining the industry’s foundational knowledge and providing tactical guidance.
Facebook and IAB Tech Lab missions are similar in that both stress collaboration, engagement in the digital media community, and development of foundational solutions and standards. Common challenges include regulations, privacy, safety, integrity, misinformation, and disinformation. Solutions must be holistic and will impact targeting and optimization, effectiveness measurement, marketing best practices, and budget allocation.
Dennis also showcased a recent Facebook-academic collaboration that compared over 1,000 randomized ad experiments to non-experimental proxy metrics and machine learning models, to provide insights into when and how to apply non-experimental measurement to evaluate the success of an ad campaign. This research reveals that lack of data does not mean lack of measurement since proxy methods can be used. Simpler last-click metrics can work well for advertisers in the absence of experiments. Click-based proxies can guide similar go/no-go decisions based on cost per incremental conversion. With just one experiment, an advertiser can calibrate last-click and last-view metrics to achieve better estimates.
Key takeaways:
- The challenges represent shared issues that are critical, but historically, each player has operated independently. Different parties have unique pieces of the puzzle, including valuable data, insights, and tech, as well as understanding and perspectives of the issues.
- Collaboration is essential. No one company can tackle the challenges of the digital media ecosystem alone. Facebook hasn’t been a terrific collaborator with the industry to date, Dennis noted. However, it’s an opportunity for Facebook to be an industry ally going forward.
- Research/data will be a key factor in assessing paths and solutions going forward.
- Accurate measurement is vital, it drives advertising dollars. However, no single solution exists. A portfolio of solutions is needed to the existing targeting, measurement, and optimization challenges.
- Dennis provided an open invitation to the industry to collaborate to help solve these challenges together.
DOWNLOAD FACEBOOK’S PRESENTATION
INTERVIEW
The Evolution of Measurement and Looking Towards a Privacy-Centric Ecosystem
Gaz Alushi — Global Head of Measurement Research and Development, Snap, Inc.
Interviewer: Paul Donato — Chief Research Officer, ARF
Summary:
The ARF’s Paul Donato interviewed Snap’s Gaz Alushi about the impact of privacy-centric ecosystems on measurement and the consumer. From his perspective of working within a walled garden, Gaz speaks to how Snap puts the consumer first without conflating tracking with measurement in capturing the impact of an ad campaign. Paul and Gaz also discussed the industry’s response to privacy regulations, working with first- and third-party data, the MRC’s cross-platform standards, and how Snap works with partners across the spectrum to measure its audiences.
Key takeaways:
- The MRC standards work for Snap as it makes it more fair across the spectrum. Gaz believes there will be a natural evolution where the industry will start translating time spent to actual attention, and how much time is required for attention based on audience, demographic, or product that’s being advertised.
- Biggest question Snap has been getting is making sure there’s efficiency across the spectrum, but the biggest demand is for cross-platform measurement, to measure at parity with other platforms.
- Most likely the industry will see aggregation across multiple sources to ensure that there is a fair broker of objective measurement and data consolidation and aggregation. Gaz predicted that the concept of tracking an individual user is going to fall away and the industry will re-pivot to looking at their audience as a whole in making their decisions.
FIRESIDE CHAT
A Holistic Measurement Journey
Ashley Eckerlin — SVP of Commercial Strategy, Planning & Analytics, SharkNinja
Mike Menkes — SVP, Analytic Partners
Summary:
Change has been the only constant of the past 18 months, but for brands that know how to adapt, this change represents opportunity. Holistic measurement – from understanding incremental impact across channels to building in considerations for macro factors like consumer behavior shifts and the pandemic – is key to sustained growth.
SharkNinja and Analytic Partners shared how they honed their advertising investment strategy as media mix has become more diversified and how they are measuring impact now. Mike Menkes and Ashley Eckerlin presented the journey and evolution of analytics at SharkNinja, a company focused on kitchen appliances, cleaning, and home care products.
Ashley stated that the firm has come a long way from their pre-analytic measurement. Previously the company measured their infomercial DTC business via direct response sales (phone and web). Using retail POS trends to evaluate TV sales was a simplistic measurement.
SharkNinja needs commercial marketing mix modeling to understand the increasing complexities of the total business. As a result, the company’s measurement analytics have evolved significantly, and its media mix has expanded tremendously.
Marketing mix modeling lets the company know where to put their next dollar. The data can be parsed to uncover what’s working, not working, account for halo effects, as well as understanding how product and portfolio messaging work together.
Ashley spoke about Amazon’s complicated ecosystem. Using direct response analytics was insufficient. Commercial marketing mix modeling enabled SharkNinja to see Amazon’s impact on the business.
Key takeaways:
- According to Ashley, commercial marketing mix model represents a beacon, allows for the use of external data sources, and provides ideas for new investments. This model enables SharkNinja to answer its main question: Is marketing driving ROI at retail? Incremental analysis informs the marketing mix and health of the brand.
- Getting leadership on board with the modelling represented a full 180-degree change from a direct response to a complex model. The approach, methodology, and proof points had to be reiterated. It is not a black box. Drive change management and commit to whatever achieves the highest ROI.
- Knowing the role of COVID vs. the impact of marketing on sales of SharkNinja was critical. Analytic Partners examined the role of non-marketing factors driving business performance and measured their impact. Using the model helped parse data and allowed for accurate forecasts.
- Marketers should seek growth and new ways of analyzing and measuring business. Test and learn is critical to moving forward.
SPEED PAPERS
Understanding Consumer Sentiment Towards Data & Privacy in Advertising
Bridget Bidlack — SVP of Product, Lucid
Summary:
Lucid is a research technology platform that delivers programmatic access to first-party survey data that can be used to drive business initiatives. Using survey-based methodologies to help clients understand the accuracy and effectiveness of their advertising efforts.
In her presentation, Bridget Bidlack shared recent research Lucid conducted on consumer sentiment on data privacy as it relates to the concept of a “free Internet” and sharing of data. Although the Internet is “free” to users, there is an inherent quid pro quo requirement: users must view ads and share data to use the Internet, and this “contract” isn’t always transparent.
Accordingly, Lucid wanted to explore consumer sentiment to determine how different user groups feel about sharing their data (i.e., do they know how their data is shared?). Lucid marketplace surveyed users across 10 countries and reported the findings below.
Key takeaways:
- The majority of users want more transparency about why their data is being collected but, on average, feel they have a good sense of how their data is used to provide personalized experiences.
- Users want to be incentivized for sharing their data, a practice already in place in the Lucid marketplace.
- The reputation of a company is important when deciding whether or not to provide information.
- Women reported being more concerned with privacy issues, however, men are more likely to report using software to block collection of their online data.
- American respondents cared least about data collection compared to the other nine countries surveyed.
Infrastructure Evolution for Next Generation Market Research
Raj Mann – Sr. Data Scientist, Microsoft
Summary:
Microsoft’s goal was to create a solution to bring independent data sources together. Their database contains 2+ years of data on millions of Microsoft’s identified customers, usage across 10+ core consumer products, and revenue (both direct and indirect, across these products). This database is also GDPR compliant. Protecting this data is a priority.
Raj Mann reviewed the evolution of Microsoft’s infrastructure. In the past, research was slower due the need for a lot of data manipulation and quality control challenges. The investments have increased extensibility of data and removed obstacles that prevented data access for researchers and improves collaboration. Microsoft’s infrastructure investments resulted in a 70% reduction in time for data analysis on the 600 studies Microsoft conducts annually.
In addition, the investment:
- Makes it possible to see connections between customer sentiment and revenue generation.
- Continues to deliver on some of the critical consumer-focused market leading research without additional resources.
- Maintains security, privacy, and customer trust around the data.
- Allows uncovering actionable insights from machine learning.
CONCURRENT SESSIONS – WINNING PAPERS & SOLUTIONS
Post-Pandemic Trends: Agility & Adaptability
The Future Goes Viral: Brand Challenges and Opportunities Triggered by the Pandemic
David Tice — Senior Consultant, HUB Entertainment Research
Jon Giegengack — Founder & Principal, HUB Entertainment Research
Summary:
The presenters summarized insights from HUB’s on-going series of surveys on viewers’ ownership, behaviors, and opinions about all aspects of TV and video. Since they have data going back to 2018, they are able to analyze long-term trends, including the impact of the pandemic.
The presentation focused on how the pandemic affected trends and what the data indicate about the future. For example, the adoption smart TVs and of streaming services has led to consumers “stacking” subscription services. Growth exploded during the pandemic. However, the data show that growth is slowing and there are no indications that a majority of viewers is replacing “traditional” TV with streaming. In other words, the pandemic accelerated this trend, and that trend is likely to persist, but not at the accelerated rate.
Key takeaways:
- Surveys on trends in viewing and viewing sources used by consumers show that the pandemic has accelerated trends that were evident prior to the pandemic: the growth of viewing sources used in general and especially of streaming services.
- The data indicate that exclusive content (such as first-run movies on HBO Max) is a strong driver of viewers’ adoption of and continuing subscription to streaming services.
- The researchers conclude that their findings suggest some likely permanent changes: the continuation of relatively high viewing levels, less linear TV and more streaming. The data also indicate that, while some are eager to return to movie theaters, watching (and paying for) first-run movies on TV will also remain high.
How High-Frequency Data Offers Consumer Insights
Joanna Piacenza — Head of Industry Intelligence, Morning Consult
Summary:
When will things return to normal and what will our new normal look like? To answer these questions, Morning Consult is tracking how consumer attitudes are shifting across a wide range of categories to gain greater insight into not only when consumers are ready to return to their normal activities but how their habits have changed. This presentation provides a case study on travel and hospitality with insights drawn from their Tracking the Return to Normal project.
Key takeaways:
- There was a rise in consumer comfort in vacationing in 2021, but progress has flattened with the rising Delta variant – over half of U.S. consumers are very concerned.
- Currently, 70% of U.S. consumers feel comfortable taking a road trip, but comfort in shared mobility, although doubled since the beginning of the year, never crossed the 50% mark.
- Although leisure travel has rebounded more quickly than business travel, there are higher levels of comfort in traveling among business travelers. Share of comfort in international travel is around 40%, which is closer to early January levels.
- Morning Consult predicts that most consumers would feel more comfortable traveling by Q2 2022, but travel and hospitality brands should focus on marketing more locally.
The Performance (Media) Must Go On
Tess Erickson — Director of Research & Strategy, Broadbeam Media
Summary:
Broadbeam Media found past assumptions of consumer behavior being upended by the pandemic. To better understand the transformation happening in the American home, they conducted proprietary research to focus on the respondent’s perceptions of the last year, through sight, taste, hearing, smell, touch, and feel, to influence respondents to think about their last year in a fresh way. The following includes what they learned through the research.
Key takeaways:
- Segment behaviors, not people. Mutually exclusive consumer segments were not providing the full picture as anyone that fell between the extremes were getting lost. Broadbeam Media found it more useful to cluster the data by behaviors and how they were interacting.
- Platforms are differentiated by content, not devices. Linear was still dominant for time-sensitive content and legacy reality shows. One type of unscripted content more popular on streaming were documentaries and investigative reporting. Streaming is taking over high-budgeted scripted space (e.g., recent movie releases, high budget action and fantasy, family content).
- Remote work changed not just people’s schedules but also their media schedules. For instance, overall media consumption was up, but remote workers watched significantly more linear TV, streaming, and social video later at night. Also, even without commutes, 65% of remote workers were listening to audio.
Cross-Platform: Measurement & Identity
Extending the Audience Scope: Brand in a Multisource Data Context
Stephanie Grassa — Head of Research and Operations TGI France, Kantar
Gilles Santini — Consultant, Vintco
Summary:
This presentation describes a project to enhance analytics on print and digital audiences in France. The main sources of data for print and digital audiences in France are:
- ACPM/One Next: References audience measures for magazines and newspapers.
- Mediametrie: Reference for broadcast and web media audiences.
- Kantar Media Division/TGI: Established to analyze consumer behaviors, brand preference, and market trends for advanced segmentation, planning, and activation.
The goal of the project was the integration of the media brand inside the French TGI system in order to enable analysis of media brand consumption combined with product and brand usage.
A probabilistic approach was decided upon, and the methodology involved matching, ascription, calibration, and brands. This methodology resulted in candidate probabilities to which a final calibration algorithm was applied.
Once this process was completed, the brands were integrated for use in TGI.
A limitation of this methodology exists because the probability of the brand is constructed from the print and web probabilities. This methodology accepts, at an individual level, the independence of behavior between print and web. According to the presenters, this is clearly not true, but to do any better is very challenging.
Have We Been Mis-Measuring Basic Demographic and Media Consumption?
Jon Puleston — VP, Innovation, Profiles Division, Kantar
Summary:
Inaccurate responses to basic demographic and media consumption questions asked in surveys are very problematic, since these responses form the bedrock data points for all demographic research and media consumption measurement.
Jon Puleston analyzed some of the reasons why respondents are not truthful in surveys and the challenges raised. As a result of inaccurate respondent responses, the industry has been mis-measuring basic demographics and media consumption for decades.
Additionally, each company and country has its own methods of measuring basic demographics and media consumption. Growing levels of international cross-market research require accurate and consistent data.
Fifty cross-market experiments and desk research provided a framework to address some of these fundamental flaws and formed the basis for Jon’s proposals:
- Adopt more inclusive and precise approaches to how demographic and media questions are asked to improve the accuracy of responses. Establish an ESOMAR committee to develop best practice methods made up of representations from the key panel supply companies, leading market research firms, and key buyers of international market research.
- Upgrade how basic demographics and media consumption questions in surveys are measured. Pledges are needed from key players in the research industry to work to adopt and implement the use of these best practices.
- Create an international measurement standard. Support is needed from advertising professionals and ARF members.
The Digital Divide: Inclusion of a Mobile Population (To Have & Have not: Part 2)
David Kingsbury — Analytics Consultant III, IRI
Summary:
David Kingsbury reviewed data from his 2019 ARF AUDIENCExSCIENCE presentation, and IRI’s subsequent advances to improve accessibility, targeting, and measurement of the digital invisible. This research has helped uncover a sliver of those invisible, i.e., households that move and are lost (either temporarily or permanently).
Research during the summer of 2021 revealed that the demographics of the digital invisible remained the same. Digital invisible consumers are more likely to skew low income, no kids, and under 34 years of age. They are part of the 13% of the population who move annually. Additionally, some consumers don’t want to be tracked. This population is not receiving targeted advertising and their purchases are not being measured.
Key takeaways:
- Approximately 20% of all shoppers exist outside normal PII coverage, and 17% of frequent shoppers (shoppers who regularly use loyalty cards) are invisible.
- IRI found that there were opportunities to recover or claw back 9% of these frequent shoppers via deterministic matching using virtual consumer identifiers. Mapping these households gives advertisers an opportunity to include them in digital messages and measurement in a privacy safe environment.
- Ad effectiveness measurement indicated that the digitally invisible had the highest sales lift. Inclusion of the digital invisible consumer in ad targeting, activation, and measurement improved the brand’s sales outcomes and ROAS.
Data Deprecation & Rising Privacy Concerns
IAB State of Data Initiative 2020
Chris Bruderle — Sr. Director Research & Analytics, Interactive Advertising Bureau (IAB)
Summary:
The IAB Programmatic & Data Center examined the impact of the loss of third-party cookies and identifiers on the digital advertising ecosystem, particularly on brands and publishers. This report measures the industry’s sense of preparedness, expectations for revenue and data-driven tasks, and progress into leveraging first-party data. Two studies were executed, conducted by Ipsos and McKinsey, surveying and interviewing key data leaders across categories.
Key takeaways:
- Overall, 67% of the industry reported being prepared but there’s disparity by company type with most prepared being ad tech and data companies (86%) and least prepared being brands (48%).
- However, their concerns about executing essential tasks in the post cookie ecosystem (“table stakes” tasks such as targeting, measurement, and user tracking) suggest that their belief in preparedness isn’t justified.
- There is a large blind spot about the financial impact caused by the post-cookie ecosystem: 76% expect that their revenue will not be impacted. However, McKinsey estimates that publishers can look at up to $10 billion loss in ad revenue.
- A lot of works needs to be done on the first-party data front: there’s currently minimal amount of first-party data being collected, and most are not being leveraged for advertising and marketing.
- There needs to be more and highly focused collaboration to implement privacy centric addressability solutions for targeting and measurement.
Understanding & Using Ethical Data Attitudes to Improve Consumer Relationships
Mark Truss — Chief Research Officer, Wunderman Thompson
Tom Corey — VP, Consulting Services NA, Wunderman Thompson
Brad Audet — CMO, Mazda North American Operations
Summary:
A Wunderman Thompson survey of 1,000 consumers each in the U.S. and U.K. found that concerns about data privacy rank 4th—behind only the economy, the COVID pandemic and healthcare. Most consumers see the collection and use of their personal data as an opaque process, creating a “negative halo” for companies they associate with this. Consumers generally do not perceive the benefits to them of sharing personal data and desire greater control—without necessarily knowing how this could be done. Mazda’s Brad Audet said that best practice in communicating about data privacy must start with a focus on the consumer, not the company’s data needs. Citing the example of Amazon, he urged that companies “co-create opportunities with customers” to use data for personalization that creates real value for them. Otherwise, their concerns will add to the current “cascade of distrust” in government and other large institutions.
Key takeaways:
- Consumers’ lack of understanding about how personal data is collected and used exacerbates their fear of the unknown.
- While a company’s use of such data to personalize offers (e.g., Amazon) can create value that is perceived and appreciated by consumers, use by third parties crosses a threshold.
- Companies must rely on their general reputation for integrity to overcome these concerns.
- Managing personal information is just one aspect of a company’s CRM.
- As Brad Audet says, “How a brand behaves is its entrée to the data.”
Understanding Audiences
Finding Mountain Dew’s Sweet Spot
Joe Conte — VP, Sales, Media Center of Excellence
Carlos Cruz — AVP, East Coast Sales, LoopMe
LoopMe’s Carlos Cruz and IRI’s Joe Conte review a case study for Pepsi’s Mt Dew using IRI’s real-time deterministic sales data and LoopMe’s in-flight optimization. Highlighting the precision of IRI’s retailer loyalty card data from 117M households (HH), Joe provided background on and execution of the campaign that increased dollars per HH sales lift by 40%. Aligning the right ads at the right time by leveraging IRI’s actual purchase data, Carlos outlined the attributes that drove brand lift and Pepsi’s sales goals to attract new buyers.
The campaign results from this data and optimization partnership drove effective sales lift, efficiency/ROAS, and a 20% lift in new buyer acquisition.
Leveraging A/B Testing to Understand Consumer Behavior
Vidyotham Reddi — Director, Mkt. Intel & Analytics, Mars, Inc.
Summary:
Vidyotham Reddi shared insights from Mars’ approach to A/B testing as their gold standard of learning. Framing it within the timely “virus” context, Vidyotham reinforced A/B testing’s dependability, versatility, and precision for marketing and understanding consumer behavior as part of Mars’ overall strategy.
Applying these learnings in eComm channels to predict overall sales, Mars created an ecosystem with intent to measure stimulus and response with A/B testing. Featuring a case study that included creating a digital A/B testing capability that informed overall multi-media pre-campaign and in-flight investment decisions, the presentation underscored how Mars’ collaborative success resulted in a new standard tool that measures multiple creative and content in cost-efficient ways.
Key takeaways:
- A/B Tests are the “Gold Standard” and the cleanest way of measuring ‘stimulus’ on sales.
- Access to a platform tying stimulus to actual purchase behavior affords tremendous flexibility and accelerates reaction time to trends.
- Leverage a platform like this, no matter which part of the globe, no matter how big or how small. Start with building for purpose, test it, try it out, and then scale it.
The Power of Big Data
Kimberly Gilberti — SVP, Product Management, Nielsen
Summary:
Big data’s power can only go so far, as cautioned in this presentation from Nielsen. Envisioning a future where big data’s integration in measurement is calibrated with panel assets, Nielsen’s Kimberly Gilberti addressed big data’s gaps in TV sources and usage, like CTV, video games, and smart TV’s native apps. This brief overview of how Nielsen uses big data detailed the enrichment needs for MVPDs and ACR in addition to the power of people-based panels that fill in the missing pieces.
Key takeaways:
- Granularity of big data provides added stability and ability to detect linear addressable ads.
- Big data in isolation is not census.
- Representative panels account for bias, enrich the data, and address gaps.
- Machine learning models trained on representative data deliver reliable and accurate demographics
Building a Cross-Device, Multi-Source, & Verified Behavioral Data Exchange Tool
Dana Budzyn — Head of Strategic Projects, YouGov
Alie Cirgenski — Head of Growth, YouGov
MODERATED TRACK DISCUSSIONS
Post-Pandemic Trends: Agility & Adaptability
Session Chair: Steven Millman — SVP, Research & Operations, Dynata
Summary:
This discussion addressed three presentations that described insights from surveys on how the pandemic affected consumer behavior during the pandemic and to what extent observed changes will persist. The presentations addressed trends in behaviors and attitudes regarding media (such as linear TV vs. streaming), travel, and also shopping, cooking, and education.
Key takeaways:
- Panelists agreed that the pandemic had changed behaviors and lifestyles, but most of those changes represented acceleration of existing trends and shifts in time spent, rather than completely new behaviors.
- Among the many changes, these stand out: the rapid adoption of streaming services, watching first-run movies at home (often for an extra charge) and online grocery shopping. At least to some extent, those changes will probably persist.
- Were there surprises? Despite the growth of smart TVs and streaming, most consumers have not abandoned linear TV, cable, etc.
- Finally, frequent surveys are a great way to track changes and discover trends, but we have to be careful not to “over-survey” consumers. Also, researchers should anticipate possible disruptions that could affect the methods they use for data collection.
Data Deprecation & Rising Privacy Concerns
Session Chair: Ryan Boh — Head of Activation & Identity, Product Strategy, Oracle
Summary:
All agreed that just because something is technically feasible does not mean that it should be done—or will be understood and appreciated by consumers. (Example: Concerns that Amazon’s Echo is always listening in the background.) As alternatives to the current practice of behavioral tracking (with or without consumer awareness and understanding), ask: What are replacement forms of data? How can these be collected? At what cost? How will you activate such data? YouGov’s surveys point to the need to compensate consumers in exchange for their personal data. Next year’s IAB survey can be a “report card” on how well companies are managing this transition.
Key takeaways:
- 70% of companies surveyed do not expect the loss of behavioral tracking data (aka “cookies”) to have a financial impact. (This seems naïve.)
- What are alternatives to today’s behavioral tracking and other personal data—and how will you collect, pay for and use such data?
- The advertising ecosystem needs interoperability across data sources and standard taxonomies for identification to allow recommendations across platforms (walled gardens).
- Companies need customer data platforms (CDPs) and machine learning/AI to take advantage of external data matched to their own first-party data.
- Data from streaming media can not only increase cost efficiencies but also help create better content.
Cross-Platform: Measurement & Identity
Session Chair: Sue Hogan — SVP, Research & Analytics, IAB
Summary:
In this moderated discussion for the track, Cross-Platform: Measurement & Identity, the speakers answered questions from the attendees about their respective presentations and discussed cross-platform challenges.
- Finding good data on minority audiences is difficult (Sue, IAB).
- Ethnic audiences skew a little more towards the invisibles segment. Possible reasons include more relocation by these consumers or changing devices frequently. (David, IRI)
- Audience targeting needs accurate measurement. Challenges exist due to GDPR, CCPA, cookie deprecation, etc. What solutions should be focused on in a test and learn environment? (Sue)
- There is a need to lean on opt-in identifiers and combining different onboarders to boost deterministic audiences. IRI has a few clients set up with clean rooms and plans to beef them up. (David)
- Jon Puleston from Kantar is spearheading changes in screening questions for global media consumption data, as presented during this Track. Is global standardization on our horizon? (Sue)
- Standardized global data would make it easier and more accurate to compare studies across countries. Studies would be repeatable. Standardization would enable a single measurement policy. (David)
- The number of times print issues are read and when they are read over time has been included in French print statistics since last year. It is the standard. For print audiences, MPX and velocity are being combined to generate actual “eyes-on” on multiple occasions and to produce “real” target audience impressions over time. (Gilles, Vintco)
- After years of discussion, we are seeing some industry consensus and forward movement. (Sue)
Understanding Audiences
Session Chair: Peter Sedlarcik — Managing Partner, Data & Analytics, HMG Health Practice, Havas Media
In a follow-up discussion for the “Understanding Audiences” track, Havas Media’s Peter Sedlarcik delves deeper into the ways the panelists are measuring for their clients, from the challenges of creating custom platforms and how technology’s rapid advances are affecting how they reconcile data, to balancing rigorous methodology with dynamic measurement approaches.
Key takeaways:
- From an adoption standpoint, more and more CPG brands are catching up to connecting brand and media metrics in real-time to purchase data as it is more readily available. (Joe, IRI)
- In applying learnings from A/B testing to predicting overall sales, Mars found one of its big challenges was the differences in brands’ penetration online and offline. Pay close attention to conversions offline. Leverage the upper funnel metrics with A/B testing—it’s cheaper and highly reliable. (Vidyotham, Mars)
- Nielsen looks at balancing rigorous methodology and dynamic measurement in three different contexts: resiliency, coverage, and comparability. Resiliency in the confidence that it will work now and, in the future, and be privacy-centric. Coverage in streamlining and onboarding as many data sources as possible for responsible measurement, while understanding strengths and weaknesses. And comparability in that the metric is apples-to-apples, enables speed, and is painless as possible (Kimberly, Nielsen).
PANEL
Identity Resolution Solution – How Close Are We?
Moderator: Alice Sylvester — Partner, Sequent Partners
Panelists:
Travis Clinger — Addressability and Ecosystem, LiveRamp
Jason Manningham — CEO, Blockgraph
Matt Spiegel — EVP Marketing Solutions, Head of Media Vertical, TransUnion
Summary:
Few would argue that a solution for identity resolution is needed to effect cross-platform video measurement. However, connecting different digital devices to unique users and households is fraught with legal, technical, and organizational barriers. How likely are we to see some kind of interoperable ID system that all industry participants can and will adopt?
Identity resolution is a difficult and organizational barrier for the industry. Alice (Sequent Partners) presented the following POV and asked for reaction from the panelists: We are not close to an identity resolution solution. We are not going to get there. Lots of data is needed for household targeting, such as third-party data and PII addresses. We cannot connect across every single consumer.
- Jason (Blockgraph) responded that he believes the industry will get to an identity resolution solution. The availability of signals (IP addresses, emails) may change, the scope will change.
- We will get close to the desired end state. The challenge is not building an ID scale or connection, but how it will be applied. The functionality and business applications are challenging, according to Matt (TransUnion).
- We will get there, according to Travis (LiveRamp). It is going to get much more complex with display, CTV, mobile, etc., plus the world is changing. It is not a tech problem, but a privacy problem. Regaining consumer trust is critical.
- Jason stated that as the media industry has fragmented, there are more players in audience targeting and measurement and different interpretations of privacy. It is complex. Unlikely there will be a global strategy and standards across the ecosystem. It’s a tech meets business challenge. There is a need to work across different constructs.
Will we be able to identify targets, optimize marketing plans, and measure against targeted audiences? (Alice)
- We are going to do all those things., according to Matt. The holy grail is for every impression to be measured and 1:1 targeting and measurement. We should not chase that goal. Micro-segmenting of small cohorts will be enough. Micro-segmenting is better than broad age and gender targeting. Manage limitations and make trust possible.
The appeal to do identity resolution is very strong. Opt-in subscribers are crucial. How many providers can provide scale? (Alice)
- It depends on the media platform. In the TV arena, there are a small number of vendors who can provide scale in terms of apps, linear, and addressable. (Jason)
- We will get to 1:1 personalization. Consumers will opt in and share ids. Not at 100%, but around 30%, which will be sufficient. More measurable than it is today, and a much better experience. (Travis)
Will that 30% be a biased sample of people who don’t understand their privacy rights and opt in? (Alice)
- Not a big risk, according to Matt. There will be three to five identity graphs in any market. Key tools will be interoperable. Plenty of consumers will be reached on a targeted basis.
- Jason agreed that the number will not be 100% for addressability, but accurate impression-based data is critical. A panel may be part of the solution, but not sufficient. Accurate identity resolution at the household level is critical.
- Marketers will pay more for that targeted impression. Media planning will change (Travis).
How do we develop best practices to measure accuracy? (Alice)
- Travis noted that best practices will be different for each marketer, for each brand.
- Matt noted that marketers should data test and to ask a lot of questions of partners about signals, sources, matching technology, and data science capabilities. Marketers must look under the vendor’s hood.
Licensing data from multiple sources makes it difficult to know about the identity of those sources. (Alice)
- Jason (Blockgraph) added that there is a scale vs. accuracy challenge. Need 3rd party services in certain circumstances.
- Even email matching can present challenges, Travis (LiveRamp) pointed out. Using identity graphs may help in many situations.
- Jason (Blockgraph) commented that the industry needs trusted organizations to provide 3rd party studies.
PANEL
What does the Converged TV Ecosystem Need From a “Calibration” Panel?
Moderator: Howard Shimmel — President, Janus Strategy & Insights, LLC
Panelists:
Caroline Horner — Chief Product Officer, 605
Molly Poppie — SVP, Data Science, Nielsen
Michael Vinson — Chief Research Officer, Comscore, Inc.
Summary:
Howard Shimmel of Janus Strategy & Insights, LLC. guided this expert panel through the critical questions surrounding the importance of calibration panels to the measurement industry. Howard asked why calibration panels are needed, what they bring to measurement, and what is sacrificed if the industry does not support them moving forward.
Key takeaways:
- Calibration panels fill in behavioral gaps and observe the whole relationship between persons, devices, and homes. Marketers want to see the overlap to understand how families use devices, concurrent usage, and co-viewing. The panels help solidify the full view of the consumer as an individual but also the total consumer UE – they retain the connections to the people within the household. (Caroline, 605)
- Marketers need something that measures across all behaviors—there are going to be populations that are not represented in big datasets that are unique – calibration panels increase accuracy and minimize bias. (Molly, Nielsen)
- Calibration panels go further to get brand effects and attitudinal metrics that are hard to extract from big data. For outcome measurement, panel surveys are indispensable. (Michael, Comscore)
- Not having calibration panels would mean a loss of control of forecasting, according to Caroline. There is also a risk of losing an understanding of people and how they interact – that will go away if you think you can depend on machine learning and large datasets by themselves. Michael states that we would lose the ability to see the underrepresented populations and do a better job of representing them.
THE LAST WORD
Josh Chasin — Chief Measurability Officer, VideoAmp
Jane Clarke — CEO, Managing Director, CIMM
Seth Duncan — Chief Data Officer, Real Chemistry
Mike Menkes — SVP, Analytic Partners
Summary:
There was a combination of both upbeat and downbeat takeaways from the anchor commentators on the last day of the conference. The following are edited highlights from the conversation:
- There’s dissonance between marketing and advertising and consumer expectations on privacy and identity, and Josh (VideoAmp) wondered if there will be a day of reckoning. He noted the paradox between tracking consumer behavior more granularly while protecting privacy. He argued that just because technology is evolving shouldn’t necessarily mean it should be inevitable, especially in regard to big data and its capabilities. He was pleased at seeing the consensus around taking individual data and aggregating it.
- It’s up to brands to make marketing and advertising a better experience for the consumer, stated Seth (Real Chemistry). He expressed some skepticism in the belief that increasing consumer literacy around their data will encourage them to opt-in. Jane (CIMM) agreed about better literacy being a lofty goal, but noted YouGov presenting that they were able get people to share all kinds of data in their panel, not only through incentives but a value exchange of providing customized viewing recommendations.
- We may never reach the holy grail of advertising measurement, but it will be good enough. Seth was struck by number of today’s speakers noting that measurement will never be perfect but it’s good enough where we will land. He also observed the increasing acceptance of combination of panel based insights projected onto large data sets. Jane agreed that all data has error in them – we need to figure out how to put them together.
- There are differences of opinions on the optimism for an interoperable ID system, observed Mike (Analytic Partners). He noted that while it’s reasonable that walled gardens can do this within their gardens, it’s unlikely to think that this valuable information will be shared broadly across digital ecosystems and offline environments. There are still a number of challenges that are being glossed over, but he thinks that there are different parts of the industry that are starting to really understand those watch-outs and the limitations. Mike concluded that we need to better understand why we want to have an identity (e.g., is it for targeting and ad-serving or is it for measurement?) and what those dynamics bring to the table.
CLOSING REMARKS
Scott McDonald, Ph.D. — CEO and President, ARF