INTRODUCTION
Online marketing has grown rapidly over the past years and has become a key component when making marketing mix decisions. Over 2014, internet advertising revenues in the U.S. increased 16% from the previous year to a total of $49.5 billion (IAB 2015). In addition, according to Online Publisher Association, video advertisements have reached the majority of the Internet users (66%), and have resulted in an action from almost half of the users. The significant business opportunities provided by the online environment are reflected as the increasing amount of attention for online advertising in academic research (Mei et al. 2008). It is yet to be researched how online videos and the accompanied online video advertisements interact and how advertisement customization can play a role to possibly optimize user experience and therefore the effectiveness of online video advertisements. Particularly, congruity forms a key opportunity for customization in embedded online video advertising. Congruity is defined as consistency between the ad and video or the advertised product and video (Moore et al. 2005). Evidences in extent literature have shown the positive effects of congruity on, for example, the processing fluency of the ad, the consumer attitudes towards the ad and the persuasiveness of the ad (Moore et al. 2005). However, in the context of online video advertising, limited research has been conducted on this topic. The main goal of this research is to explore the effect of online video advertising congruity on consumer attitudes towards the online video, advertisement and advertised product. Therefore, this research attempts to examine the information processing of a consumer and how congruity acts in this process to affect consumer's preferences. Due to the nature of this research question, an online experiment is used to test the hypothesized relationships of the conceptual model. The collected data in this study were examined using partial least squares (PLS).
CONGRUITY AND PROCESSING FLUENCY
Congruity can be defined in terms of similarity. The perceived similarity between two given stimuli of equal importance, determines their congruity (Rokeach & Rothman, 1965). Extensive research has been conducted on the effects of congruity in advertising. For example, studies have focused on congruity between involvement types of ads and TV programs (Sharma, 2000), viewer's mood-ad content congruity (Kamins, Marks, & Skinner, 1991), and program context- advertised product congruity (De Pelsmacker, Geuens, & Anckaert, 2002). Previous research from (Moore, Stammerjohan, & Coulter, 2005) has shown that congruity has a positive effect on consumer attitudes. Research on online banner advertisements on websites has indicated that advertisements should be consistent with the website brand (Newman, Stem Jr, & Sprott, 2004). Appeal congruity can then be achieved when the appeal of the online video and the appeal of the embedded ad are similar: an emotional (affective) video paired with an emotional ad can be deemed congruent. Similarly, an informational ad will be more effective when embedded in an informational video. In this study, Ad-and-Video congruity represents such type of congruity. Based on extent research on processing fluency, it can be assumed that online video advertising congruity, either in terms of appeal or content, will result in more processing ease of the ad as the information provided by the video and ad is similar. The following hypotheses are formulated:
Hypothesis 1: Congruity between the appeal of the online video and the appeal of the advertisement has a positive effect on processing fluency.
Hypothesis 2: Congruity between content of the online video and the advertised product has a positive effect on processing fluency.
We further propose that the above process is influenced by ad/video appeal or execution format (emotional vs. informational). An informational ad format is defined as an ad execution designed to appeal to the rationality of the receiver by using objective information describing a brand’s attributes or benefits, while an emotional ad format as an ad execution designed to appeal to the receiver’s emotions by using drama, mood, music and other emotion-eliciting strategies (Yoo & MacInnis, 2005). When an individual is watching an emotional ad or video, the possible effect of Ad-video or Product-video congruity on a consumer's processing fluency will be attenuated. The following hypotheses are formulated:
Hypothesis 3a: Congruity between the appeal of online video and the appeal of embedded advertisement has a stronger effect on processing fluency for informational appeals than for emotional appeals.
Hypothesis 3b: Congruity between content of the online video and the advertised product has a stronger effect on processing fluency for informational contents than for emotional contents.
Involvement and processing fluency
With regards to online advertising, research has shown that the degree to which a consumer is involved with the topic of a website also affects the processing of online advertising. It can be argued that when an individual is not involved in watching an online video, its effect on a consumer's processing fluency will be absent. On the other hand, when a consumer is highly involved in watching an online video, it is more likely that the viewer experiences processing fluency. As a result, in this situation, involvement may actually affect processing fluency of the viewer. Based on these findings, the following hypothesis is formulated:
Hypothesis 4: Involvement has a positive effect on the processing fluency of consumers.
Post-viewing attitudes
Research has also indicated the mediating role of an individual’s affective response from processing fluency to its evaluation (Winkielman & Cacioppo, 2001). A viewer of an online video is shown an ad that has the same appeal as the video, making it rather easy to process. The viewer appreciates this as it takes less effort to view the ad and to relate to advertised product to the content of the video and the viewer's interests. As a result the viewer is more likely to generate a positive attitude towards the ad, the advertised product and the video. The insights lead to the formulation of the following three hypotheses:
Hypothesis 5: Processing fluency has a positive effect on the attitude towards the online video.
Hypothesis 6: Processing fluency has a positive effect on the attitude towards the advertisement.
Hypothesis 7: Processing fluency has a positive effect on the attitude towards the advertised product.
Figure 1 shows the conceptual model, which summarises the hypothesized relationships.
<Insert Figure 1 about here>
METHODOLOGY
We conducted an experiment in the form of a 2 (type of appeal: informational vs. emotional) × 2 (Ad-Video congruity: congruent vs. incongruent) × 2 (Product-Video congruity: congruent vs. incongruent) full-factorial between-subject design. Video and advertisement appeal are each measured by the three items derived from the study by Moore et al. (1995) on the individual difference in response to advertising appeal. Processing fluency is measured by two items of the study by Lee and Aker (2004). Three items of the study from Zaichowsky (1985) are used to measure the involvement of the participants. The three items related to the attitude towards the video, advertisement and product are derived from the respective studies from Russell et al. (2004) and Kozup, Creyer and Burton (2003). Returned questionnaires numbered 276 responses with 24 incomplete data, resulting in 252 valid responses for data analysis. The collected data in this study were analyzed using partial least squares (PLS).
Measurement model
A measurement model was used to evaluate reliability, convergent validity, and discriminate validity. A structural model was used to determine the significance and association of each hypothesized path, and the explained variance. Reliability was examined by composite reliability values above the 0.70 benchmark (Fornell and Larcker 1981). Results show all values above 0.70, indicating satisfactory reliability. Convergent validity was examined by all indicator loadings that were significant and exceeding 0.7 and average variance extracted (AVE) by each construct exceeding 0.50 (Fornell and Larcker 1981). Results shows all indicator loadings above 0.70 and all AVEs exceeding 0.50, indicating satisfactory convergent validity. Discriminant validity is demonstrated by the square root of the AVE for each construct exceeding the correlations between the constructs (Chin 1998). These results indicate that the proposed models have good reliability and validity.
Structure model
Figure 2 shows the standardized path coefficient, the significance of each hypothesized path and the variance explained (R²). H1 and H2 investigate the effect of Ad-Video congruity and Product-Video congruity on processing fluency. Results show that Product-Video congruity is positively related to processing fluency (β=0.115, t-value=2.131, p<0.05), supporting H2. However, the effect of Ad-Video congruity on processing fluency is not significant ((β=0.022, t-value=0.377, p>0.05). Thus, H1 is not supported. H3a and H3b examine the moderating effect of the type of appeal. The hypothesis testing of H3a and H3b is discussed under the multi-group analysis (PLS-MGA). H4 proposes that involvement has a positive effect on the processing fluency of consumers, which is supported by the results (β=0.399, tvalue= 6.778, p<0.05). H5, H6, and H7 investigate the effects of processing fluency on post-viewing attitudes. Processing fluency is positively related to attitude towards the video (β=0.192, t-value=2.838, p<0.05), advertisement (β=0.426, t-value=7.607, p<0.05) and the focal product (β=0.422, t-value=7.171, p<0.05). These results support H5, H6 and H7.
<Insert Figure 2 about here>
We hypothesize that the type of appeal (informational or emotional) moderates the effects of Ad-Video congruity and Product-Video congruity on processing fluency in H3a and H3b. Specifically, the effects on under informational appeal condition are stronger. To test for group differences, we applied a multi-group analysis (PLS-MGA) approach. The PLS-MGA results for both informational and emotional groups, show that that the positive effect of Ad-Video congruity on processing fluency is only significant for informational appeals (β=.150, t-value=1.99, p<.05) but not emotional appeal (β=-.010, t-value=.15, p>.05), supporting H3a. Similarly, results show that that the positive effects of Product-Video congruity on processing fluency is only significant for informational appeals (β=.170, t-value=2.27, p<.05) but not emotional appeal (β=.050, t-value=.57, p>.05), supporting H3b.
CONCLUSION
Results show that congruity is related to the improvement of processing fluency only for informational ads/videos. This finding provides some support to the processing fluency theory which holds that less discrepant information should result in a higher processing fluency and extends the existing literature by investigating the appeal/execution format as a boundary condition. Involvement itself has been found to be positively related to processing fluency. Thus, when a consumer is initially more involved in watching a video, the processing fluency increases. This is in line with previous research that proved that involvement positively affects the attention that is given to advertisements and ultimately also the attitude of consumers.

• Zhiying Jiang(Singapore University of Social Sciences, Singapore)
• Chong Guan(Singapore University of Social Sciences, Singapore)
• Meilin.Zhang(Singapore University of Social Sciences, Singapore)
• Ivo. L. de Haaij(Erasmus University Rotterdam, Netherlands)