hairong

[hairong]

AAA Interview of Hairong Li: My Research Experience

Dr. Anthony (HoYoung) Ahn, head of the American Academy of Advertising (AAA) communication committee, interviewed Dr. Hairong Li of Michigan State University, the 2022 recipient of the Ivan Preston Award for Outstanding Contribution to Research. Below are the interview questions and his answers.

Question 1: What does AAA mean to you?

AAA means a great deal to me for both personal growth and career development. I first attended the AAA annual conference in 1991 as a doctoral student, and I’ve since missed it only a few times. Over the years, I’ve come to know many of the AAA members. I always have the feeling of a big family reunion at each annual conference, especially during the outing. Socializing is always fun when catching up with old friends, making new friends, and chatting about our lives and things happening around us. Of course, I attend to share my research and to learn what new studies other members are doing for inspiring research ideas. As a result, many of my collaborative projects were initiated at the AAA annual conference.

For example, I got to know Professor John Leckenby of the University of Texas at Austin in the late 1990s when I was a junior faculty member, and our shared interest and passion led to the launch of the Journal of Interactive Advertising in September 2000, which has shaped my research thrust. Also, the AAA funded one of the research proposals I coauthored with my colleague Steve Edwards on the perceived intrusiveness of display ads, resulting in two Journal of Advertising publications, which have been cited over 2,000 times.

The AAA annual and global conferences also were the venue for several pre-conferences and special topics sessions that I chaired or co-chaired with other AAA members, like mobile advertising in 2004, big data in 2014, computational advertising in 2017, and AI for advertising in 2020. These are just some examples of the impact of AAA on my research. I’ve learned so much from these events. I don’t think my career would have been the same without being an AAA member for these years.

Question 2: When you reflect on the past 22 years of academic research regarding digital advertising, what are the findings that you found most interesting regarding the unique effects and implications of digital advertising?

There are so many significant findings in digital advertising research over these years, as highlighted in several recent literature review articles, including those published to celebrate the 50th anniversary of the Journal of Advertising. As digital advertising is an essential part of the digital economy, which plays a significant role in our society, any findings that promote the healthy growth of digital advertising are inherently intriguing to me. More specifically, since our studies on the perceived intrusiveness of display ads in the early 2000s, I have believed that the future of advertising lies in its creativity to offer value, not annoyance, to the consumer.

Thus, I’m interested in the findings that advance our understanding of how consumers perceive and respond to various digital advertising formats. Because digital advertising always runs on digital media, any findings that reveal the attributes, characteristics, and capabilities of the website, mobile app, social media, streaming video, and video games, as well as intelligent media (e.g., virtual assistants, wearables, and chatbots) and virtual media (e.g., AR, VR, and metaverse) are all fascinating. Further, many of these media and associated ad formats are increasingly powered by artificial intelligence and regulated for privacy protection; many studies on the role of AI and regulation in advertising are also appealing to me. A customer-centered approach to studying various uses should produce the most relevant and valuable findings to advance our discipline. Among the recent remarkable are the studies on native advertising and influencer marketing that show these brand messages are less intrusive and more entertaining and thus more effective advertising formats, studies on targeting approaches and methods using anonymous audience identities, studies on strategic consequences of customer privacy concerns, studies on the use of AI for deeper consumer insights and advertising creativity, and studies on the functions of emerging media technology in enhancing real and virtual customer experiences. Of course, there are many more studies of interest beyond this shortlist.

Question 3: When we consider the gap between academic advertising research and the needs of the advertising industry, do you see interesting opportunities for future research?

Academic advertising research has advanced dramatically in recent years, but the gap remains. To address this challenge, I moderated a special topics session on bridging the gap in advertising research between academia and practice at the 2006 AAA annual conference. Although academic research and applied research are supposed to generate knowledge about advertising, each side has its agenda, as the panelists pointed out. Advertising scholars tend to improve the theoretical understanding of the advertising discipline and the precision of research methods, compete for promotion and tenure, and elevate the reputation of their programs. In contrast, advertising managers seek to understand the interactions of various market forces, develop solutions to real-world advertising problems, and generate brand sales and revenues. One of the ideas from that session was that, even if the agendas are different, advertising scholars can do more for their research to be more relevant to those who are trying to make a living in practice, including attending industry events, interacting with managers, collaborating with practitioners, and making research findings more accessible to advertising professionals.

In 2012, we received a grant from Marketing Science Institute in Boston to develop a website for sharing usable research findings. I worked with my colleague Bruce Vanden Burgh and our students to select and curate recent academic articles in the hope that the findings that are easy to understand and use for advertising professionals may have a more significant impact. Due to the lack of continuous funding, the project lasted only two years and ended quietly. To this end, I’m glad to see in recent years, some academic journals have disseminated research findings through podcasts, online webinars, and other formats to extend the influence of academic advertising research to a greater community, including advertising professionals.

Academic studies should be more relevant to advertising professionals. Recently, many online webinars hosted by trade organizations in collaboration with adverting technology companies to address emerging issues, such as the upcoming deprecation of third-party cookies, data clean-rooms for consumer privacy, connected consumer identity, AI-based audience targeting, branding in the metaverse, and empathy for CX success. Some of these issues are worth academic research and have the potential for theoretical advance. Research on these issues may help bridge the gap between academic research and advertising practice.

Question 4: What tips or advice do you have for graduate students and promising/emerging scholars and professors?

Many factors help the development of a successful research career. Based on my experience, I have something to share. First, be forward-looking when selecting research directions and topics. Digital advertising is dynamic and calls for timely research on new, engaging, and relevant topics. Thus, it’s essential for advertising researchers, especially young scholars, to keep abreast of the industry evolutions to identify and explore the topics that have become increasingly important in theory and practice. For example, many future-oriented studies, including our articles on display advertising, virtual experience, and mobile advertising, have received more citations in recent years partly because they addressed subjects still current or pertinent to emerging issues.

Second, taking a multidisciplinary approach to research projects. While psychological concepts and theories remain significant in conceptualizing many phenomena of digital advertising, data science and computer science are more critical in digital advertising, especially in programmatic, computational, and intelligent advertising. Integrating theories and methods from different disciplines into digital advertising research can help open new research territories and result in breakthrough findings. For example, in a recent article published in a major marketing journal, my coauthors and I reviewed and integrated the literature from several fields, such as computer science, psychology, robotics, and communication to develop a conceptual framework for artificial empathy in marketing interactions. The conceptualization and proposition development was only possible with being multidisciplinary.

Third, collaborating with other researchers. Doing rigorous research is often challenging, so working with those with common interests, unique perspectives, and complementary skills can make the research more productive and fun. A good team will motivate each other and mitigate possible frustration from research difficulties. I’m fortunate to have had many coauthors over the years, and I always cherish the days of working as a team member on different projects and sharing the excitement of seeing our articles in the publication.

Besides these suggestions, the most important are hardworking and perseverance. Developing a research program and carrying it out like operating a production line, with some projects in ideation, some in data collection and analysis, some in writing, and some in submission and revision. Submit each paper to the most suitable journal, never give up when a paper is rejected, embrace an invite for even a risky major revision, and take the reviewers’ comments seriously and revise accordingly. I hope these ideas can help graduate students and emerging scholars become more productive and enjoy more of their research.

(Originally published on the AAA website in August 2022)

[hairong]

Metaverse: The Role of Virtual Affordances

The metaverse has become a new star in the digital media galaxy since Facebook renamed itself to Meta Platforms in October 2021. It is intriguing to many businesses and consumers, and one of the reasons is: “You can do things that you couldn’t do in the physical world,” as Nicola Mendelsohn of Facebook said (Savitz, 2022). We observed this potential when we explored virtual experience in e-commerce twenty years ago, which we termed “virtual affordance” (Li, Daugherty, & Biocca, 2002). Compared to early 3-D advertising, the metaverse can simulate more immersive virtual experiences. What makes a virtual experience unique is often the role of virtual affordances. As the call for more studies of advertising in the metaverse is put out (Kim, 2021), virtual affordances have the potential for new theories of virtual experiences in the metaverse. To this end, let me review some issues on virtual affordances in the hope that such a review may be useful for future research.

The affordance concept was conceived by psychologist James J. Gibson. In theorizing the relationships between the environment and an animal, he (1977) suggested that the affordance of anything is a specific combination of the properties of its substance and its surfaces in reference to an animal. Gibson (1979) wrote: “The verb to afford is found in the dictionary, but the noun affordance is not. I have made it up” (P. 119, italics in original). Affordances gained popularity after the design theorist Donald A. Norman made a distinction between real or physical affordances and perceived affordances. Norman (1998) wrote:

The set of possible actions is called the affordances of the object. An affordance is not a property, it is a relationship that holds between the object and the organism that is acting on the object. The same object might have different affordances for different individuals…. In the design of objects, real affordances are not nearly so important as perceived ones; it is perceived affordances that tell the user what actions can be performed on the object and, to some extent, how to do them. (p. 123, italics in original)

For example, a door knot is a real affordance, and turning it to open the door is a perceived affordance, which little babies may not have acquired.

To study how consumers interact with and respond to 3-D product visualization in e-commerce in the early 2000s, we conducted a series of experiments on different products. Geometric products are those whose purchase decisions only need a visual inspection, such as paintings and sculptures; material products are those whose selection often involves touching, smelling, or tastings, such as sweaters, perfumes, or wines; and mechanical products are those that trialing is preferred prior to purchase, such as bicycles and cameras. In an experiment with a virtual engagement ring, the participants were amazed when they found it possible to engrave someone’s name in the inner band to see the effect immediately, something not easy to do with a real ring. This finding prompted us to propose the concept of virtual affordances, a set of possible actions of a virtual object (Li, Daugherty, & Biocca, 2001). Our follow-up studies (Li et al., 2002; Li, Daugherty, & Biocca, 2003) confirmed that, when virtual affordances exceed real and perceived affordances, a virtual experience is more likely to generate positive product attitudes and purchase intentions. For example, a jacket was a material product by our definition, which would require haptic inspection by hand for a confident purchase decision. When it was simulated in 3-D visualization, however, many participants reported the intent to purchase after viewing it in different colors and styles, with no need to touch a real jacket.

In 3-D advertising two decades ago, the ability to design and simulate virtual affordances was limited. Now, blending the latest technologies of augmented reality and virtual reality on more powerful computing platforms with massive spatial and temporal data, the metaverse can simulate almost infinite virtual affordances. In a typical metaverse scenario, people in the form of avatars interact with other avatars or virtual objects in simulated spaces, like designing a product with coworkers in the office, playing cards with friends at home, or shopping with kids in a strip mall, all virtually. These digital avatars, virtual objects, and simulated spaces can be designed to have various virtual affordances, so imagination seems the only limit in rendering virtual experiences in the metaverse.

Interestingly, Neal Stephenson, the novelist and author of Snow Crash that coined the term  “metaverse” in 1992, was interviewed at the 2022 IAB Annual Leadership Meeting (IAB, 2022). When he was asked how the metaverse today is different from the metaverse in his novel, he said there are several differences. Many metaverses have emerged while his novel had only one metaverse, so interoperability could be an issue, and each metaverse has different ways to play and different revenue models. He noticed that some games in a metaverse allow the users to buy virtual goods, like gears and weapons, to equip themselves in the game, but such purchases would not alter the probability for these users to win a game. That is important because it would be unethical for “rich” users to win all the time. This is a caveat, as ethical and even legal considerations are essential for the design of virtual affordances in the metaverse.

Among the popular metaverses are Decentraland, Horizon Worlds, Roblox, and the Sandbox. Many leading brands have established a presence in these metaverses, including Adidas, Burberry, Coca-Cola, Nike, Louis Vuitton, and Samsung (Hoogendoorn, 2022). More recently, Estée Lauder participated in the first Metaverse Fashion Week in Decentraland, and Wendy’s and Chipotle each opened a virtual restaurant in Horizon Worlds and Roblox, respectively (Alcántara, 2022). It is likely that some brands are performing better than others in these metaverses, and a tipping point could well be the virtual affordances of these brands—are they exceeding, matching, or trailing the real and perceived affordances?

References

Alcántara, Ann-Marie. (2022, April 5). Restaurants’ Virtual Stores Test Consumers’ Appetite for Metaverse Marketing. The Wall Stree Journal. Retrieved from https://www.wsj.com/articles/restaurants-virtual-stores-test-consumers-appetite-for-metaverse-marketing-11649160001

Gibson, James J. (1977). The Theory of Affordances. In Robert  Shaw & John Bransford (Eds.), Perceiving, Acting and Knowing: Toward an Ecological Psychology (pp. 62-82). Hillsdale, NJ: Lawrence Erlbaum Associates.

Gibson, James J. (1979). The Ecological Approach to Visual Perception. Boston: Houghton Miffin.

Hoogendoorn, Robert. (2022, January 12, 2022). 10 Major Brands Stepping into the Metaverse. Retrieved from https://dappradar.com/blog/10-major-brands-in-the-metaverse

IAB. (2022, February 9) A Fireside Chat with Neal Stephenson/Interviewer: Anthony Katsur. IAB ALM.

Kim, Jooyoung. (2021). Advertising in the Metaverse: Research Agenda. Journal of Interactive Advertising, 21(3), 141-144. doi/10.1080/15252019.2021.2001273

Li, Hairong, Daugherty, Terry, & Biocca, Frank. (2001). Characteristics of virtual experience in electronic commerce: A protocol analysis. Journal of Interactive Marketing, 15(3), 13-30.

Li, Hairong, Daugherty, Terry, & Biocca, Frank. (2002). Impact of 3-D Advertising on Product Knowledge, Brand Attitude, and Purchase Intention: The Mediating Role of Presence. Journal of Advertising, 31(3), 43-57. doi/10.2307/4189225

Li, Hairong, Daugherty, Terry, & Biocca, Frank. (2003). The Role of Virtual Experience in Consumer Learning. Journal of Consumer Psychology, 13(4), 395-407.

Norman, Donald A. (1998). The Invisible Computer: Why Good Products Can Fail, the Personal Computer Is So Complex, and Information Appliances Are the Soluation. Cambridge, MA: MIT Press.

Savitz, Eric J. (2022, April 25). Facebook Is Broken. Execs Say a Fix Won’t Come Fast. Barron’s. Retrieved from https://www.barrons.com/articles/facebook-broken-metaverse-ads-tik-tok-fix-51650651013

(Originally published in the American Academy of Advertising Newsletter, June 1, 2022)

[hairong]

Artificial Intelligence for Advertising Research

AAA 2020 Pre-Conference, San Diego, CA March 26, 2020 at 1:00-4:30 p.m.

One of the latest advances in advertising is the rise of artificial intelligence (AI) in recent years. Comprising largely machine learning, natural language process, artificial neural network, voice recognition and computer vision, AI has great potential for betterment in many aspects of advertising, such as consumer insight generation, advertising creativity, media planning and execution, and campaign optimization and effectiveness assessment. It is likely to add more humane elements to the increasingly popular programmatic advertising, which still generates large amounts of irrelevant and annoying ads. Thus, AI has become a new subject for advertising research and education.

Growing interest in AI and advertising is already manifested in the advertising community. A special topics session on the role of AI in advertising at the 2019 American Academy of Advertising conference, chaired by Harsha Gangadharbatla, attracted a roomful of audience even though it was on Sunday morning and generated lively discussion among the attendees and panelists. A special section of the Journal of Advertising on AI and Advertising, guest edited by Hairong Li, was released in September 2019. The special section consists of four articles that cover the impact of AI on the process of advertising, the role of AI in advertising creativity, an algorithm for personal ad creation, and a recommender system for two-sided media platforms. Intelligent advertising is defined as consumer-centered, data-driven, and algorithm-mediated brand communication in the special section. Shelly Rodgers, incoming editor in chief of the Journal of Advertising, recently distributed a themed issue call for papers on promises and perils of artificial intelligence and advertising. The call states that AI research is gaining momentum in many fields such as marketing, communication, psychology, technology and health, but research on AI and advertising is limited, suggesting significant opportunities.

To address the need for ideas to stimulate research on AI and advertising, this half-day pre-conference will explore some fundamental issues, such as most advertising-relevant AI technologies, intelligent advertising as a hybrid of salesmanship and salesmachineship, the role of AI in consumer experience, AI-powered creativity, personalization, and the effectiveness of recommender systems vs. human recommendations. Seven thought leaders who have done research in these areas will share their studies and perspectives. They are Harsha Gangadharbatla, University of Colorado, Boulder; Louisa Ha, Bowling Green State University; Jisu Huh, University of Minnesota; Scott Koslow, Macquarie University; Cong Li, University of Miami; Yuping Liu-Thompkins, Old Dominion University; and Guohua Wu, California State University, Fullerton. 

For more information, please contact the pre-conference chair Hairong Li, Michigan State University via email: hairong@msu.edu.

[hairong]

Eight Aspects of Interactive Advertising

Reflections and Expectations from the Founding Editors: Hairong Li and John Leckenby

Interactive advertising has experienced a sea of change since the journal was launched. These data present an opportunity to reflect on this transformational process while calling attention to what we believe are eight aspects of interactive advertising that should be covered within JIAD in the future. Two of these were the initial focus of the journal, which are interactivity and modality. A key difference between interactive advertising and conventional advertising is the level or degree of interactivity. Subsequently, JIAD has helped distinguish three dimensions of interactivity over the years, including the depth, width, and speed of response, which is the degree of speed a user may customize a message. Interactivity has also evolved from clicks to touchscreens to gestures in virtual reality, with endless new ways emerging. Regarding message modality, a parallel development has taken place from text, graphic, audio and video, to 3D, virtual reality, and mixed reality. Thus, the journal has adequately explored the effects of interactivity and modality.

Advances in technology involving interactivity and message modalities have also ushered in the emergence of new devices, which began with personal computers (PC) and mobile phones, only to make way for smartphones, tablets, interactive TV, outdoor LED/LCD, and, more recently, wearables, such as Apple Watch and Samsung Gear. In fact, new or alternative forms of interactive advertising always seem to coincide with the acceptance of new devices. For example, mobile advertising, or what was commonly referred to as text ads, which started with short message service (SMS), has become almost obsolete today. In a similar way, the theory of virtual experience in nonimmersive 3D visualization as simulated on the PC has been updated involving immersive-mix reality on wearable devices, like Oculus Rift and HTC Vive. Digital media and information technology–based devices will continue to innovate the way users engage with content, including advertising messages.

Another important aspect of interactive advertising not to be overlooked is that of context. With the rise of mobile computing (e.g., laptops, smartphones, tablets, wearables) location-based services (LBSs) have become a natural feature of mobile apps and mobile advertising. By considering time and location as the macro context of interactive advertising, identifying editorial messaging of the ads as the micro context becomes natural. Better congruency of interactive advertising with its micro context gives rise to what is called “native advertising,” which is advertising that is very similar to its surrounding messages in both content and format. In addition, the effectiveness of native advertising is often enhanced by personalization, which is discussed next. As such, new research on both macro and micro contexts is highly recommended.

Related to the context is the platform in which advertising appears. Interactive advertising first appeared on websites, which were populated with publisher-generated content. New genres of platforms have since emerged in that each mobile app or social media outlet constitutes a platform. Platforms within a genre then share common affordances as well as possess unique affordances. For example, Facebook differs from Twitter in terms of why and how consumers use these platforms, despite the fact that both are saturated with user-generated content. Thus, a native ad that was originally created for Facebook would not be the same when it is transplanted to Twitter or Pinterest; each native ad must be developed in light of a platform’s unique affordances. More important, as the use of social media outlets has increased in our daily lives, users have formed various virtual groups. A networked society has been established with its fundamental ramifications for advertising not fully understood yet and certainly requiring more exploration.

When the journal first began, display advertising was firmly in place as the dominant form, so advertising delivery received little attention. Partly because banner ads were first “hard coded” in an HTML page and then linked and served from ad servers, with minimal programmatic activity compared with today’s standards. A few years later, though, search advertising emerged from obscurity given the popularity of Google AdWords and marked the beginning of computational advertising. The journal published one of the first articles on Google AdWords when the Google Online Marketing Challenge (GOMC) started in 2008. Computational advertising, which takes advantage of user profiling, algorithms, and various tools, has evolved from auctions involving search/display advertising to programmatic buying of many forms, including within digital television. The use of computational methods has not only changed the way advertising is placed but also how advertising is created, resulting in the emergence of the computational creative practices richly in need of further study.

Like advertising delivery, a focus on data or what is commonly referred to as “big data,” is a natural byproduct of digital media. Such online data can be overlaid with offline transactional information to develop richer user profiles so that ad delivery is more precise. In addition, online data, especially involving social media, have been increasingly utilized for consumer insights in almost every phase of business, such as new product development, advertising campaign planning, promotional offer design, and service quality assessment. For researchers, big data have opened a new route of discovery and, along with “small data,” allowed scholars to achieve the breadth and depth of exploration in almost all areas of advertising in unprecedented ways.

The final aspect that needs further consideration is the outcome of interactive advertising. With big data, outcomes of an ad or campaign can be assessed not only on a large scale but also in more granularity. For example, search advertising can be analyzed in terms of the path to purchase, including exposure, clicks, navigation patterns, conversions, purchase, and even return. Such a dynamic process exists in all forms of digital media and deserves more research. Further, the engagement of social media advertising often consists of static, one-shot measures, such as likes, comments, and shares. In contrast, it is important to point out that engagement is a process, which can be better illuminated with interactive advertising. For instance, on WeChat, the most popular social media platform in China, an ad allows users to see likes and comments along with emojis posted only by their friends, and when new likes or comments are added, the users will see a notification on the top of the newsfeed, often resulting in repeated interaction, thus enhancing the effect of the ad. Such a dynamic process can be revealed in field experiments in collaboration with service providers designed to measure both positive and negative outcomes of interactive advertising.

In summary, because the many aspects of interactive advertising continue to change, research exploring these areas must advance accordingly. New forms of interactive advertising will no doubt emerge in the years to come. If we can observe the latest developments of these eight aspects—interactivity, modality, device, platform, context, delivery, data, and outcome—while carrying out innovative research that extends our knowledge, we can aid in the healthy growth of the discipline of interactive advertising. It is essential to keep in mind that these aspects are often intermingled in any given interactive form. Thus, isolating the impact of one aspect while holding other aspects constant is oftentimes conceptually necessary. Ideally, though, examining the impact of interactive advertising in a natural setting using a combination of big and small data is likely to make new discoveries involving these aspects possible, which should be the goal of research moving forward. We certainly hope the journal will continue to play a distinctive role in the future, as it has in the past, and elevate the paradigm of interactive advertising research to new heights.

Excerpted from Terry Daugherty, Vanja Djuric, Hairong Li & John Leckenby (2017), “Establishing a Paradigm: A Systematic Analysis of Interactive Advertising Research,” Journal of Interactive Advertising, 17(1), 65-78.

[hairong]

Digital Analytics: Using Google’s AEK in the Classroom

The profession of advertising has seen a paradigm shift in recent years in that big data is playing a greater role in the creation, placement and evaluation of advertising, especially in the ways of consumer engagement. Consumer insights through digital analytics, real-time in some cases, are shaping almost every aspect of advertising practice. As a result, the knowledge about and skills in collecting, processing, managing, analyzing and reporting digital data have become a sought-after competency of advertising professionals and advertising students. Big data related topics such as computational advertising, social listening, programmatic buying and advertising analytics are introduced in more advertising programs in the US and other countries. Teaching digital analytics in the classroom, however, is challenging because shareable materials to build a fundamental understanding of analytics, including realworld business data and the opportunity to perform a diverse range of analysis using an analytics tool, are not easily accessible to most instructors and students.

Analytics Education Kit

The Analytics Education Kit (AEK) by Google provides a number of resources to overcome this challenge. Online videos featuring industry experts are provided that help viewers build a thorough understanding of digital analytics. Access is given to an analytics account with live data from a real ecommerce business, which has ongoing transactions that are facilitated with advertising campaigns. The live data allows students to use analytics to explore how consumers interact with the business’s website, including the effects of advertising and the paths to purchase. The AEK also provides real world analysis exercises for students to perform that will make them familiar with common features in analytics tools and how business insights can be derived from analytics data. In addition, the AEK includes guidance notes for instructors so they are knowledgeable about the business and can assist students through the exercises.

Google Analytics is the primary analytical tool part of the AEK. It is considered the most popular analytics tool in the world with over 10 million websites using it. It has been available since 2005 and has continually evolved to meet the growing analytics needs of businesses. Given its popularity and availability, it is likely students will go onto use the tool for their own business or for another business. The theoretical knowledge and practical skills students gain from the AEK not only prepares them to use Google Analytics within a business but also prepares them to use other digital analytics tools. Even if they use another analytics tool, the AEK helps to build knowledge and skills in students that they can apply to other analytics tools. Taking this into consideration the AEK is an excellent program for teaching digital analytics.

Use of AEK in the Classroom

I was invited to test the AEK in its pilot stage. I integrated it in my media planning and interactive advertising courses in the spring of 2016, and based on students’ feedback, I developed a new online course of digital analytics using the AEK in the summer of 2016, with the help of Google executives Deepak Aujla and Jingtao Ji. Most students who finished the course reported having enjoyed the course and learned a great deal. A student commented at the end of the course, “I recently met with a potential employer and they were highly intrigued to know that I had experience with Google Analytics. I envision this class moving to the position of a core class for marketing and advertising students in the future.”

As we know, digital analytics is a large and complex subject and includes site analytics, mobile analytics and social analytics. With the AEK, my course focused on the concepts and uses of site analytics but it also covered parts of mobile analytics and social analytics that were directly related to site analytics. For example, traffic sources from mobile devices and social media were explored, as they are becoming increasingly important in understanding multichannel users.

The online course consisted of seven modules, one module per week. The first module defined digital analytics, its role in business and several key concepts. The second module introduced the ways Google Analytics collects, configures, processes and reports data as well as the Google Analytics account from the AEK to students. Students familiarized themselves with the interface of Google Analytics and the live data in the account. The next four modules covered four main sets of analysis—audience analysis, acquisition analysis, behavior analysis and conversion analysis. The last module introduced advanced analytics, such as filter use, custom reports and dashboard design, in addition to how to create a Google Analytics account and integrate Google Analytics into a website or a mobile app.

Each module included a lecture, a quiz and a writing assignment. The lecture used many short video tutorials from the AEK featuring Justin Cutroni, Analytics Evangelist at Google, and step-by-step exercises that were developed by myself to explain important topics and demonstrate how specific analyses can be carried out. Each quiz consisted of ten questions that sampled the lecture content and additional readings, which allowed students to see how well they mastered the module. The writing assignment was an essay for the first two modules and the last module, and was a report for the third to sixth module, which each reported the results of analysis for three required questions and two elective questions.

For example, the report for the third module was on audience analysis, which included identifying the ten countries besides the US and Canada that had highest sessions in May 2016, the changes in the sessions and revenues between April and May 2016, and the sessions and revenue from mobile devices in May 2016, as well as two elective questions that the student believed were most intriguing as they focused on helping the owner of the website grow their business. Students were able to use these questions to demonstrate their creativity in digital analytics. Other reports had similar types of questions on the theme of each module.

Higher Level of Skills

As part of the essay for the last module, students were asked about what else they would wish to have learned from the course. Some students reported that, beyond the skills of learning to use Google Analytics to extract relevant information and answer specific questions, they also wanted to know more about how businesses, big and small, use digital analytics to solve daily issues of operation. Such additional cases have now been included in the AEK.

A few students also indicated, as they planned to run a business in the future, analytics skills were essential for them to set up measurable goals for success. One student commented: “Understanding Google Analytics and learning about its main functions has allowed me to not only comprehend and analyze data that I am given, but it has also taught me how to search, transform and problem solve with data that I am generating. The use of Google Analytics has proved to make it easier and more convenient to let your website or app generate its data from the users and compile it for the business owner in an organized and timely manner.” My reading of this comment is that thinking more like a business owner should be the mindset of a digital analyst and our teaching should nurture that mentality.

If you are interested in learning more about the AEK from Deepak Aujla please complete the form available at https://goo.gl/mn9awh.

(Deepak Aujla and Jingtao Ji of Google contributed to this article. Originally published in the September 2016 issue of the AAA Newsletter)

[hairong]

Big Data and Small Data

Innovative paths to integration of big data and small data are explored in this blog. 

Big data is one of the most important developments in our field in recent years. It has not only driven the research on consumer insights and the proliferation of precision marketing, but it has also affected many aspects of the advertising business, research, and education. With the increasing popularity of big data in our profession, however, we have seen a surprising lack of discussion on the integration of big data with “small data,” the type of data that have been used in advertising business, research and education for years. Although some practitioners have argued that small data remain critical, little conceptualization is available on the issue of big data and small data integration. Thus, the purpose of this analysis is to explore the differences between big data and small data, the innovative paths to their integration in light of current practice, and the importance of integration for the future of advertising research and education.

Differences Between Big Data and Small Data

Vast amounts of user data are cumulated as a byproduct of normal consumption of interactive media. When users visit a Web site, search for a product online, post a note on Facebook, make an online purchase, or use a location-based service from a smartphone, they leave their digital footprints in these media and constitute what is called big data (Li and Huang, 2013). Other types of big data can be membership data, data from telephone calls, data from patients, etc. Big data can be a combination of different types of data such as combing membership with media usage. In comparison with small data, big data has strengths and weaknesses from an advertising researchers’ perspective. Although there are no commonly accepted definitions, we refer to small data as data that are collected for particular purposes from surveys of subject samples. Examples include consumer panels, satisfaction surveys, and various public opinion polls. As presented in Table 1, small data and big data differ in several aspects.

Innovative Paths to Integration

As discussed in the previous section, big data and small data each have their characteristics, and they should be integrated to supplement each other. In reviewing the current business practice, we have identified five paths to integration: namely conception adoption, matched combination, cross calibration, coordinated creation, and synergistic consolidation.

Concept Adoption. Concept adoption refers to modifying conventional concepts that are good for small data to use for big data. There are many concepts that are developed with small data have proved valid in advertising, such as gross rating point (GRP), effective frequency, and audience fragmentation, and these concepts can be enriched with the integration of small and big data. For example, GRP has been used in television advertising for decades. It is based on sample data and is thus a small data concept. Since 2009, AdMonitor, a service of Miaozhen Systems in China, has used a hybrid mechanism of integrating an online panel with Internet population tracking to effectively evaluate the reach and frequency of online advertising, resulting in a new concept of Internet gross rating points (IGRPs). The new measure has proved to be capable of significantly increasing the efficiency of advertising delivery, especially for online video ads. Nielsen and ComScore have offered similar measurement services in the United States. Another concept is customer satisfaction, which has been traditionally measured with surveys, but IBM is reportedly replacing customer satisfaction surveys with the increased use of tracked behavior and word of mouth of customers.

Matched Combination. Matched combination refers to adding information from big data to a single-source panel in order to offer new insights on consumers. For example, we know the demographics, brand attitudes, and purchases from surveys. At the same time, with tracked data on media usage, ad exposure, and search strings for the same panel members, we can combine variables from small data and big data to generate a richer bank of information and build various media decision models. In practice, digital television program viewing behaviors, which are recorded automatically with service providers, can now be overlaid with purchase behaviors of the households and even individual members, and such matched data is able to reveal the true impact of television commercials. In a data mining study by Li, Tan, and Zhao (2009), tracking data of display ads, search strings, and survey data were integrated to demonstrate the influence of exposure to display ads on category and brand search behavior, category need, and purchase intent of a sample of Internet users in China.

Cross Calibration. Given the possible biases of both data types, efforts must be made to integrate big data with small data to minimize potential biases of either data type. This is what we refer to as cross calibration. Millward Brown’s Verve is a way to harness big data alongside traditional research techniques. It enables an analysis of social commentary in parallel with traditional survey data sets. This, in turn, allows researchers to understand the interaction with other elements of the brand experience and marketing campaigns. Such cross calibration can help reduce possible biases in big data (Hollis, 2012).

Coordinated Creation. Coordinated creation refers to creating something new by integrating small data with big data in coordinated ways. For example, brand equity is always considered to exist in the consumer’s mind and consist of brand experiences in consumers, including knowledge, loyalty, and actual purchase. When online brand interactions such as brand searches, comments, and reviews have increased dramatically, this digital content becomes part of the equity of a brand. Baidu, the top search engine in China, and Millward Brown have cooperated to integrate brand measures that are developed in small data with all the digital brand information Baidu has captured. This effort has created a new service, a “brand digital equity leaderboard,” as a comprehensive measure of brand value for hundreds of major brands, consisting of multiple indicators.  

Synergistic Consolidation. Synergistic consolidation refers to sophisticated ways to combine information in multiple sources and forms to generate better insights on consumers, media strategy, and brand promotion. For example, Millward Brown has developed the Mingboard application, which can automatically integrate assorted data to show key aspects of brand intelligence in real time, including ad spending, brand metrics, search analytics, social graph, and sales, among other indicators. Such consolidation helps garner the true value from large volumes of big data and small data.

Future Integration

The innovative paths to integration between big data and small data have largely emerged from the advertising business. Academic researchers in advertising and marketing are certainly in pursuit of such innovative ways to integrate big data and small data for new discovery. However, the proprietary nature of big data makes it difficult for academics to access; thus, we need to explore and solve the access issue. Academic associations have a role to play in building “data gardens” for member researchers to use, where they can benefit from big data. Big data is here to stay and grow, along with the rise of the Internet of Things, as more devices are interconnected in our society and people spend more time on digital media and content. Thus, the future of advertising will depend on how well we embrace this important advance in our profession.

(Excerpted from a paper written by Hairong Li and Peking Tan and presented at the ICORIA 2014).

[hairong]

Call for Papers: Big Data in Advertising

Journal of Advertising Special Issue on Big Data in Advertising

Manuscripts are currently being solicited for an upcoming special issue of the Journal of Advertising (JA) dedicated to using big data in advertising research.

Digital environments such as the Internet, social media, mobile devices, wearable technology, and the Internet of things produce large data sets by recording, often in great detail, interactions between customers and the brand throughout various phases of the customer experience. These data sets create opportunities for advertisers, have the potential of disrupting aspects of the advertising industry, and become a goldmine of information for academics to test advertising theories.

The purpose of this special issue is to stimulate research into using these new data sources in advertising. More specifically, this issue aims to explore how big data contribute to theory building in advertising research. Multidisciplinary collaboration between advertising scholars and scholars from data-oriented fields is encouraged. Potential research topics that may be addressed include (but are not limited to):

Identifying customer insights using big data 

New methods of measuring audiences using big data sources

Measuring and monitoring brand associations from social media

Data mining and text analytics in social media

Personalizing messages with big data

Testing advertising theories with big data sources

Integrated use of big data and small data

Optimizing the selection of media delivery channels and platforms

Using big data without violating customer privacy or ethical standards

Big data in qualitative advertising research

Submission Guidelines Submissions should follow the manuscript format guidelines for JA at http://www.tandfonline.com/toc/ujoa20/current. Papers should be no longer than 30 double-spaced pages in length (including references, tables/figures, and appendices).

The submission deadline is January 31, 2016.

All manuscripts should be submitted through the JA online submission system, ScholarOne, at http://mc.manuscriptcentral.com/ujoa, during January 1-31. Authors should select “SPECIAL ISSUE: Big Data in Advertising” as “Manuscript Type.” Please also note in the cover letter that the submission is for the Special Issue on Big Data. Manuscripts will go through a peer review process, and the special issue is planned to appear as the last issue of 2016.

Guest Editors Edward Malthouse, Northwestern University Hairong Li, Michigan State University

For additional information regarding the special issue, please contact the guest editors at ja.guest.editors@gmail.com.

[hairong]

Computational Advertising

Computational advertising refers to an approach of advertising to serve the right ad to the right consumer in the right context based on the deep consumer knowledge that is enabled by computational methods.

A technical definition of computational advertising is “a new scientific sub-discipline, at the intersection of information retrieval, machine learning, optimization, and microeconomics. Its central challenge is to find the best ad to present to a user engaged in a given context, such as querying a search engine (“sponsored search”), reading a web page (“content match”), watching a movie, and IM-ing” (http://labs.yahoo.com/areas/?areas=computational-advertising). Andrei Broder and Vanja Josifovski, both of Yahoo Research, co-taught a course “Introduction to Computational Advertising” in fall 2009 at Stanford University (http://www.stanford.edu/class/msande239/). A panel at the 2014 AAA Pre-Conference on Big Data in Atlanta focused on computational advertising, representing one of the early dialogs on this subject among advertising educators.

What It Means to Advertising Educators Computational advertising seems to be Yahoo engineers’ term of what “interactive advertising” means to advertising educators. In other words, we may consider computational advertising as “an approach of advertising to serve the right ad to the right consumer in the right context based on the deep consumer knowledge that is enabled by computational methods.”

We are no strangers to the first element of the definition—“serving the right ad to the right consumer in the right context,” as advertisers have always intended to do that. We are familiar with part of the second element, too, which is “consumer knowledge.” Advertising has always been run on what marketers know about their target audiences. 

Deep Knowledge by Computational Methods  What is really new in computational advertising is the last element—the “deep” consumer knowledge that is enabled by “computational methods,” which is what engineers, mathematicians, and economists at Google, Yahoo, Microsoft and other digital media do so well for living.

Computational methods are a mix of information tracking (e.g., cookies, registration, and personalization), mathematical modeling (e.g., PageRank, quality scores, and auctions), and computer applications (e.g., AdWords, AdSense, Analytics). These methods make it possible to obtain real-time feedback, speed up the whole process, and maximize the precision. More recently, Facebook integrates its online user knowledge with offline user knowledge through cooperation with third-party vendors, resulting in even “deeper” consumer knowledge. All of this has combined to create a new mode of advertising, which is transforming the entire advertising business.

For students in advertising programs across the country, they probably would not become “inventors” of these computational methods, but they can certainly become “creative users” of these methods. More importantly, advertising majors should be able to understand, think, and create future advertising in the spirit of computational advertising and to lead for meeting the new challenges as a result of the transformation of the advertising business.

Three Domains of Computational Advertising If we all like the concept “computational advertising,” we can have our version of computational advertising. It may consist of three domain:

1. Computational Creative 2. Computational Placement 3. Computational Evaluation

Computational Creative Computational creative explores the knowledge, applications and implementation in creating ads on the feedback of users. For example, Gatarski (2002) explores the use of an automated design system based on genetic algorithm to design banner advertising. Just as genes encode traits in living organism, so this algorithm propagates features of the most successful banner ads and extinguishes the least successful traits. Besides individual ads, ad groups may be tested and selected in Google AdWords. A set of lessons on the computational creative shall examine innovative applications in major platforms of digital media including social media, mobile phones and iPods that help the creation of effective advertising. Amazon’s dynamic ads is a more recent example of computational advertising. 

Computational Placement Computational placement deals with innovative applications of embedding ads in various kinds of digital content that takes advantage of information about the users, either in the real-time or in a delayed mode. The more informed of the users is a placement mechanism; the better it can serve the users. Contextual and behavioral placement, such as Microsoft Advertising Media Solutions, Yahoo Publishers Network, and Google AdSense, and other popular targeting mechanisms will be examined in the lessons on this theme. Potential side effects of computational placement such as privacy threats and data security will be reviewed, along the ongoing debates, in addition to the potential benefits to users of computational placement.

Computational Evaluation Computational evaluation focuses on the new metrics of measuring the effectiveness of advertising in digital media. For example, Google Analytics generates useful information about many aspects of a Web site, including the amount and sources of traffic, bounce rates, and navigation patterns. Research firms such as Comscore, Nielsen, and Dynamic Logic all have unique commercial applications that track and analyze user data. Besides, data mining, in and of itself, has become a domain of research and application and it should be introduced to our majors.

In sum, computational advertising is transforming many aspects of the advertising practice and theory. As advertising educators, we must embrace this new trend so that our students can have a bright future career.

[hairong]

Why Usablefindings.com?

Usablefindings.com was launched with the announcement of a contest for best summaries on February 14, 2014. Some friends asked me why we developed this knowledge sharing platform. It’s not easy to answer briefly because there was a long story behind this project. If you are curious, please read this blog.

Knowledge Gap Narrowing the knowledge gap has long been an issue in the fields of advertising and marketing. For example, there was a discussion on Elmar in 2005 on “Most Influential Theories or Most Useful Findings.” The discussion touched the issue of the practical relevance of academic research. It is obvious that some marketing academics are constantly seeking solutions for producing both scientifically rigorous and practically relevant research outcomes. The Marketing Science Institute and Journal of Marketing Research once announced a joint sponsorship for a research competition, a conference, and a special issue of JMR that would focus on rigorous research that addresses problems of importance to marketing managers. The call stated:

There is increasing concern about the possible divergence between the research conducted by academics and the problems faced by managers. One aspect of this divergence is that executives are voicing concerns about the relevance of academic research for important industry problems. This issue has many contributing factors, and it will be difficult, if not impossible, to resolve all the underlying concerns.

And the call also recognized the differences between practitioners and academics:

Whereas practitioners and academics share an interest in customers, competitors, and markets, practitioners are more concerned with actionable results, specific knowledge, proprietary advantage, and “how-to” knowledge (e.g., how to increase sales). In contrast, academics tend to focus on general knowledge, theory development, dissemination of results, and “why” knowledge (e.g., the process by which advertising affects sales). It is also well known that the incentive and reward systems for members of the two groups are disparate.

Therefore, at least two fundamental questions are facing academic researchers of advertising and marketing. Is managerial relevance of academic research really relevant for academics? If the answer is yes, how can academic research be more managerially relevant?

2006 AAA Panel To answer the above two questions, we organized a panel for the American Academy of Advertising annual conference in Reno in 2006, entitled “Bridging the Gap in Advertising Research between the Academia and the Practice: Challenges and Solutions.” Among the suggestions from the panelists were: work in teams of academic and practitioner researchers to do research that translates basic research into research that is useful to the practice of advertising; build databases and index them so that they are convenient to use by practitioners; make the relevance of our research a priority within the academy; and identify the most timely and relevant issues for younger academic researchers to pursue.

Trial Run To put some of the ideas from the 2006 AAA panel into action, the Department of Advertising, Public Relations, and Retailing at Michigan State University launched in 2007 a “Putting Research to Work” service, aimed at bridging the knowledge gap between academic researchers and leaders in our professional fields. We saw the service as adding value to the degrees of our alumni while providing a useful service to the advertising industry. Two student assistants browsed through the table of contents and abstracts of a few dozen academic journals for the initial selection of relevant articles, and two senior faculty members selected the final articles and wrote the briefs in a “lay” language that was easy for practitioners to digest.

The electronic service sent twice a month briefs of usable research findings to more than 1,000 executives within the nation’s top advertising agencies and public relations firms, as well as alumni of the department, and faculty members and graduate students in our college. The responses were positive and indicated to us a number of things. First, recipients wanted a way to discuss the studies. Second, respondents wanted to know how to get the actual articles. And third, they wanted more of these briefs than we were providing.

MSI Grant As user-generated content continued to gain favor, we planned to develop a social media platform for authors to contribute such briefs of their own studies. Due to the lack of financial support, however, we had put this idea on hold until we saw Marketing Science Institute’s call for “Ideas Challenge” in celebration of its 50th anniversary in 2011. We applied for and received a grant from the MSI to build this usablefindings.com platform.

Essence of the Platform As readers of the “Insights from MSI” for years, we certainly like these summaries of cutting-edge studies. We feel, however, the publication is geared more toward academic researchers and many entries tend to be long and not easy for practitioners to “grab and use.” Thus, a more interactive platform of usable findings can probably carry out the mission of the “Insights from MSI” in a more effective way.

We have considered using Facebook and even Google+ to create such a platform. An initial exploration convinced us that such generic sites lack the features we would like to see in a new platform, such as tagging briefs with any keywords, different layouts for easy browsing, and alerts by keywords, etc. We also looked at some content sharing sites and found that entries to these sites are largely difficult for readers to interact with, including searching, commenting, and emailing. However, many features from these established sites can be integrated in our new platform to make it a user-friendly and powerful site for researchers and practitioners in need of sharing research findings and ideas.

The content on the platform is generated and managed through the technique of crowdsourcing. This community-based approach to content generation allows a much larger amount of studies to be posted onto the platform. Maintaining a high level of quality content is the responsibility of the community and curators by rating studies and posting comments. An important factor in this platform’s success is its method of building user credibility and accountability, which are two factors we have taken into account for the design and user interface. Thus, login is required. The platform is integrated with popular social media sites, so users may login using their social media credentials to submit or rate summaries. Summaries that receive higher ratings are displayed more prominently and thus likely viewed by more visitors.

This platform functions as an archive of usable findings. Many usable findings of academic studies in our fields are published in brief via various outlets, such as news stories, newsletters, special reports and individual blogs. These outlets are highly valuable and, if their contents can be summarized in a single location and linked to the original text, more users can locate them and share, use and comment easily. As such, these studies should be more accessible and impactful.

As a platform of user-generated content, its future depends on user contribution and engagement. As our goal is to narrow the knowledge gap between academe and practice, we hope that scholars will post usable findings, practitioners will browse or search such findings, and both scholars and practitioners will comment and explore new research issues.

We encourage you to visit usablefindings.com to share, use and comment.

[hairong]

Using “Real” Data for Managerial Relevance

I received a grant from Google/ WPP in 2009 to explore the impact of online advertising through mining Chinese Internet users’ data. My collaborators and I collected brand and category search strings from one panel and overlaid them with tracking data from another massive panel on an individual basis using a unique identifier. This enabled us to obtain single-source measures of 14,500 users, including millions of records of ad exposures and search strings over a six-month period, on 28 campaigns for 10 brands in seven categories, in addition to the demographics of these anonymous users. It was a huge, multi-dimensional data set and challenged us in data transformation and analysis. In the course of this project, I often considered the issues of data mining in advertising research.

Gold Mine of Consumer Insights The rise of interactive media has fundamentally altered the way of advertising and marketing research for the past decade. As we know, vast amounts of user data are cumulated as a byproduct of normal consumption of interactive media and, with appropriate privacy protection, they are a gold mine for consumer insights. When users visit a Web site, search for a product online, post a note on Facebook, make an online purchase, or use a location-based service from a smartphone, they leave their digital footprints in these media. Such real-time data have greatly benefited the operations of numerous Web sites, the services of major research firms such as Nielsen and comScore, and the business of a new generation of data aggregators and warehouses.

The power of these tracking data has been manifested in the proliferation of precision marketing practices and many other renovations of the advertising industry. For example, advertising giant WPP recently launched a new service, Xaxis, which manages what it calls the “world’s largest” database of individuals’ profiles, including demographic, financial, purchase, geographic, and other information collected from their Web activities and brick-and-mortar transactions. Xaxis uses the database to personalize ads consumers see on the Web, socialnetworking sites, mobile phones, and ultimately, the TV set (Steel, 2011). On the education front, data mining, Web analytics, and even computational advertising courses are offered to computer science students at Stanford University and marketing students at Peking University in China, as far as I know. The Journal of Interactive Advertising, for which I am an editor, published two articles in its fall 2010 issue by Chinese researchers (Li, Pan, & Wang, 2010; Wang, Zhang, Li, & Zhu, 2010) using data mining methods. These examples suggest a pattern of increased use of real-time data on consumer behaviors in the practice, education, and research of our field.

Characteristics of Real-Time Data Compared with self-reported data from conventional surveys or experiments, real-time data have unique characteristics from a researcher’s perspective. First, they are behavioral. Clickstreams, search strings, postings, check-ins, online purchases, and friends and circles are all indicants of what users have actually done, from which users’ knowledge, interests, preferences, and even future behaviors may be inferred. Although past behaviors can be measured with self-reported questions, the accuracy of such measures obviously is not comparable with that of tracking data. Second, real-time data are often contextual. In addition to time and locale, tracked user behaviors usually take place in a specific setting, with precedent and consequent paths. For example, data of anonymous users for our Google/WPP project contained six months’ history of sites visited, ads exposed, searches by users, and URLs and titles of visited sites from search result pages, etc. Causal relationships can be explored from these sequential measures. Third, realtime data can be huge in size. Because of the interactive nature, media use and online behavior can be tracked continuously, generating endless streams of data. Depending on the needs of analysis, researchers may deal with large sizes of user samples, say thousands of users and millions of records, and thus increase the robustness of analyses and results. These characteristics of real-time data can help elevate academic research to an eminent height.

More of Managerial Relevance A major reason I see for using real-time data is to enhance the managerial relevancy of academic research. An issue facing advertising and marketing academics for years is that executives often voice concerns about the relevance of academic research to important industry problems (Ankers & Brennan, 2003; Lundberg, 2004). To address this issue, I moderated a panel at the 2006 AAA conference, which called for making the relevance of our research a priority within the academy and suggested several possible remedies, including “working in teams of academic and practitioner researchers to do research that translates basic research into research that is useful to the practice of advertising” (Li et al., 2006). Using real-time data in academic research can certainly help in this regard. For example, a finding of our Google/WPP project was that exposures to display ads increased both brand and category searches among users who were not in need of the studied categories. An executive who attended my presentation at the Google/WPP summit later tested and confirmed our finding in his firm’s own study, as I heard from his associates.

The idea of using real-time data for managerial relevance can be expressed best using Marla Royne’s words. She writes in her chapter for an advertising theory book to be published in 2012:

With the seeming demise of the “grand theory” across different disciplines, we must accumulate the important data collected and used by practitioners in an effort to come up with a better understanding of our field. While data-driven research is hardly the approach advanced and taught by scientists, it appears that much of “science” may be going that way, and, hence, this may be the best way for advertising academicians to contribute to the discipline. Partnering with practitioners and their data may help practitioners understand the consumer by using theory to explain marketplace happenings in a more systematic way. Helping the industry understand these occurrences through the use of theory to explain consumers and their behaviors has the potential for advertising academicians to advance the advertising discipline as a whole.

Access to Real-Time Data The proprietary nature of real-time data makes it difficult for academics to access; thus, we need to explore and solve the access issue. I have observed different ways real-time data are made available for academic researchers in recent years and some may have implications for the American Academy of Advertising. First, firms make their data public for analysis. For example, in October 2006, Netflix launched a three-year contest that offered a huge data set of 100 million movie ratings to researchers from all over the world to develop the best predication model for a reward of $1 million (Lohr, 2009). Second, university initiatives seek industry cooperation in making proprietary data available for analysis. An example is the Customer Analytics Initiative in the Wharton School at the University of Pennsylvania. It intends to be “the thought leader in data-driven customer-level analysis, applying these methods in a wide range of industries including interactive media, financial services, pharmaceuticals, telecom, nonprofits, and other areas where the use of detailed customer-level datasets is a key driver for business success” (Anonymous, 2011). Third, research firms make their proprietary data available to academics on a fee basis. An example is that comScore offers through Wharton Research Data Services (WRDS, 2011) its years of data on Web-wide visitation and transaction behavior based on a random sample from a cross-section of more than two million Internet users in the United States, including panelist-level browsing and buying behavior from tens of thousands of Internet users. Advertising researchers may take advantage of these sources of real-time data or find other innovative ways to access them for their projects.

In sum, mining real-time data of digital media and online behavior in various scenarios may lead to breakthrough advertising research that is more relevant to the practice of advertising. I share the vision of Micu et al. (2011) that the future marketing research will resemble “a river of information” from which continuous, organic flow of knowledge is “fished” to solve 80 percent of the real-world issues. Indeed, it is time for advertising academics to embrace this exciting growth in our profession.

References Ankers, Paul and Ross Brennan (2003), “Managerial Relevance in Academic Research: An Exploratory Study,” Marketing Intelligence and Planning, 21(1), 15–21.

Anonymous (2011), “About WCAI” (http://www.wharton.upenn.edu/wcai/about-us.cfm).

Li, Hairong, Helen Katz, Bruce Vander Bergh, and Richard T. Cole (2006), “Bridging the Gap in Advertising Research Between the Academia and The Practice: Challenges and Solutions,” in Richards, Jef I. (ed.), Proceedings of the 2006 Conference of the American Academy of Advertising, 75–79.

Li, Ji, Rui Pan, and Hansheng Wang, “Selection of Best Keywords: A Poisson Regression Model,” Journal of Interactive Advertising, 11(1), 27–35 (http://www.tandfonline.com/doi/full/10.1080/15252019.2010.10722175).

Lohr, Steve (2009), “A $1 Million Research Bargain for Netflix, and Maybe a Model for Others,” The New York Times, September 21, (http://t.co/VBDwCHu).

Lundberg, Craig C. (2004), “Is There Really Nothing So Practical as a Good Theory?” Business Horizons, 47(5), 7–14.

Micu, Anca Cristina, Kim Dedeker, Ian Lewis, Robert Moran, Oded Netzer, Joseph Plummer, and Joel Rubinson (2011), “The Shape of Marketing Research in 2021,” Journal of Advertising Research, 51(1), 213–221.

Royne, Marla B. (2012), “Toward Theories of Advertising: Where Do We Go from Here?” in Rodgers, Shelly and Esther Thorson (eds.), Advertising Theory, Routledge.

Steel, Emily (2011), “WPP Ad Unit Has Your Profile,” The Wall Street Journal, June 27 (http://t.co/DxDagpc).

Wang, Feng, Yin Zhang, Xiaoling Li, and Huawei Zhu (2010), “Why Do Moviegoers Go to the Theater? The Role of Prerelease Media Publicity and Online Word of Mouth in Driving Moviegoing Behavior,” Journal of Interactive Advertising, 11(1), 50–62 (http://www.tandfonline.com/doi/full/10.1080/15252019.2010.10722177).

WRDS (2011), “Welcome to WRDS.” (http://wrds-web.wharton. upenn.edu/wrds/).

(Originally published in the September 2011 issue of the American Academy of Advertising Newsletter)

[hairong]

Big Data for Advertising Research and Education

A Pre-Conference for the 2014 AAA Conference Atlanta, March 27, 2014 Registration at http://aaoa.wildapricot.org

Chairs Hairong Li, Michigan State University Michelle R. Nelson, University of Illinois at Urbana-Champaign Jisu Huh, University of Minnesota Harsha Gangadharbatla, University of Colorado at Boulder

This is a full-day pre-conference on March 27, 2014, consisting of four 90-minute sessions and covering the major areas of big data for advertising research and education. The first session will focus on the issues of data acquisition by exploring different methods, approaches and techniques in data tracking, collection and aggregation. The second session will explore the methodological implications of big data for advertising research through exemplifying empirical studies. The third session will introduce computational advertising as an innovative use of big data in advertising education. And the fourth and last session will be for vendors to demonstrate useful software programs, applications and tools for data analysis and visualization. Panelists from different disciplines and parts of the world should make the pre-conference an ideal forum for AAA conference attendees.

SESSION THEMES, CHAIRS, AND PANELISTS

8:00 – 8:30 am, Registration

The sessions will start after a brief introduction by Dr. Hairong Li, Professor of Advertising at Michigan State University.

8:30 – 10:00 am Session 1: Methods, Approaches and Techniques in Big Data Acquisition

The first step in using big data in advertising or marketing usually involves acquiring the said data. There are multiple ways in which advertising professionals, researchers, and academics can acquire and aggregate big data. This session focuses on two such data collection methods: (1) the use of syndicated sources and (2) acquisition through various tools, software, and programming languages. Topics and issues covered include data acquisition through third-party sources like ComScore, Google, and Acxiom; the advantages and disadvantages of working with syndicated data; issues associated with merging syndicated data and internal data; use of APIs in data acquisition; the various ways in which social media data can be collected and organized and its use in academic research; other data acquisition methods; and ways in which acquired data can be presented (data visualizations) and used in advertising, marketing, and non-profit environments.

Chair/Moderator/Discussant

Dr. Harsha Gangadharbatla, Associate Professor of Advertising, the University of Colorado at Boulder, email: harsha.gangadharbatla@colorado.edu. Dr. Gangadharbatla’s research interests lie at the intersection of technology, business, and communication. Prior to joining academia, he was a software programmer at Cognizant.

Panelists

Dr. Martin Paul Block is Professor in the Integrated Marketing Communications Division of the Medill School at Northwestern University. Dr. Block will address the use of syndicated data in advertising and marketing. He will focus on various available sources such as BIGinsight, ComScore, Google and Acxiom and talk about the role of metrics in advertising. He will discuss the various problems associated with syndicated data and the issues that arise when combining internal data with syndicated data.

Dr. Peter Neijens is Professor of Persuasive Communication in The Amsterdam School of Communication Research (ASCoR) at the Communication Science Department of the Universiteit van Amsterdam. Dr. Neijens is past president of the European Advertising Academy (EAA). His research interests include persuasive communication and media & advertising. His presentation addresses the collection and use of social media data for academic research.

Theo Araujo is PhD Candidate in The Amsterdam School of Communication Research (ASCoR). His research project investigates the impact of content, network and individual characteristics on the diffusion of messages created by brands on Social Network Sites, employing automated data extraction, content and network analysis. He holds a Research Master degree in Communication Science from the University of Amsterdam.

Hyeoncheol Lee is PhD Candidate at the department of computer and information sciences, Towson University. His research interests include data mining in social networks and big data analysis. Lee will talk about various data collection methods as they relate to advertising and marketing, particularly on social networking sites. He will explain the use of Application Programming Interface (API) in data collection on online social networks particularly Twitter and YouTube. He will also talk about other existing data collection tools and how they can be used in advertising.

Dr. Scott Hendrickson is Data Scientist at GNIP based in Boulder, CO. He holds a PhD in Physics from the University of Colorado, and has worked with various startups and established software companies on data analysis, machine learning, data visualization and data-centric strategy projects. He will present on the emerging field of data science with a particular focus on the various data acquisition methods available to marketers and non-profits. He will also talk about data visualization and its use in disasters.

10:00 – 10:30 am, Coffee Break

10:30 am – 12:00 noon Session 2: Methodological Implications of Big Data for Advertising Research

Advertising research has always been shaped and transformed by the emergence of new media and new technologies for measuring the audience and tracking their responses to advertising messages. Now, we are experiencing another round of sea change in the way the media audience and advertising effects are measured and studied. This panel session will present cutting-edge trends in research in advertising and related fields utilizing big data and new analytical tools, and discuss methodological implications of big data for advertising research.

While big data seems to offer opportunities for improving the execution and evaluation of advertising and may provide valuable exploratory research insights, there are many issues and questions. Is big data really better than medium or small data? How will big data change the process of doing advertising research? How will advertising research and practice benefit from big data? To address these and other questions, four panel members and the audience will share and discuss various interdisciplinary approaches to big data analysis and the implications of big data for methods and measurements used in advertising and relevant research areas. The panelists include experts from the advertising, marketing, mass communication, and computer science fields, with extensive research on the topic of big data and web analytics.

Chair/Moderator/Discussant

Dr. Jisu Huh, Associate Professor, Director of Graduate Studies, School of Journalism and Mass Communication, University of Minnesota, Affiliated Faculty Member of the University of Minnesota’s Social Media and Business Analytics Collaborative (SOBACO), Email: jhuh@umn.edu.

Panelists

Dr. Shintaro Okazaki, Associate Professor of Marketing, Universidad Autónoma de Madrid, Spain. Dr. Okazaki earned his PhD in Marketing from the Universidad Autónoma de Madrid. His research focuses on interactive and international marketing. He serves on the editorial boards of Journal of Advertising Research, International Journal of Advertising, Journal of Interactive Advertising, Journal of Electronic Commerce Research, and Internet Research, among others. He has been appointed as incoming Editor of the Journal of Advertising. His presentation will focus on a methodological framework for using opinion mining to analyze comments on Twitters. A series of procedural recommendations is described for researchers who are unfamiliar with this research area. The major steps include brand selection, determination of a classification scheme and categories, human coding, programming of the automated classification algorithm, and evaluation of the classification results. He then will present the results of a pretest that examined the content of Tweets about IKEA. He performed a social network analysis to identify interrelationships among the Tweets. In closing, methodological implications and utility for advertising research will be discussed.

Dr. Edward C. Malthouse, Theodore R and Annie Laurie Sills Professor of Integrated Marketing Communications and Industrial Engineering and Research Director for the Spiegel Institute on Digital and Database Marketing, Medill School, Northwestern University. Dr. Malthouse earned his PhD in Computational Statistics from Northwestern University and completed a post doc at the Kellogg Marketing Department. He was the co-editor of the Journal of Interactive Marketing between 2005-2011. His research interests center on media marketing, database marketing, advertising, new media, and integrated marketing communications. He teaches undergraduates, graduates and executives and has been a visiting professor at universities in Japan, China and Europe. He will discuss what is meant by “big data,” the sampling properties of big data and its appropriate uses in advertising research, with some relevant examples. He will also contrast the uses of big data with those medium-sized customer databases with millions or tens of millions of cases, and small-sized data of hundreds or thousands of cases.

Dr. Itai Himelboim, Assistant Professor in the Department of Telecommunications at the Grady College of the University of Georgia. Dr. Himelboim earned his PhD in Mass Communication from the University of Minnesota. He studies social media networks and their implications for democratic and civil societies. He pursues this interest by examining patterns of interactions and identifying key actors and communities in Twitter, YouTube, discussion forums and other social spaces. He will present his research on Applying Network Analysis to Examining Twitter Data, discuss various issues and challenges in content analysis of social media, and provide some suggestions.

Dr. Wenyu Dou, Professor and Associate Dean of the School of Business, City University of Hong Kong. Social media in China has been witnessing explosive growth during the past few years and its number of users has exceeded 500 million in 2013. Millions of data records about user interests, interactions with brands, even purchases, are generated each day. For companies and brands, how to harness the power of “big data” and deliver targeted messages has become a major challenge. Professor Dou, an expert on social media marketing in China, will introduce and discuss a number of key analytics tools (e.g., Kong Ming, Doodod), which can be effectively applied to understand the interests and behaviors of social media users in China. Specifically, he will focus on analytics for Weibo, China’s counterpart to Twitter, with 400 million users. He will conclude with recommendations for advertising researchers on possible avenues to conduct studies using data from these Weibo analytics tools.

12:00 noon – 1:00 pm, Lunch

1:00 – 2:30 pm Session 3: Computational Advertising and Big Data Implications for Education

Computational advertising… “seeks to put the best ad in the best context before the right customer” (Essex 2009, p.16). This definition sounds like advertising strategy. The computational part of this ‘new discipline’ is the new part. Using data mining, machine learning, computational techniques, microeconomics, and psychology, computational advertising can help media producers, media sellers, and businesses focus on using data and information for crafting new models of content delivery that can actually benefit the consumer. This new paradigm requires both the mechanics and science of data manipulation (Computer Science) as well as the art, sociology and psychology of understanding messages and people (Advertising).

Yet, in a world where 2.5 quintillion bytes of data are created daily – where billions of individual ads, search queries, page impressions, photo uploads and purchases are analyzed by Google, Facebook, and Amazon to identify, understand, and predict social relationships and patterns - there is virtually no related course content or academic research within the field of advertising.

This panel session will discuss the definition and future of computational advertising as it relates to advertising education. What knowledge and skills are desired in industry? How much “computational” is required to develop, implement and understand computational advertising? [How] can computer science and advertising come together in cross-disciplinary education? The panel is comprised of experts in computer science who have already taught related cross-disciplinary course, from an expert in the media industry who works with ‘big data’ and regularly hires advertising students, and from young industry professionals with a combination of computer science and advertising degrees and skills.

Chair/Moderator/Discussant

Dr. Michelle R. Nelson, Associate Professor, Charles H. Sandage Department of Advertising (ADV), University of Illinois at Urbana-Champaign, email: nelsonmr@illinois.edu. Dr. Nelson co-wrote a proposal with the head of Computer Science (CS) at UIUC to bring a “computational advertising” professor to campus. CS and ADV are working together on a joint CS+Advertising undergraduate degree.

Panelists

Dr. Ram Akella, Professor of Information Systems and Technology Management, Director, Center for Knowledge, Information Systems and Management of Technology, University of California Santa Cruz. Dr. Akella was co-instructor for what is believed to be the first course on computational advertising: (ISM293 Introduction to Computational Advertising, University of California, Santa Cruz) in Spring 2009. Dr. Akella will share his experience teaching the course and his ideas for the future of computational advertising from the perspective of computer science and industry.

Dr. Hari Sundaram, Associate Professor, School of Arts, Media and Engineering + Computer Science, Arizona State University. Dr. Sundaram helped form the nation’s first digital culture B.S. at Arizona State University. He has regularly taught courses for computer scientists and non-computer scientists related to social media, data structures for digital culture and social and economic networks. His research combines computer science and persuasion to find ways to enable cooperative behavior adoption on a large scale (e.g., sustainability, healthy lifestyles).

Dr. Helen Katz, SVP Research Director, Starcom MediaVest Group. Dr. Katz works with digital analytics, social media monitoring, return path data, and addressable advertising and other forms of ‘big data’ in her role as Senior Vice President of Research for the Starcom MediaVest Group. She’ll talk about the ways that industry use “big data” and the knowledge and skills that students should have to be successful.

Yuri Kleban, Google Advertising Operations. Mr. Kleban graduated from UIUC in 2011 with an Advertising major and Communication minor. While in college, he started his own YouTube show, worked with a fellow Illini on Phlint (web & app development startup), and co-created a polling app on Facebook that attracted 500,000 users (his colleagues went on to create striking.ly). Upon graduation, Yuri was hired to work for Google in Global Advertising Operations. While at Google, Yuri began learning programming (JS and Python) and has created an Internal Tools team. He is currently the Tech Lead of the Global Customer Services team.

2:30 – 3:00 pm, Coffee Break

3:00 – 4:30 pm Session 4: Vendor Demos of Software Programs, Applications and Tools

This session will consist of four company representatives and they will demonstrate the functions and use of major programs, applications and tools for data analysis and visualization. These companies have agreed to send the representatives to the pre-conference although the names of these representatives are still to be decided as time is near.

Chair/Moderator/Discussant

Dr. Harsha Gangadharbatla, Associate Professor of Advertising, the University of Colorado at Boulder, email: harsha.gangadharbatla@colorado.edu. Dr. Gangadharbatla’s research interests lie at the intersection of technology, business, and communication. Prior to joining academia, he was a software programmer at Cognizant.

Panelists/Vendors

Four companies—Loudpixel, IBM, SiteMinis, and HootSuite—will do a demonstration of the various tools they offer for collection and analysis of data in advertising.

Allie Siarto, the co-founder of Loudpixel, a social media monitoring and analysis company with a focus on understanding social media context, sentiment and trends in order to drive business and communication decisions. Allie has launched monitoring programs and executed social research across a variety of industries, including restaurants, finance, auto, technology and consumer goods. She teaches a class on monitoring and measuring social media at Michigan State University, and her work has been published in Forbes, PC World, Washington Post, Yahoo!, Small Business Advisor, MSN, VentureBeat and Under 30 CEO. Loudpixel helps companies analyze hundreds of thousands of social media conversations in order to understand who their customers are and what they care about. Allie will discuss how social media research can be used to compliment other research and consumer insights initiatives.

Marci Troutman, Founder and CEO of SiteMinis, a mobile technology company in Atlanta, will review the trends in mobile analytics, introduce the tools for collecting, analyzing and curating data from mobile devices, and discuss the managerial implications of mobile analytics.

Marva Bailer is the IBM Business Unit Executive for North America Smarter Planet Industry Solutions. Her team utilizes all of the software offerings that IBM offers in an industry context. Key business areas are Social, Analytics , Mobile, Fraud, Audience Sentiment, Optimization, Service Management, Software Defined Networks, Facilities Management, Customer Service, Demand Marketing, and Commerce. Smart Grid, Media/ Content Delivery, OSS/ BSS, are key focus areas.

Paula Cusati, Program manager responsible for all aspects of HootSuite’s Higher Education program including strategic planning for scale, international expansion, marketing campaigns, account management, content development and reporting. HootSuite’s Higher Education program offers professors and their students access to tools and resources that support social media education.

4:30 – 5:00 pm Q&A, Summary and Conclusion