Native Advertising on TV: Effects of Ad Format and Media Context

Native advertising has become popular on websites and social media platforms in recent years. Now the TV industry is starting to develop its own native ads/sponsored content. The current research examined this newer form of native advertising – native video ads on TV. Using advertising format (video vs. text/image) as a within-group variable and media context (news vs. entertainment) as a between-group variable, a repeated measures was run to explore the effects on viewers’ attitude-toward-the-ad, ad trust, brand interest, as well as perceived differences between native ads and traditional advertising formats. In general, participants reported a more favorable attitude toward video native advertising than text/image advertising. Native ads embedded in news content were perceived as more trustworthy than those embedded in entertainment content. Advertising format also seems to affect how participants perceived the differences between native advertising and traditional advertising. if the TV show and ad are similar in emotional content. on brand placement more brand-related and


INTRODUCTION
Since the beginning of broadcast and cable television, these industries have relied on advertising to stay in business and to earn a profit. For decades the most popular format for television ads has been 30 and 60-second ads. These ad formats have brought TV networks millions of dollars in revenue year after year. However, as with other media, the Internet has disrupted the TV industry's business model and is forcing the industry to invest in new ways in which to provide advertisers with an effective experience that will help them sale their products and services.
One new advertising format that TV is testing is native advertising. This ad type was originally used on the Internet. On the web, native ads try to match the design and content of the host webpage (Hill, 2013). Many researchers (e.g., Becker-Olsen, 2003;Tutaj & Van Reijmersdal, 2012) argued that this ad format was more informative and less irritating, and therefore more effective. It is predicted that, by 2021, native display ad revenue will make up 74% of total US display ad revenue, up from 56% in 2016 (Meola, 2016). Industry experts believe native ads would work well on TV, especially for news and entertainment programming (e.g. Wolk, 2015). The ads can be inserted within a TV program via character dialogue or on-screen graphics, or featured outside the show as a short video presentation that discusses the product. With this being a new trend among TV broadcasters, there has been very little research conducted on television native ads and how audiences react to this type of advertising.
The current study is one of the first to test the effectiveness of television native ads. Video native ads on TV are compared to text/image native ads on websites to explore which ad format will generate higher ad trust, more favorable attitude-toward-the-ad, and stronger brand interest. In addition, this study seeks to provide data on whether ad format and media context (news/entertainment) influence the participant's perception of native ads in comparison with traditional advertising.

Native Advertising on TV
Advertising has always been the central revenue creator for TV networks. For decades TV stations have adopted the 30 and 60-second advertisement format. In the digital age, TV has seen new competition from the Internet, which offers more non-traditional options to help advertisers reach their potential customer base. One of the newest trends in digital advertising is native advertising.
Native advertising, also referred to as sponsored content, is a term that is used to describe paid advertisements that are produced to have the look of editorial content from the publisher itself (Wojdynski & Evans, 2016). Native ads evolved from the convergence of advertorial concepts and social media. During the rise of social media, these sites provided advertisers with the ability to create large followings of potential customers that they could promote their products to for very minimal cost. Brands have discovered that traditional ads that outright promoted a product did not receive as much feedback as native advertising messages that were more subtle and more playful and conversational (Campbell & Marks, 2015).
One website that is the leader in this type of advertising is Buzzfeed. The website is popular among Millennials as it offers lists of pop culture references, viral videos and top headlines of the day. It also has a large amount of native advertising. An example of native advertising that you might see on Buzzfeed would be "15 Different Ways to Eat Oreo Cookies." This article would contain on the byline "Sponsored Content from Oreo." While it looks like a legitimate article, it is actually an advertisement for Oreo cookies. Recently, it's not just Buzzfeed that is producing native ads. Well-known and established news brands are also beginning to invest in producing sponsored content for their digital assets, including The New York Times, Wall Street Journal, and The Huffington Post (Moses, 2014).
While Internet native advertising continues to grow, the TV industry is starting to develop its own native ads/sponsored content. Advertisers are provided with new options beyond the traditional 30 and 60-second ad format to reach their customers on television networks. Smit, Reijmersdal, and Neijens (2009) described it as in-program sponsoring, product placement or brand placement, which is the presence of products within TV content. Previous case studies have shown that consumers have a higher brand recall with these ad formats than with traditional advertisement formats (Reinares & Reinares, 2013). One of the first TV shows to integrate native ads is the 2016 TBS police satire comedy Angie Tribeca. TBS decided to try and copy the Netflix business model with this show's first season premiere by offering all episodes of the first season in a 25-hour commercial free binge-a-thon on the cable channel. In reality, the "commercial free" premiere actually contained commercials, just not traditional ads. While in between episodes, the channel offered two minute features of content from Dunkin' Donuts, TurboTax and Redd's Apple Ale. TBS also offered advertisements within episodes as well. In the pilot episode, each time a vehicle was shown, the Ford logo appeared on screen. Another episode offered two characters discussing the joys of eating a Snickers bar, while one character holds up a Snickers wrapper so the viewer can clearly see the candy bar logo. TBS isn't the only Turner network that plans to cut down on the amount of traditional ads. Turner Broadcasting announced their plans to decrease the amount of traditional ads shown on TruTV as well as TNT and instead offer more programming (and more native ads). TNT plans to cut traditional ads by 50% in new dramas. Other networks have also tried this new type of ad format as well. FOX offered a three-episode story line in their popular show "Empire" that incorporated Pepsi. Procter & Gamble purchased all the ad inventory for a set of Friday night movies on the Lifetime network (Steinberg, 2016). NBC News also began using native ads on their websites including Today.com, CNBC.com, and potentially NBCNews.com (Taintor, 2013).

Advertising Format and Vividness Effects
When it comes to native advertising, one prominent difference between online ads and TV ads is ad format, because many online native ads contain mostly text and images. It was only recently that we started seeing online native ads in video format. One construct that has been frequently used by scholars to explain differences between video and text messages is the level of vividness. According to Steuer (1992), vividness is "the representational richness of a mediated environment as defined by its formal features; that is, the way in which an environment presents information to the senses (p. 11)." It has two dimensions -breadth and depth. Breadth refers to the number of sensory dimensions presented (such as colors, graphics, etc.), and depth is the quality and resolution of the presentation. In this sense, audio, motion images and videos are highly vivid content and written text is low in vividness.
Vivid information stimulates multiple perceptual systems and therefore should be better perceived than information that only appeals on single perceptual systems (Li, Daugherty, & Biocca, 2002). For example, Appiah (2006) used a sample of 296 college students (ages 18-40) to review audio/video and text/picture testimonial ads on websites. After viewing the ads, participants were asked to report their attitudes towards the ads and the advertised product. The results showed that the participants felt more strongly towards the audio/video testimonial ads than the text/picture ads. They also rated websites that contained audio/video testimonial ads more favorably than those containing text/picture ads or no ads at all. Spalding, Cole, and Fayer (2009) also reported that advertising campaigns that used rich media advertisements, especially in video format, had much stronger effects on viewers, compared to campaigns that only used pictures, GIFs or simple Flash layouts. Flores, Chen, Ross (2014) examined the effect of banner ads on Internet users' attitude, in terms of the type of ad (display ad with photographs or text-only ad), the shape (horizontal or vertical), and the language of the ad (English vs. Spanish). They found that the ad for a high-involvement product was viewed as more appealing when advertised with a display ad instead of text-only.
Even though specific studies on the vividness of native ads could not be located, it is expected that it should demonstrate a similar pattern as discovered in previous literature. Therefore, the following hypothesis was proposed: H1: Video native advertising will generate (a) more favorable attitude-toward-the-ad, (b) higher ad trust, and (c) higher brand interest, compared to text/image native advertising.

Influence of Media Context
Native advertising is often criticized as too deceptive, especially when used within news content (Colhoun, 2015;Sonderman & Tran, 2013). This was a concern the Federal Trade Commission (FTC) had and it held a workshop in 2013 that dealt with consumers recognition and understanding of sponsored content (FTC, 2013b). Others are concerned as well that native ads are crossing the church-state divide, which is an attempt to keep editorial and commercial interest separate in newsrooms, as sponsored content looks more and more like editorial content (Benton, 2014;Carlson, 2015). For example, in 2013, The Atlantic published an article on the Church of Scientology that was paid for by the organization itself. The only signal that it was not editorial content was a small yellow banner at the bottom of the page that stated it was sponsored content. There was a quick backlash against the article online and the website took down the story in a matter of 12 hours after it was posted (Wemple, 2013). A study by Contently reported that 59% of respondents view news websites with sponsored content as less credible and 67% felt deceived after they found an article was sponsored by a brand (Lazauskas, 2014). Entertainment based content, however, does not have these concerns, as news based content is considered to be for more informational purposes than entertainment content that is simply for entertainment purposes. Therefore, the current study examined how media context, news or entertainment, might affect perceptions of native ads.
Pelsmacker, Geuens, and Anckaert (2002) defines media context as "the characteristics of the content of the medium in which an ad is inserted (e.g., articles in a magazine, spots in a television program), as they are perceived by the persons who are exposed to them (p. 49)." Research has consistently suggested that media context could affect advertising effectiveness. For example, Coulter (1998) examined the effects of emotional responses to television shows on the participant's attitudes towards the ads. Two 30-second automobile ads from a fictitious company were used, one positive in tone and the other negative, during a commercial break for two shows (M*A*S*H and Highway to Heaven). The results showed that program liking facilitates the effect of program-induced affect on ad evaluations; and the liking of the ad is greater if the TV show and ad are similar in emotional content. A study on brand placement by Van Reijmersdal, Smit, and Neijens (2010) found that brand placement in TV programs with a higher perceived informational value generated higher brand recognition and more brand-related behaviors. Brand recognition and behaviors were also affected by program genres (special interest program vs. general interest programs). Pelsmacker, Geuens, and Anckaert (2002) studied viewers' response to TV and magazine ads and stated that ads presented in a highly appreciated context resulted in a more positive attitude toward the ad. Involvement moderated the effects of media context. Low-involvement viewers perceived ads embedded in a congruent context as clearer and more likable, while high-involvement viewers reported more favorable attitudes toward ads embedded in a contrasting context.
The role of media context has not been tested in a native advertising environment. Therefore, the following research question was developed: RQ1: Does media context influence the effects of video and text/image native advertising? In addition, the researchers in the current study were also interested to see how ad format and media context influence the comparison between native advertising and traditional advertising.
RQ2: How do audiences compare video and text/image native advertising to traditional advertising?

Method
A 2 (advertising format: video vs. text/image) × 2 (media context: news vs. entertainment) mixed between-and within-factorial design was utilized to explore the effects of native advertising format (within-group factor) and media context (between-group factor) on audiences' attitude-toward-the-ad, ad trust, and brand interest.

Research Stimuli
A pretest with 12 college students was conducted to help identify stimulus ads and develop the questionnaire. Four previously published native ads were selected for the current study -CNN Heroes (Subaru), The New York Times (Ford), Saturday Night Live (Pepsi), and BuzzFeed (Pepsi). Real ads that aired on TV, rather than researcher-manipulated ads, were used because the purpose of the study was to explore viewers' responses to different native ad formats, not specific design elements. The video ads that were streamed for the participants were the exact same ads that aired on the TV networks. In addition, even with researcher-designed ads, it would be difficult to manipulate differences between video and text/image ads. The researchers did try to make the ads relatively comparable by focusing on two product categories -automobiles (Subaru vs. Ford) and soft drink (Pepsi).
CNN Heroes is a feature that highlights those that are making a difference in their community. The CNN Heroes video ad (https://www.ispot.tv/ad/AoYM/subaru-a-cnn-heroessuccess-story) runs for a total of 120-seconds and tells the story of a woman that sought help for her substance abuse at the Phoenix Multisport gym, which is a place that provides help for those suffering from abuse problems and gets them involved in sports to make a positive change in their lives. Subaru partnered with the organization to offer financing for the gym.
The New York Times display ad (https://paidpost.nytimes.com/ford/sowing-the-seeds.html) is clearly marked at the top of the page with "Paid for and posted by Ford" to let readers know this is a piece of paid content. The ad contains text, interactive graphics, and large photos to show how the company is trying to make their vehicles greener by using less plastic. There is also a video to reinforce the material covered in the text and graphics.
The Saturday Night Live (SNL) sixty-second video ad aired during an episode of SNL (http://www.fanpop.com/clubs/richard-dean-anderson/videos/4237358/title/saturday-night-livemacgruber-pepsi). The ad contained cast members from the show in character in a MacGyver like skit with the main character, MacGruber, trying to diffuse a bomb, while discussing Pepsi cola.
The BuzzFeed text ad (https://www.buzzfeed.com/pepsi/a-thank-you-letter-to-summer) was marked at a Pepsi brand publisher article and is written as a thank you letter to summer with memes, gifs and links to Pepsi's social media pages on the side of the article.

Participants and Procedure
Undergraduate students at a large southern university in the United States served as research samples in the experiment. The use of a student sample was appropriate because the goal of the study was to test if the expected effects appeared, not generalizability of a larger population (Lang, 1996). The researchers provided a brief description of the experiment and assurance of confidentiality. The researchers also explained to students how the questionnaire should be answered. Students who expressed interest were then asked to sign a consent letter prior to the study. The total sample size was 82, with approximately 97% in the age group of 18 to 25. Each subject was assigned randomly to one of the two groups -Group 1 (native ads embedded in news content; n=44) and Group 2 (native ads embedded in entertainment content; n=40). Within each group, there were two ads representing different ad formats (video vs. text/image). CNN Heroes (Subaru) and The New York Times (Ford) ads were included in Group 1. Saturday Night Live (Pepsi) and BuzzFeed (Pepsi) ads were included in Group 2.
Two surveys (one for each group) were designed on Qualtrics and tested with a small group of faculty and students to ensure effectiveness of the questions. These two surveys contain the same information except for the stimulus ads. During the experiment, students were asked to view the ads and answer questions after viewing each ad, including their attitude-toward-the-ad, ad trust, and brand interest. They were also asked to compare native ads to traditional television or print/display ads. Participants were not able to move on to the next page until they completed questions for the first ad. They were also asked to answer an open-ended question for each ad to get more in-depth information on their response to native advertising strategies.

Measures
The measurements for main variables (attitude-toward-the-ad, ad trust, brand interest) in this study were adapted from previous studies with necessary modifications.
Attitude-toward-the-ad. Muehling and McCann's (1993) five-item semantic differential scale was used to measure participants' attitude toward each ad. Participants were asked to rate the ad as "good-bad," "like-dislike," "favorable-unfavorable," "interesting-uninteresting," and "appealing-unappealing". The coefficient alpha was .94. Ad trust. Using a modified version of the five-item, seven-point semantic differential scales developed by Ohanian (1990), participants were asked to indicate whether they thought the ad was "undependable-dependable", "unreliable-reliable", "untrustworthy-trustworthy." The coefficient alpha for trustworthiness was .91.

Brand interest.
Participants' interest in the advertised brand was measured with a four-item, seven-point Likert scale (Machleit, Madden, & Allen, 1990). Participants were asked if they were intrigued by the brand, if they would like to know more about the brand, if they were curious about it, and if they thought learning more about the brand would be useful. The coefficient alpha was .92.
Comparison with traditional advertising. After viewing each ad, participants were asked, when compared to traditional advertising (television or print ads), whether the native ad they saw was more informative, entertaining, credible, persuasive, or enjoyable (on a seven-point scale).

RESULTS
Using media context (news vs. entertainment) as a between-group variable and advertising format (video vs. text/image) as a within-group variable, a repeated measures test was run to explore their impact on attitude-toward-the-ad, ad trust, brand interest, as well as the perceived differences between native ads and traditional advertising formats (television commercials and print/display ads). Main effects and interaction effects were observed and analyzed. Even though product characteristics were not the focus in the current study and participants were told not to take into consideration product-specific information in their responses, product involvement was measured for both automobiles (M = 4.05, SD = 1.59) and soft drink (M = 4.12, SD = 1.70) and there was no significant difference between these two products.

Role of Media Context (News vs. Entertainment)
The first research question (RQ1) investigated potential influences of media context. Significant main effect was found for ad trust, F(1, 82) = 6.18, p < .05 (see Table 1). Native ads embedded in news content were perceived as more trustworthy (M = 3.71, SD = .85) than those embedded in entertainment content (M= 3.30,SD= .86). No significant interaction effects, however, were found between advertising format and media context for attitude-toward-the-ad, ad trust, or brand interest.
Media context also affected how participants perceived the differences between native ads and traditional ads. Native ads embedded in news content were seen as more informative (M = 5.07, SD = 1.64), compared to traditional ads, than ads embedded in entertainment content (M = 3.13, SD = 1.71). Native ads embedded in news content were also seen as more credible (M = 4.82, SD = 1.63), compared to traditional ads, than ads embedded in entertainment content (M = 3.60, SD = 1.52). However, native ads embedded in entertainment content were seen as more entertaining (M = 5.24, SD = 1.48), compared to traditional ads, than ads embedded in news content (M = 4.48, SD = 1.86).
In terms of interaction effects, significant results were found for only one category -"enjoyable." There was no difference between video and text/image native ads for the news group. For native ads embedded in entertainment programming, however, video native ads (M = 5.65, SD = 1.67) were seen as more enjoyable than text/image ads (M = 4.38, SD = 2.06), when compared to traditional advertising formats F(1, 82) = 5.46, p < .05 (see Table 1).

Qualitative Analysis of Comments on Native Advertising
Participants were asked to comment on the practice of native advertising in an open-ended question after viewing each ad.
Overall. The comments were positive, in general. About 86.9% of participants (73 out of 84) thought the stimulus ads were creative and entertaining, although some did feel that these ads could be a little deceiving and sometimes take away attention from the product. Some of the comments include: "I prefer this type of ads rather than the traditional." "It captures attention more than a regular ad because of the interactive nature." "I like how there's actual story behind it and they aren't just telling you to buy the product." "As long as it's marked as an advertisement not a news story or proper documentary, it's fine." Ad Format. Participants expressed slightly stronger love for video native ads, compared to text/image ads. Eleven out of 84 participants left neutral or negative comments for text/image ads, while only eight out of 84 participants did so for video ads. Generally, text/image native ads were deemed as less entertaining than video ads. Some of the comments include: "This ad is less entertaining as it's mostly facts and number… this method of advertising is less effective than video advertising." "It's informative but boring." "Much less likely to engage with this content than a video." However, a few participants did indicate that text/image ads would be more informative and credible. For example: "It seems more credible than the TV ad because there aren't actors making it seem fake." "I think I'd prefer a web-based ad. This one was more creative and provided a lot more fact." Media Context. There was not much criticism for native ads embedded in entertainment content, except that a few participants thought television is becoming too much commercialized. However, for native ads embedded in news content, there were more concerns about journalism ethics and misleading information. Some of the comments include: "Didn't even want to read it. Didn't like how they were tricking me." "Some people may be angered if they are fooled to believe that it is a news article." "It's harder to identify as an advertisement."

DISCUSSION
The current research examined a newer form of native advertising -native ads on TV. The effects of video ads and text/images ads were compared and the role of media context was investigated. Participants' perceptions of native ads, when compared to traditional ads, were also analyzed in relation to advertising format and media context.

Video vs. Text/Image Native Ads
The world of communications is becoming more visual with the help of digital media. Netflix, Amazon and other video streaming sites keep attracting new customers. Facebook just recently rolled out a new platform for video -"Watch." There is an urgent need for media scholars to expand research in the effects of video content.
In the current study, even though advertising format did not seem to affect ad trust or brand interest, participants reported more favorable attitude-toward-the-ad toward video native advertising (M = 5.29, SD = 1.50) than text/image advertising (M = 4.85, SD = 1.57). This finding was further reinforced in participants' comments. The majority of the participants thought the video ads were more entertaining and enjoyable than text/image ads. However, a few participants did indicate that text/image ads were more informative and credible. Previous studies have shown that messages with high level of vividness could have positive effects on people's attitude (Ching, Tong, Chen, & Chen, 2013;De Vries, Gensler, & Leeflang, 2012). Vivid content may also be able to generate an aroused and excited feeling (Hutter & Hoffman, 2014). This might help explain the results of the current study.
Advertising format also seems to affect the comparison between native advertising and traditional advertising. Text/image native ads were seen as more informative than traditional ads, when compared to scores for video native ads. However, video native ads scored higher on entertaining, enjoyable, and persuasive. Again, this is consistent with previous research in vividness effects. This is good news for marketers who are planning to invest in video native ads on TV and online. Favorable attitude-toward-the-ad often leads to favorable brand attitude and stronger purchase intention (MacKenzie, Lutz, & Belch, 1986;Marchand, 2010), even though these factors were not tested in the current study. However, as many participants mentioned in the comments, to be effective, video native ads should not be deceptive or take away attention from the product.

Media Context
According to Pew Research Center (2016), television is currently the most widely used news platform, as 57% of U.S. adults get news from local, cable, or network television. Approximately 38% of U.S. adults get news from online news sites or social media. News content is as important as entertainment content for the TV industry and how these two types of content affect native advertising effectiveness is worth investigating.
Even though no interaction effects were found between ad format and media context, native ads embedded in news content were perceived as more trustworthy than those embedded in entertainment content. Media context also affected how participants perceived the differences between native advertising and traditional advertising. Native ads embedded in news content were seen as more informative and credible than traditional ads, compared to the scores for video ads. This is probably because the current research used example ads from two highly respected news organization, CNN and The New York Times. As suggested in previous literature (e.g., Becker-Olsen, 2003;Tutaj & Van Reijmersdal, 2012), the credibility and authority of a website would be transferred to the ads placed on the websites. However, in participants' comments, many were concerned about the impact of native advertising on news organizations in terms of media credibility.
For news organizations that plan to adopt native advertising for news programming, the reputation of the organization/program may add credibility to the ads, but publishers must be able to maintain a balance between profits and news objectivity. For example, NBC News EVP of Digital Media, Peter Naylor, stressed that the network has two rules for the native ads -make sure the audience understands the difference in editorial content and no one in the editorial department works on the native ads (Taintor, 2013). This is an important separation that media brands need to make in regards to sponsored content. Because native ads look so similar to regular news or editorial content, viewers might identify the article as a real news story if it wasn't for the identification of "sponsored content" on the byline, which could be easily overlooked.

Limitations and Directions for Future Research
The following limitations must be considered when evaluating findings of the current study. First of all, undergraduate students were used as samples. A more diverse group of participants would help provide insight on the effects of TV native advertising on individuals with different demographic backgrounds such as age, education, and income. In addition, the current study only focused on text/image native ads on websites. As we are seeing more and more video native ads online, it would be necessary to compare native ads on TV to video native ads online, including websites as well as social media. Finally, real ads were used in this study. Even though it adds authenticity to the research procedure, some of the factors might not be highly controlled, such as design elements, brand attitude. Future research may include stimuli designed for specific research goals in order to generate better results.