Brand+TV+iPad: the new research triangle

Editor's note: Charles Young is founder and CEO of Ameritest, an Albuquerque, N.M., research firm.

What if, instead of looking through a one-way mirror of a focus group facility, the advertising manager of a major brand could look through the flat-screen TV hanging on the consumer’s living-room wall and watch her in her real-world setting? These days, he would most likely see her watching the moving images on the big screen while intermittently looking down at the tablet computer in her lap, before looking back up at the big screen.

If the ad manager were then to send a text message to the consumer asking her what she was doing on her tablet, she might say she was doing something unrelated to the TV content, like reading the news, because what was on television was not engaging at that moment. She might reply that she was sending a text message to a friend on Facebook about the program that they both were watching at that time. She might say she was looking on IMDb to see what other movies the actor she saw in the program had been in. She might say that she was interacting with the program by voting for her favorite performer on a reality show. Or, she might even report that she had just bought something online that she had seen advertised on the TV.

Each of these answers not only has implications for the job of the advertising manager but also serves to remind us how the world of television advertising has changed.

In the new world of multiplatform marketing, there are many devices – TVs, computers, smartphones, tablets, singly or in combination – that can be used to connect with consumers in a wide variety of settings. However, at the moment, for major advertisers, by far the most important in terms of advertising dollars at risk is the situation just described: the consumer sitting in the living room watching a big screen TV on the wall with a tablet on her lap.

The arrival of an iPad that competes for attention with the TV set in the living room is not just a problem of distraction, with attention given to one device coming at the expense of the other. If this were so, it would be a simple problem for researchers to quantify by observing consumer behavior – for example, researchers might simply use eye-tracking cameras to measure how much time is spent looking at one device versus another.

But for advertisers who seek the power of multiplatform marketing, this novel situation represents an opportunity, in the form of a new research triangle: TV + tablet + brand.

Like Euler’s three-body problem in physics, understanding potential interactions between the new triangle of TV + tablet + brand is complex. It calls for research that moves beyond behaviorism to understand the psychology of these three things interacting in a real-world setting.

Produced interesting insights

A number of interesting studies of the new research triangle have been done in a laboratory setting. Turner Networks, for example, has conducted research in its new media lab in New York on the contextual effects of multiplatform audience engagement. Turner’s research has produced interesting insights, such as demonstrating the importance of audio in cueing consumers busily engaged with their tablets to shift their attention to something interesting that is about to happen on the TV screen.

However, given the high stakes, it is important for the advertising industry to quickly validate such learning and hopefully produce new insights by proving results in real-world settings. To that end, researchers can strive to design studies that meet the following three conditions:

  1. Research that is collected in the place where the triangle actually exists; that is, in the consumer’s living room.
  2. Research collected at the time the consumer is actually engaged with media (i.e., in real time or as close to it as possible).
  3. Research collected non-intrusively on the devices the consumer is actually using – TV set or tablet computer (or both).

Fortunately, the technology for researching the triangle is rapidly evolving. And while smart TVs equipped with cameras that might be used for eye-tracking or facial response studies have not yet reached penetration levels that are practical for these kinds of studies, the tablet – in most cases, the iPad – has already reached household penetration levels projected to surpass 50 percent in the next year, so that even now we can begin to conduct scientifically projectable studies of the multi-platform triangle.

Which group for the iPad?

Now imagine you are a respondent in a focus group and you are asked to sort different computing devices into groups based on their similarities. Into which group would an iPad fall? Would it fall into a group containing laptop and desktop computers, because of its computing power and ability to perform complex tasks on its larger screen? Or into a group containing smartphones, because of its mobility?

Our answer is that the iPad falls into its own distinct third category, at least for our purposes in studying multi-platform brand-building in the living room. We base this conclusion on research we conducted when we began the process of adapting our advertising pre-test methods for use on the iPad.

Interestingly, although it’s a mobile platform, the iPad is significantly more likely to be used at home for answering researchers’ questions than is a desktop or laptop. Notably, among consumers recruited from survey panel companies, nearly one in four computer-based interviews are actually taken away from home, usually at work.

Moreover, on the iPad, the consumer is three times more likely to be simultaneously watching television while taking the survey. While for some research purposes, having an undistracted consumer sitting in front of a computer while answering survey questions may be desirable, for studying the context effects of the multi-platform triangle, the iPad is superior to a computer (see Figure 1).

Of course, the consumer may be just as likely to be sitting on the couch in front of the television with a smartphone as with an iPad. However, given the current size of most smartphones, the screen is too small for anything but the most simplistic kind of ad interview.

For this reason, the iPad turns out to be a better device than a phone for conducting ad research. The screen is certainly large enough for watching an ad in high definition. In fact, these days, many younger consumers are increasingly watching television content on an iPad instead of on a TV set. Perhaps more importantly, the finger-touch interaction with the larger screen is ideal for deploying the visual diagnostic questions for which our company is known.

Having been in the ad testing business for many years, we had a number of research-on-research questions to answer before we could migrate our methods to an iPad. Would the changes in device, in setting and in context – while the TV might be on; while other people might also be in the living room – produce norms different from those we built up over time with our conventional online methodology? Given other distractions in the room, would consumers be more or less likely to complete our interviews? How might our moment-by-moment picture-sort graphs of attention, emotion and meaning be affected by shifting the picture-sort task to a touch-screen device? And perhaps most importantly, how could the user experience of our respondents be improved in order to improve the overall quality of the data we were collecting?

All of the issues were answered

After conducting experiments to compare iPad methods to conventional methods, all of the issues were answered to our satisfaction. Norms are indeed comparable, well within a small range of statistical tolerance. Completion rates are, if anything, higher. And redesigning our picture-sort interview to fit the finger drag-and-drop capabilities of an iPad touch screen leads to a more game-like experience during the picture-sort questions – much like playing a game of solitaire or Scrabble on the iPad.

And while the diagnostic Picture Sort Flow this data produces is surprisingly unaffected – virtually no differences at all – the user experience is radically improved. Indeed, because the experience is rated more game-like, an iPad interview is rated as a “very positive” experience 59 percent of the time, versus 43 percent for a conventional interview. Consumers prefer an iPad interview to a conventional interview 71 percent to 29 percent.

Finally, regarding the distractions that are present in an iPad-based interview – TV program on in the background, the social context of other people in the living room and even, as we shall see shortly, the emotional state of mind of the consumer while taking the interview – we realize that the iPad now creates two choices for designing an ad study.

First, it is still possible to conduct conventional ad pre-testing or tracking studies using an iPad. Researchers can let these variables fall where they may and set them aside in analyzing the same kind of data that is produced by computer-based interviews. Remember, conventional norms continue to be comparable.

But now there is also a second option. And that is to more carefully control the timing of when data is collected to better understand how contextual variables affect consumers’ responses to brand-building advertising embedded in real-world programming.

Important new factor

Speed is an important new factor in advertising research. The days when advertisers and their agencies would have to wait three or four weeks for copy-testing results are fading fast. The speed of ad research continues to accelerate: A five-day turnaround time for a full report is increasingly becoming the new normal.

The impetus for such speed is to drive the tempo of creative team decision-making. Our original insight was that if we could minimize the turnaround time for research, we could reduce the long-established barriers between creatives and researchers and thus reposition pre-testing in a more positive light. Instead of being a drag on the creative process it could be used as a simple feedback loop that drives an iterative process of creative optimization.

A hidden benefit of this automation-driven pre-testing process is that we realized that if we could concentrate data collection within a single day, as opposed to the conventional practice of spreading interviewing across a week or so, we could begin to provide insights into media placement. And, in particular, we might begin the journey of exploring the new research triangle.

Coordinating the sample and timing of research data collection with the media plan for the debut of a new television commercial can be a tricky problem from a logistical standpoint. But last year’s Super Bowl gave us an excellent opportunity for testing out our idea.

Last year, a number of advertisers released their ads to the Internet about a week before the big game in order to prime social media for generating buzz. This made it possible to identify an ad, in advance, that we knew would be airing during the game. For our triangulation experiment, we selected a humorous Taco Bell commercial called “Viva Mas.”

In order to study the contextual effects of the Super Bowl, we used a simple test-and-control design. We pre-recruited two groups of 100 consumers to take a standard Ameritest interview on an iPad at home. The first group of consumers took the interview on the Saturday before the game – this was our control cell. Interviewing for the test group was timed to start on Sunday immediately after the commercial had aired during the game. The group in this test cell had the option of taking the interview either while they were watching TV or immediately after the game. All of the interviews for both groups were completed on the same day for which they were scheduled.

What we learned from this experiment is that given the context of the Super Bowl, the Attention Score for the Taco Bell ad was significantly higher – but the Motivation Score was lower.

This result should not be so surprising. Everyone is hyper-aware of the advertising during the Super Bowl – hence the heightened level of attention or engagement. That’s the reason advertisers pay big bucks to air their commercials during the game.

But the Super Bowl viewers have become self-anointed ad critics who analyze the ads so they can talk about them later. This stiffer competition from other ads in the game leads to a heightened awareness of weaknesses in the ads. Our picture-sort diagnostics for “Viva Mas” showed a strong level of unintended negative emotion towards a middle section of this ad showing old people dancing in a disco like teenagers – and it was easy to show in analysis that this was suppressing motivation.

To validate our findings, we turned to Tucson, Ariz., research firm Communicus, whose approach to tracking uses a unique longitudinal sampling methodology to isolate the effects of different variables in the marketing mix. Communicus’ conclusions regarding the effectiveness of the Taco Bell Super Bowl airing were that “Attention was strong, branding was average and motivation was weak – producing a relatively weak overall ranking.” Thus, an independent assessment of market performance confirmed our report-card findings about the contextual effects of the Super Bowl on the Taco Bell advertising.

Wide range of emotions

One of the interesting things about the Super Bowl is that it produces a wide range of emotions in its audience. Because different members of the audience are rooting for different teams there may be both positive and negative emotional reactions to the same moments in the game.

In this respect, sports may be different from a lot of television programming wherein the storyline of a show is intended to synchronize or focus the emotions of the audience. A comedy is expected to generate a certain range of emotions that is different from what a drama or a romance or an action movie is expected to evoke. Consequently, one of the creative aspects of a media programmer’s job is to match a brand’s advertising to the appropriate programming content so that congruent emotions generated by both program and ad can amplify the effectiveness of the advertising in terms of motivating consumers to buy the brand. We have published research with Turner Networks to show that this emotional congruence effect on advertising motivation is quite real.

Of course, the range of emotions that consumers are likely to experience throughout the day as they are exposed to advertising is likely to be much wider than those the consumer is feeling when they are in the mood to sit in front of a computer to take a conventional research survey.

Media Behavior Insights (MBI) is a New York research company start-up that has recognized this and developed a new product to address what it feels is an important information void for mobile media programmers. Under MBI's approach, emotions that consumers are feeling and the media they’ve been exposed to are measured every 15 minutes. For our Super Bowl experiment, we borrowed (with MBI's permission) the emotional battery it uses to track consumer emotions throughout the day.

The emotional battery covers a wide range of states, such as confident, excited and hopeful on the positive end to angry, frustrated and worried on the negative. For our analysis, we combined these into a Positive Feelings Net and a Negative Feelings Net in order to study how the emotional context of the emotions generated by the Super Bowl game impacted the response to the Taco Bell ad.

The impact of the emotional context associated with the game on consumer response to the ad was significant. Consumers who were feeling positively while watching the game were more likely to be motivated by the ad and rated the ad higher on a variety of related diagnostics.

Importantly, consumers in a positive mood were also much more positive toward how well they thought the ad fit the brand and toward the particular brand perceptions being generated (Figure 2).

Consistent with this, we see that the happy viewers of the game were feeling much more positively toward the ad itself. Our Flow of Emotion showed that the emotional response to the ad from beginning to end was much more positive, 55 percent positive versus 42 percent (Figure 3).

One reason for this, which we could see in our Flow of Meaning (Figure 4), is that those in a positive frame of mind were much more likely to find the ad to be funny than were those in a negative frame of mind.

To summarize, looking across our research we can see that we have one more solid example showing the impact of the emotional context generated by media on the effectiveness of an ad.

Early days

These are, of course, early days for understanding the hidden power that resides in the triangle of TV + tablet + brand. Looking out from one corner of the triangle, we can now see that researchers have a new ability to measure advertising at the real-world point of media consumption. That is, in the living room.