Content and Feedback Analysis of YouTube Videos: Football Clubs and Fans as Brand Communities

Abstract

The use of Web 2.0 tools has been transforming the interaction between companies and their clients, especially for those that are selling emotional products. Consumers are generating and sharing contents concerning their favourite products on the web. Even if this process has been widely acknowledged, only a few studies have been specifically devoted to the analysis of both the contents and the feedback the consumers receive from other users.

This article analyzes the online presence of sport brands through contents that are generated by sport clubs (official contents) and their fans (User Generated Content, UGC) on YouTube. After a description and classification of video contents, it examines the factors that influence the performance of the videos in terms of passive (videos views) and active behaviour (any kinds of interaction with videos) among the viewers.

In order to carry out this analysis, 125 YouTube channels were considered thereby accounting for a total of 375 videos.

Results show that official contents are those preferred by the users/consumers and that if the video displays a passive/purely informative content, the chance of getting an active behaviour from the users tends to decrease.

These findings may help companies manage their online presence, creating awareness about contents and information that should be spread and shared on the web.

Keywords

Social media internet marketing Web 2.0 Sport YouTube

Introduction

The use of Web 2.0 tools has been significantly transforming the interaction between companies and their clients, especially in the corporate communication of certain sectors where brands are seen as ‘love brands’ (Albert, Merunka & Valette-Florence 2008) and products are mainly emotional such as in sport where fans are particularly close to their favourite team, and therefore, do intervene in the management of the club itself.

Online users have been extending the time they spend on the net thereby also deeply changing their role. They are no longer passive actors and they are contributing to the generation of much online content (Users Generated Content, UGC) (Bernoff & Li 2008) in some of the most popular Web 2.0 websites such as Wikipedia, YouTube, Pinterest, Flickr, etc. According to Nielsen data (2012), the constant growth of their traffic confirms that such contents are valued and being used.

Considering that some Internet spaces are promoted directly by brands/companies (Pace 2008), it is clear that UGCs are somehow competing with those directly generated by the companies (Marketers Generated Content, MGC). Moreover much UGC across various media is brand-related and has the potential to shape consumer brand perceptions (Smith, Fischer & Yongjian 2012) but, to our knowledge, even if there are some studies in literature devoted to UGCs (Cheong & Morrison 2008; Krishnamurty & Dou 2008), only a few analyze the success of UGCs contents in terms of consumer views and feedback.

Although some previous research has already examined the role of Web 2.0 tools and the possible uses and benefits for companies (Constantinides & Fountain 2008; Hawkins 2009; Laroch, Habibi & Richard 2013; Tuten 2008), also in the sport field (Waters et al. 2011), contributions remain mostly descriptive and anecdotal. Considering the most famous online video sharing platform (YouTube) there is a lack of studies concerning which kind of video is most appreciated in terms of views by users and which kind of video generates any kind of reaction from the users. Only one study (Kruitbosch & Nack 2008) provides some evidence that professional videos are more popular than the users’ created ones.

This study is focused on the sport field for two, main reasons: first, because supporters are particularly keen to be active, and they often create communities (fan clubs) around their favourite teams, so interaction among fans and between the fans and their favourite football club is a standard for this sector; second, because sport companies have become more professional over the years and they are increasingly investing in their brands, which are often considered the most important asset of sport clubs (Bauer, Sauer & Schmitt 2010).

This study aims to understand the consumers’ preferences towards sports brand-related contents, generated by both supporters and clubs. We can refer to video success, and so measure it, from two main perspectives: passive and active. In fact, online content could be: (a) passively watched/read by users or (b) create engagement, that is, users can actively interact in some ways. Both behaviours generate value for the concerned companies, as confirmed by metrics developed and used to evaluate Internet websites and by online advertising models (Danaher 2007).

Literature Review

Even if there are only a few studies that examine the reason of success in terms of views (Kruitbosch & Nack 2008) and consumers’ active reactions to Web 2.0 contents, the existing literature provides nevertheless some useful references for other closely related topics. Studies concerning online and brand community (Ahonen & Moore 2005; Armstrong & Hagel 1996; Devasagayam & Buff 2008; Dholakia, Bagozzi & Pearo 2004; Grant, Heere & Dickson 2011; McWilliam 2000), consumer behaviour on social media and Web 2.0 ecosystem (Berthon et al. 2007; Berthon, Pitt & Campbell 2008; Daugherty, Eastin & Brigh 2008; El Ouirdi et al. 2014; Hanna, Rohm & Crittenden 2011; Kucuk & Krishnamurthy 2007), UGC and word of mouth (Cheong & Morrison 2008; Daugherty, Eastin & Brigh 2008; Ding et al. 2014; Krishnamurty & Dou 2008) were analyzed in order to provide a first framework for our research. Most of these studies are not specifically related to sport companies but scholars have become accustomed to use and/or adapt such frameworks related to corporate brand in order to analyze this particular kind of businesses (Bristow & Sebastian 2001; Grant, Heere & Dickson 2011). In fact, football clubs, similar to any other companies that operate with their names on the market, can be examined and seen as brands. As any other companies, they also sell products and more specifically their core products are events (i.e., football games) and they also can be assessed in terms of their brand value. For example Manchester United is a football club and its brand value stands at $ 269,000,000 according to Forbes.¹ The main difference with brands that do not operate in the sport field is mainly related to their customers’ (fans) behaviour since in general these are loyal (there is no brand switching) and keen to have a relationship with the company because they see themselves as part of that organization. Every football club has one or more fan clubs that are basically communities (Heere & James 2007) and at the same time they try to create their own community, so for all these reasons the literature devoted to the analysis of the behaviour patterns of users of communities and their Web 2.0 platform may be also applied to the specific topic we are dealing with.

First of all, scholars confirm that consumers are changing their role, creating online content relating to companies and becoming what they call ‘creative’ or ‘invading’ customers (Berthon et al. 2007; Vescovi, Gazzola & Checchinato 2010). They also state that online users might be divided into different categories (Bernoff & Li 2008) where two opposite groups can be identified: those who create the content and those who passively enjoy such content. This second category might also include the so-called ‘occasional makers’, since they contribute to the success of certain content in commenting, spreading or sharing it with other people. This thread confirms our research assumption that users also passively enjoy content uploaded online by other users or by companies. In defining features able to influence the success of content in terms of views and feedback we found three main threads that could be applied to this topic: (a) the subject of the content, (b) the source that creates the content and (c) the recommendations. In fact, according to Berthon, Pitt and Campbell (2008), the topic of online resources plays a key role in influencing the success of video, ads etc. The creation of online brand communities is, therefore, strictly related to the generation of online content (UGC), which is encouraged by the traditional Web 2.0 tools (Constantinides & Fountain 2008) and stands out as a very attractive marketing tool for sport clubs. Since UGC are part of the brand community contents, literature concerning virtual communities could be helpful to define the concept of categories of content. Usually communities might be subdivided into two major groups according to the kind of contents they share: ‘rational’, that is, which is mainly meant to convey information; and ‘emotional’, which deals more specifically with the interaction among users/members (Armstrong & Hagel 1996; Devasagayam & Buff 2008). The same classification is to be applied also to the UGCs as they might be emotional or rational too.

A second thread concerns the source of such contents. The source of a message in a communication process could affect the receiver’s feedback (Wells & Hackanen 1997; Wilson & Sherrell 1993), so it has a role in shaping the users’ behaviour. Knowing that Web 2.0 content is created both by the fans and the football clubs, we can assume that sources of messages can generate different feedbacks from users, but studies on that specific topic are still at an early stage of development (Goh, Heng & Lin 2013) so that it remains difficult to determine if UGCs are preferred to MGCs. To be considered UGCs, online contents need to fulfil some requirements, such as being published on a public and accessible website or social network; they need to demonstrate creativity and effort, and must be created by consumers in a non-professional way (Kaplan & Haenlein 2010); therefore, they are usually of poor quality in terms of design (Waldfogel 2009). Previous research concerning word of mouth compares content created by companies with content created by consumers (Goldsmith & Horowitz 2006; Sussan, Gould & Weisfeld-Spolter 2006) but the main subject of these studies refers to product recommendations and not to creative content. As Cheong and Morrison (2008, p. 38) state, ‘UGC has been closely aligned and often confused with eWOM, the two differ depending on whether the content is generated by users or the content is conveyed by users’, and the lack of information about consumers’ use of those two different contents points out and cements the need to study this topic. Another research that compares user and firms generated contents deals with the comments posted in response to the official advertisings and the consumers generated advertisings for a product (Pehlivan, Sarican & Berthon 2011) finding out that there was no indication of a source effect.

It is generally assumed that for an iconic brand–that is, a brand related to a so-called ‘myth’ or to outstanding, once-in-a-lifetime performances–consumers use ritual actions to experience the myth when using the products (Holt 2004) and are keen to have information about it—so in such case contents provided by a company (MGCs) are likely to enjoy more interest than those that are posted by other users (UGCs). If this assumption is confirmed, we are likely to validate the so-called theories of tribal marketing (Cova 2003) as well as those related to the brand community that encourage companies to create and provide their clients with contents and spaces. Football clubs could be considered as being able to generate much more successful content than their fans since they employ more (skilled) resources and they also have access to sources dealing directly with the products and services they provide the market with (usually games, athletes, merchandising, etc.). However, to be successful MGCs are to be developed according to the fans’ specific needs and have to differ from off-line contents, such as videos and images created for TV channels, in order to capture reach, but also intimacy and engagement.

A third thread that concerns the intrinsic feature of Web 2.0 is the content recommendation and diffusion. To be known and shared, in fact, online contents have to be spread and the traditional Web 2.0 tools are definitely enhancing such a process. Web applications allow users both to extract content from the websites where they were originally posted and to carry out their research through social tags, social bookmarks and favourites, tools that recall concepts like the general intelligence or the so-called folksonomies (Hamill, Attard & Stevenson 2009; Šnuderl 2008).

Therefore, two main research questions are posed to explore the features that affect the success of contents:

RQ1: What are the determinants of passive videos watching (videos views)?

In particular, starting from a review of the literature we can assume that three kinds of contents influence this passive behaviour even if we do not know if they affect it in a positive or negative way. These are: subject, source and recommendations diffusion. So we can split down the first main question as follows:

RQ1a: Does the content subject have any influence on the success of the video?

RQ1b: Do users prefer UGCs or MGCs?

RQ1c: Do recommendations influence the success of such contents?

Moreover while reach can be achieved in large numbers, it often does not translate into a true exchange with the users (Hanna, Rohm & Crittenden 2011). Therefore, the second research question deals with the success and return of a video in terms of active behaviour:

RQ2: Which are the content features able to produce a proactive feedback from users after they have watched a video?

Methodology

Sampling Method

In order to achieve our goals, we analyzed a series of videos—UGCs and MGCs—posted on YouTube and devoted to clubs having participated in the first and second highest divisions of the Italian football league. YouTube was chosen as the platform for this analysis because of the huge success this website has been achieving over the past year. Although some previous research has already studied the role of YouTube as a new media channel used by companies to either communicate to their stakeholders (Haridakis & Hanson 2009; Kaplan & Haenlein 2010; Pace 2008) or obtain data and contents (Huberman, Romero & Wu 2009; Waldfogel 2009), there is little research concerning the factors affecting the number of views and feedback responses (Kruitbosch & Nack 2008).

The first step of this research consisted in collecting and classifying UGCs and MGCs uploaded on YouTube related to teams that played in the first and second highest Italian soccer leagues. It should be noted that to post a video on YouTube, the user has to create his/her own channel, that represents his/her profile, and therefore, also his/her personal interests, passions and so on. This means that the channel represents the source of the videos.

In order to get a representative sample and to compare videos posted by users with those generated directly by the clubs, we searched for all channels featuring a kind of passion for a certain team competing in the Italian first or second division. We had to search for channels, instead of single videos, since they provide information on the user’s profile, as well as about videos, favourites, etc. so that we could identify who had posted the contents we were going to analyze.

Moreover, we searched for these channels through a series of tags based on the official denominations of the clubs, their traditional colours as well as the most common nicknames (such as Juve for Juventus) available on Wikipedia or on the online fan communities.

Then, considering that having at least one subscriber was perceived as the minimum threshold to identify the channel as being attractive, the analysis was restricted to the channels featuring at least one subscriber. Moreover, since users can post videos relating to different topics, we opted for hiving off those channels that featured videos portraying the passion for a certain team from those displaying also more generic contents, thereby focusing on channels that had at least five videos related to the team and where at least 50 per cent of the uploaded videos were related to the club itself.

The final sampling of our research consists of the three most viewed videos available on each of these channels, in order to identify those having enjoyed the most success among the users and at the same time to obtain a sample that stands at about 10 per cent of the total population (videos related to the identified supporters’ channels were in fact 3,845). The availability of channels related to the various clubs was assessed about a week after completion of the national leagues, in order to have stationary data in terms of access and upload that were, therefore, likely not to be influenced by the sportive success of some teams or controversies on facts and figures of the competition.

Measurement

In this research, the video is the unit of analysis, since it represents contents that users enjoy. In terms of measurement, YouTube videos can be considered as web pages because they have a single domain and metrics refer exactly to them. For each video the following statistics available on YouTube have been collected: number of views, uploading day, number of comments, total amount of ratings, number of users that tagged such videos as their favourites, number of times the video was clicked on after having been embedded on another website.²

Table 1.

Categories Identified to Classify the Contents of the Videos

Contents Description		Category
Videos displaying fans and their actions in and outside the stadium	Emotional	Fans starring in the video
Videos celebrating the team, a single player or anyway the passion people feel for that club		Celebrative content
Controversial or violent content addressed to the referees or to opposing teams		Controversies
TV video reporting on the team, highlights and additional contents related to the team and its performance	Rational	Passive content
Various videos displaying contents that are not included in the above categories	Rational	Other contents

Source: Authors’ own.

Generally, the measurement of success of a website depends on its purpose (Hong 2007) and the purposes for uploading YouTube videos are generally related to obtaining both views and interactions from users. Therefore, in order to measure the success of YouTube videos we use the concepts of passive (RQ1) and active (RQ2) user behaviour.

Contents Classification

In order to answer to the two research questions, RQ1 and RQ2, all videos have been classified according to their content. The criteria for the classification start with very specific sub-categories that were then regrouped into five major categories (see Table 1). As discussed in the literature review, online contents can be divided into ‘emotional’ and ‘rational’ (Armstrong & Hagel 1996; Devasagayam & Buff 2008). These two categories are too wide to generate homogeneous reaction from users; therefore, we decided to split the ‘emotional’ videos in three different sub-categories, that is, those displaying celebrative content fans starring in the video and controversies. The category fans starring in the video suggests, in fact, that the emotional subject is more related to fans than to the brand, whilst the category controversies suggests that we are facing a negative emotion, such us anger or impatience. Moreover, starting from the concept of brand advocacy (Cova, Giordano & Pallera 2007), videos featuring negative contents or controversies along with other teams or related to the refereeing have been separated from the others as they could generate a different user behaviour (Cova & White 2010). In the sport field feedback related to controversies takes the form of defensive comments from fans of opposing clubs or criticism of people not directly related to the club itself, such as the referees. That kind of feedback is not that relevant for the company as it does not bring any useful information or positive input.

We have also identified one category under ‘rational’ contents, named passive content, while those videos that could not be classified into the previous categories were regrouped into the last category, named other contents.

The traditional content analysis (Collis & Hussey 2003) was not applicable as the study deals with videos mainly related to fans and football matches. Moreover, analysis of web-based messages adds new complexity to the process of content analysis (Hwang, Mcmillan & Lee 2003; McMillan 2000).

The methodology for searching for and representing visual information is still at an early development stage. Although sport videos have been explored before (Rui, Gupta & Acero 2000; Sahouria & Zakhor 1999), most studies do not refer to the content in terms of topics of the video, but mainly to the content of the scene, such as shots in basketball and mainly from a structural point of view (Hanjalic, Lagendijk & Biemond 2001; Huang, Puri & Liu 2000), which is not in line with the purpose of this research. Moreover audio contents mainly provide either songs celebrating single players or the football club, or dialogues among fans that is far from our purpose of classifying the videos as a whole. Furthermore, trying to transcribe the contents could have been too subjective and not meaningful for the purpose of this study. So, to deal with the generation of contents, topics have been first classified (Table 1) in the aforementioned five categories to move then on to the analysis of videos and descriptive statistics of the sample.

Dependent Variables

In order to measure the success of videos in terms of passive enjoyment (RQ1), we used the daily number of views. This measure can be associated to the general concept of site exposure, which refers to the total number of visitor sessions at a website in a specific timeframe (Novak & Hoffman 1996). The use of metrics related to website traffic is well known in literature (Alpar Porembski & Pickerodt 2001) and number of views is the only traffic-related variable available on YouTube. Moreover, considering videos as a website page, we can state that views represent a good metric for monitoring the passive enjoyment.

The kind of feedback generated by users in response to videos they have viewed represents another way to measure the success of videos because the interaction is one of the key capabilities of the Internet (Palmer 2001). Therefore, in order to answer the second question, that is, RQ2, and to compare videos that have received different numbers of views, a specific index, called Index of Activity (IA), has been developed. This index, allows us to put the number of views beside an active feedback by the users in order to identify a relationship between these measures, and it has been calculated as follows:

{IA}_{i} \frac{V_{i}}{C_{i} F_{i} R_{i}}

(1)

where i = 1,…, N (N is the number of videos observed); V, C, F and R are the average daily measures respectively of the views, comments, favourites and ratings (see Appendix Table A1 for more details). Therefore, the IA is the ratio between the daily views of a video and the daily feedback it received (daily comments, favourites and rates).

To understand if a video was able to generate a number of interactions higher than the average one for the sector, so performed better in comparison to the competing videos, we compare the value of the index calculated for one video, that is, IA_i, with the average index $(\overset{}{IA})$ , calculated from the entire sample using the mean value of V, C, F and R. If the value of the IA_i, calculated for the ith video, is lower than $\overset{}{IA}$ , the ith video is likely to generate active interaction with the users since the amount of views that are necessary to get a comment, have it bookmarked, and have it rated stands below the average. In doing this comparison, a dummy variable (A) was calculated as follows per each video:

A_{i} \{\begin{matrix} 1 if {IA}_{i} \overset{}{IA} \\ 0 otherwise \end{matrix}

(2)

Then, this dummy variable was used as the dependent variable of the analysis created in order to answer to RQ2, for example, which are the content features able to produce a proactive feedback from users after they have watched a video.

Data Processing and Independent Variables

In order to compare results related to videos posted on different days, all data have been transformed into daily average data.

To test our assumptions, we first carried out a descriptive analysis of the sample. Then, we investigated the interaction occurring among the different variables and their influence on the total amount of views videos are enjoying (RQ1). After that, we have analyzed the interaction existing between the ‘A’ dummy variable and the contents of the video, in order to determine which topics are likely to get a more active involvement of the users (RQ2).

To achieve the first goal (i.e., to identify the factors influencing the passive enjoyment of the video—RQ1) a multiple linear regression model has been applied. As explained earlier, views are the dependent variable of this regression model. Consistently with RQ1 we chose as independent variables (see Appendix Table A1 for details):

The content of videos: the five categories have been transformed into dummy variables (Passive video, Controversial video, Celebrative video, Fans starring in the video, Other video) in order to verify their dependence from the number of views (RQ1a);

The source (clubs or users) that has been generating the video (Official) (RQ1b);

The number of times a single video has been recommended (F). We assume that if the video has been marked as a favourite resource by a single user, this is likely to increase awareness. Moreover as additional variable of recommendation we also consider the number of times a single video has been viewed after it has been posted on another website (CL). This is a kind of measure of recommendation, since users are viewing videos taken from YouTube by other users and embedded in other websites, thereby spreading the content through the web (RQ1c).

To achieve the second goal (i.e., to identify the features influencing the active enjoyment of the video—RQ2) a logistic regression model has been applied. The dependent variable is the dummy A and the independent variables are related to the video categories (see Appendix Table A1 for a description).

Description of the Sample

The YouTube channels fulfilling the earlier requirements are 125; 86.4 per cent of such videos deal with teams competing in the first division of the national league. The population consists of 3,845 videos; at first our sample included the three most viewed videos displayed on such channels, with a total of 375 units but the final sample used in this research consists actually of 374 videos because one was invalidated as data had not been correctly entered. Table 2 summarizes the main descriptive statistics related to the videos. It has to be pointed out that on average videos have been viewed 138 times per day and they have been receiving 0.30 comments per day.

Such data display a high variability: few videos have been viewed many times and the majority of the videos included in the sample have been played only by a few users, thus recalling the long tail model described by Anderson (2006).

Table 2.

Characteristics of the Sample

		Mean	Median	Min	Max	Interquartile Range
Videos*	Views (V)	138.32	24.72	0.24	8681.75	72.32
	Comments (C)	0.30	0.05	0	22.83	0.17
	Favourites (F)	0.26	0.05	0	16.75	0.14
	Ratings (R)	0.23	0.05	0	5.17	0.14
	Clicks to Link (CL)	7.57	0.44	0	610.50	2.23

Source: Authors’ own.

Note: *see Appendix Table A1 for a description of these variables.

Having a look at the 125 channels included in the sample, only four are to be considered official (MGC), since they have been created by the clubs themselves (all of them playing in the highest division), whereas the remaining channels have been generated by supporters or other people that are anyway not the owners of the brand/club. Table 3 summarizes the average values of the variables that describe the videos, divided according to their source (MGC/UGC) in order to compare these two categories.

Table 3.

Mean Value of Each Quantitative Variable Regarding Videos Depending on Official Nature of Their Respective Channels

	MGC (Official)	UGC (Non-official)
V	1479.30	93.87
C	0.78	0.29
F	2.63	0.18
R	0.96	0.20
CL	20.08	7.16

Source: Authors’ own.

Such groups show real differences, especially as far as the amount of average views per day is concerned: such values are indeed much higher for those videos that have been posted on official channels. Coming now to the contents of videos, Table 4 shows that views are mostly related to informative contents, such as videos showing news about the football clubs or performances by a single team and videos that users do not create, but take from news sites and then share on their YouTube channels. This confirms that fans do not refer to a single medium to get the information they are looking for, but actually they combine different devices such as television, tablets, smartphones and PCs. The web, therefore, becomes a tool that sports organizations are able to exploit in order to submit their contents with much more flexibility in terms of both time and space.

Table 4.

Frequency of Videos with Reference to the Category they Belong

Category	Composition (%)
Passive/informative content	36.36
Emotional content	28.34
Fans starring in the video	22.46
Controversies	5.35
Other contents	7.49
Total	100

Source: Authors’ own.

Table 5.

Average Values of the Descriptive Variables According to the Various Video Categories

		Passive	Celebrative	Fans Starring	Controversies
Video	V	121.78	211.50	51.99	67.91
	C	0.38	0.25	0.16	0.34
	F	0.14	0.48	0.17	0.12
	R	0.22	0.27	0.12	0.14
	CL	10.37	4.39	8.63	5.17

Source: Authors’ own.

About 50 per cent of the videos belong to the categories that actively involve consumers, such as videos with a celebrative content (28.34 per cent) or videos starring the fans themselves (22.46 per cent) where there is an even stronger active involvement, since such resources deal directly with the fans and the kind of activities they set up to support their teams in and outside the stadium, through—for instance—choirs or choreographies.

Coherently with the literature devoted to the brand communities, clients are interacting by sharing activities and experiences they have been living through and that were somehow related to a certain brand (Kozinets 1999). The availability of such videos seems also to validate the theories developed about the brand property, which has been conceived as belonging also to the customers (Berthon et al. 2007) since they can create channels where the real protagonists on stage are the fans themselves and their interaction with the club. Finally, a mere 5.35 per cent of the videos deals with controversies, and therefore, we can assume that generally users are not as interested in such topics as the people responsible for football clubs have traditionally believed (Checchinato & Gazzola 2009).

By crossing the video categories along with the descriptive variables (Table 5) we can notice that the highest values—in terms of daily average views (V), ratings (R) and favourites (F)—are registered for the resources that display celebrative content. It is also interesting to point out that videos having controversial contents do obtain the lowest figures in terms of daily average favourites. However, a high value registered for the daily comments may derive from customers feeling they are supposed to somehow defend their brand, thus confirming the importance of brand advocacy (Cova, Giordano & Pallera 2007).

We can also notice that videos starring fans are the least commented, viewed and rated. Such a low value registered for all those variables is probably due to the low quality of the videos’ images (Kruitbosch & Nack 2008).

Results

As for the exploratory purpose of our research to answer RQ1, we opted for verifying the possible linear correlation existing among the quantitative variables that describe videos (see Table 6). Such correlations, even though they differ in terms of their intensity, are all positive and the stronger linear relationship occurs between the average daily views (V) and favourites (F).

Since views are supposed to be an important factor in determining the success of an online resource, it is necessary to verify whether there is a linear relationship associating views to the independent variables identified.

Table 6.

Correlation Matrix Related to the Quantitative Variables Describing the Videos

	V	C	F	R	CL
V	1
C	0.39	1
F	0.91	0.18	1
R	0.46	0.75	0.41	1
CL	0.51	0.77	0.35	0.55	1

Source: Authors’ own.

The regression was estimated using a stepwise robust method in order to correct the possible heteroscedasticity of the error terms. The robust regression was estimated by using White’s robust variance–covariance matrix to generate robust standard errors for our statistics (White 1980).

Stepwise results are presented in Table 7, therefore, only significant variables are reported and the complete list of variables used in the model is described in Appendix Table A1.

Table 7.

Determinants of the Views

Independent Variables	Coefficient*	Robust Standard Error
Official (MGC)	381.30	167.36
F	392.97	68.43
CL	3.02	1.16
Fans starring in the video	−63.75	32.38
Constant	14.89	15.06

Source: Authors’ own.

Notes: *Number of observations = 374; Adjusted R² = 0.879; cut-off value a = 0.05.

The relevant independent variables are likely to provide us with good estimates about the dependent variable (adjusted R² = 0.879) and they also reflect the original assumptions: the V variable increases as long as the video was posted on an official channel and there is also a positive relation along with the CL variable as well as with the F variable; on the other hand, V decreases in case of videos featuring fans.

As far as the second research question is concerned, a robust logistic regression model has been adopted to verify whether the category the video belongs to is likely to stimulate an active feedback from the ith user (A_i= 1).³

This condition is verified through a single category (pseudo-R²of McKelvey & Zavoina (1975) = 0.125): if the video displays a passive/informative content, the chance of getting an active behaviour from the users tends to decrease by 76 per cent (Table 8).

Table 8.

Robust Logistic Regression Model Estimates about Active Feedbacks Coming from the Users

Independent Variables	Coefficient* (Odds Ratio)	Robust Standard Error
Passive/informative	−1.42 (0.24)	0.23
Constant	0.45	0.13

Source: Authors’ own.

Notes: *^{Number of observations = 371; McKelvey and Zavoina’s R²(1975) = 0.125.}

Videos that are very similar to television resources are those that enjoyed less feedback coming from the users. This result validates the assumption that to get a kind of interaction with the ‘surfers’, videos posted on the Internet do not have to be created with the same criteria that are normally exploited for television and that their success is strictly resulting from the provision of ad hoc contents (Cova, Giordano & Pallera 2007).

Discussion

The aim of this study was to examine the features that influence the success of YouTube videos in terms of passive and active user behaviours. The earlier analysis has increased the knowledge about UGCs and the users’ preferences about such resources in the sport field, providing also some information that is likely to support and help the sport brands in addressing or modifying the kind of feedback they have been receiving.

No one of the categories used to classify video contents is able to stimulate more views or active behaviours than others, but at the same time some categories generate less views and feedbacks: the category ‘fans starring in the video’ produces less views and the informative contents generate lower interactions. Therefore, we can conclude that users do not prefer watching other fans as actors in the videos and are less engaged if videos display informative content.

Concerning the contents, videos containing controversial content are more often commented than the average (0.34 versus 0.30), but at the same time they are the least chosen as favourites (Table 5). It confirms that consumers are loyal to the brand and they are interested more in getting news than in provocative contents (Kozinets 1999). Comments usually aim at defending the brand.

Results confirm that UGCs compete with MGCs in attracting the fans’ attention, as both UGC and MGC obtain views and feedbacks, but users prefer watching videos created by their favourite football clubs. These findings validate our research question concerning the role of the source in affecting the number of views registered for a certain video. Moreover this finding suggests that the source of messages still has a key role in generating feedback and so the classical communication model (Lasswell 1948) could be useful also in the new media environment even though with some updating. It is very common to find on the Internet first information generated by consumers about a certain company on unofficial pages, and only afterwards on the corporate website (Kaplan & Haenlein 2010). The fact that the fans prefer MGCs could be consistent with the issue of source credibility as in a context such as Web 2.0 people do not always know who uploads documents and writes news and comments, so a well-known sender positively affects the passive enjoyment of contents.

The important role of recommendations and contents diffusion in increasing the number of views of videos is confirmed: favourites and clicks on embedded videos have a positive and significant impact on the average views per day (see Table 7). This finding supports the idea that word of mouth is an important communication tool on the web as well (Chu & Kim 2011; Riegner 2007).

This study is an initial investigation on success of content created both by firms and users. As such there are some limitations further research will need to address this topic in the future. Even though the classification of the videos according to their contents depends on pre-determined criteria chosen by the researchers, it is still somehow associated with the researchers’ interpretation. Moreover, the index of activity used to evaluate the users’ behaviour does not allow identifying a range of values but refers simply to the average of the sample. This index is likely to be applied also to other topics, but the threshold value separating active and passive involvement should be revised according to the peculiarities of any eventual new area of study.

Managerial Implication

In spite of such limits, there are some useful managerial implications deriving from the results of this study, namely: the kind of feedback such content enjoys in terms of its views on YouTube and the active behaviour of users enjoying such content. The tendency that users display in creating online content confirms the concept of commonly owned brand (Berthon et al. 2007). If clubs are interested in using the web to interact with their fans or in creating online communities, they should therefore, develop communication patterns that are not addressed to the single fan but to the entire community. Such communication activities are, therefore, not to be conceived as two-way, but as addressed to a multi-nodal network (Rheingold 1993). This is even more evident for common platforms, as they feature both UGCs and MGCs, with the latter being the most viewed shared resources. Despite the specific field of this research the provision of an official resource is to be considered as an opportunity to start up a dialogue and communicate with the consumers, who otherwise would have to refer only to the feeds that have been posted by other users, without any feedback for the firm that has also no chance to monitor such materials in an effective way.

The results of this study show that contents generated directly by firms are the most viewed resources posted online, thus confirming that also in the sport field consumers use multiple media to keep informed about their passions and at the same time they go almost anywhere in search of the kinds of entertainment experiences they want (Jenkins 2006). Sport companies do not seem prepared to completely fulfil this need, since there were just a few—4 out of 125—official channels, but nevertheless they have been enjoying more average views than the unofficial ones. Moreover, official channels have a positive and significant influence on the course of such views (see Table 7). In comparison with users, football clubs have access to more resources, support and sources that are likely to generate content; at the same time fans appreciate and recognize the authority and reliability of the original source, especially in such a domain as football where news is coming from several different sources. This is the main difference between sport brands and other brands that are related to the previously mentioned characteristics of the field. Football companies do not need to receive recommendations by users to be credible and reliable since their products (games, merchandising, TV channel fee …) are bought for other reasons than their quality. Therefore, the main reason of success of UGC, the credibility and reliability, does not exist in the sport field.

Clubs actively entering such platforms might somehow lose a bit of control but at the same time they can collect and have access to a plethora of relevant information about fans habit. Moreover, since companies do not have any real tools to limit the brand-related actions performed by their fans, still—through the generation of their own online resources—they can attract their target to their spaces inside the not proprietary platforms and in doing so not completely lost the audience reached by UGCs. According to the results of our previous studies (Checchinato & Gazzola 2009), it should be noted that such dynamics are likely to affect also the corporate organization: within the firms there should be somebody who is in charge of managing the interaction with the users and identifying valuable content. Moreover, to fully display their potential, the online presence and the use of Web 2.0 tools have to be integrated into the general communication activities firms are implementing. Such an informative need is confirmed also in the analysis of the categories the videos included in the sample have been subdivided into; results let us understand that people are combining different sources to collect the information they are seeking for. At the same time, such contents are likely to generate a much more passive reaction among the users, since they are originally created to be passively viewed. Firms have, therefore, to differentiate the contents they are posting online from those placed offline, as they are not alternative but complementary. New professional skills, especially devoted to the generation of online contents, have to be developed accordingly (Cova, Giordano & Pallera 2007), in order to capture reach but also to generate engagement and enact the ‘prosuming’ users (Des Autels 2011). As far as views are concerned, fans are less interested in videos displaying themselves. That somehow means that fans are more interested in videos where the subject is the brand (both in an emotional and rational way) than in videos displaying other fans. It has to be pointed out that the quality of these videos is generally lower than the others and Kruitbosch and Nack’s (2008) findings reveal that this could be a reason for low performance for videos. This topic needs to be analyzed more in depth, taking into consideration all the variables that might affect users’ reaction. In order to enhance the views of a video, companies have to encourage recommendations both to bookmarks and to the sharing, and choose contents that are not have single fans as actors. On the other hand, in order to generate a reaction by the audience, videos displaying informative contents should be avoided.

Such information is very useful for clubs to improve their online presence taking into consideration the use or the implementation of simple Web 2.0 applications. Some examples are the provision of buttons that facilitate the sharing of contents on online platforms such as Facebook or Pinterest or their inclusion on blogs or websites especially devoted to social bookmarking activities such as Delicious. Since contents are positively affecting the views as well as the kind of feedback that is coming from the users, clubs should constantly monitor the online resources where supporters are discussing about their brand such as any other corporate brand. Aims and information collected in social network will be different, since sport brands do not really compete to gain customers’ loyalty as traditional companies have to do (Laroche, Habib & Richard 2013), but they could enhance their revenues listening to fans’ conversations and acting to fulfil their needs. Additional tools—similar to the index of activity—should also be developed in order to foster a better understanding of the feedback coming from the supporters, enabling the adaptation of the communication strategy accordingly. Through such a kind of monitoring, companies would provide users with the kind of contents they really appreciate, and therefore, they would also definitely get the active behaviour they are longing for.

In the coming future this research might be extended to compare different online platforms in order to achieve a better understanding of the behaviour patterns displayed by fans/supporters. There is no evidence that the results of this study may be extended also to other websites that do not display videos but different kind of online content.

Footnotes

Appendix

Table A1.

Description of the Variables that Describe Videos

Variables	Description
Quantitative
V	Number of average views per day
C	Number of average comments per day
F	Number of average favourites per day
R	Number of average ratings per day
CL	Clicks that the video enjoys per day having been embedded into other websites
Qualitative
Passive video	1 = video displaying a passive content; 0 = otherwise
Controversial video	1 = video related to controversies; 0 = otherwise
Celebrative video	1 = video displaying a celebrative content; 0 = otherwise
Fans starring in the video	1 = video starring fans/supporters; 0 = otherwise
Official	1 = video displayed on an official channel; 0 = otherwise

Source: Authors’ own.

Notes

References

Ahonen

T. T.

Moore

(2005). Communities dominate brand. Lond: Futuretext.

Albert

Merunka

Valette-Florence

(2008). When consumers love their brands: Exploring the concept and its dimensions. Journal of Business Research, 61(10), 1062–1075.

Alpar

Porembski

Pickerodt

(2001). Measuring the efficiency of web site traffic generation. International Journal of Electronic Commerce, 6(1), 53–74.

Anderson

(2006). The long tail. New York: Hyperion.

Armstrong

Hagel

J. III

(1996, May–June). The real value of on-line communities. Harvard Business Review, 74(3), 134–141.

Bauer

H. H.

Sauer

N. E.

Schmitt

(2010). Customer-based brand equity in the team sport industry: Operationalization and impact on the economic. European Journal of Marketing, 39(5/6), 496–513.

Bernoff

(2008). Harnessing the power of the Oh-So-Social web. MIT Sloan Management Review, 49(3), 36–42.

Berthon

P. L.

Pitt

Campbell

(2008). When customers create the ad. California Management Review, 50(4), 6–31.

Berthon

P. R.

Leyland

F. P.

McCarthy

Kates

S. M.

(2007). When customers get clever: Managerial approaches to dealing with creative consumers. Business Horizons, 50(1), 39–47.

10.

Bristow

D. N.

Sebastian

R. J.

(2001). Holy cow! Wait ’til next year! A closer look at the brand loyalty of Chicago Cubs baseball fans. Journal of Consumer Marketing, 18(3), 256–275.

11.

Checchinato

Gazzola

(2009). Web 2.0: un’indagine sulle società di calcio italiane. Mercati & Competitività, 2009(1), 45–70.

12.

Cheong

H. J.

Morrison

M. A.

(2008). Consumers’ reliance on product information and recommendations found in UGC. Journal of Interactive Advertising, 8(2), 1–29.

13.

Chu

S. C.

Kim

(2011). Determinants of consumer engagement in electronic word-of-mouth (eWOM) in social networking sites. International Journal of Advertising, 30(1), 47.

14.

Collis

Hussey

(2003). Business research: A practical guide for undergraduate and postgraduate students. Palgrave Macmillan Ltd.

15.

Constantinides

Fountain

(2008). Web 2.0: Conceptual foundations and marketing issues. Journal of Direct, Data and Digital Marketing Practice, 9(3), 231–244.

16.

Cova

(2003). Il marketing tribale. Milano: Il Sole 24 Ore.

17.

Cova

White

(2010). Counter-brand and alter-brand communities: The impact of Web 2.0 on tribal marketing approaches. Journal of Marketing Management, 26(3–4), 256–270.

18.

Cova

Giordano

Pallera

(2007). Marketing non convenzionale. Milano: Il Sole 24 Ore.

19.

Danaher

P. J.

(2007). Modeling page views across multiple websites with an application to internet reach and frequency prediction. Marketing Science, 26(3), 422–437.

20.

Daugherty

Eastin

Bright

(2008). Exploring consumer motivations for creating user-generated content. Journal of Interactive Advertising, 8(2), 1–24.

21.

DesAutels

(2011). UGIS: Understanding the nature of user-generated information systems. Business Horizons, 54(3), 185–192.

22.

Devasagayam

Buff

C. L.

(2008). A multidimensional conceptualization of brand community: An empirical investigation. Sport Marketing Quarterly, 17(1), 20–29.

23.

Dholakia

U. M.

Bagozzi

R. P.

Pearo

L. K.

(2004). A social influence model of consumer participation in network– and small-group–based virtual communities. International Journal of Research in Marketing, 21(3), 241–263.

24.

Ding

Phang

C. W.

Tan

C.-H.

Sutanto

(2014). The role of marketer- and user-generated content in sustaining the growth of a social media brand community, 47th Hawaii International Conference on System Sciences, 1785–1792.

25.

Goldsmith

R. E.

Horowitz

(2006). Measuring motivations for online opinion seeking. Journal of Interactive Advertising, 6(2), 3–14.

26.

Goh

K. Y.

Heng

C. S.

Lin

(2013). Social media brand community and consumer behavior: Quantifying the relative impact of user-and marketer-generated content. Information Systems Research, 24(1), 88–107.

27.

Grant

Heere

Dickson

(2011). New sport teams and the development of brand community. European Sport Management Quarterly, 11(1), 35–54.

28.

Hamill

Attard

Stevenson

(2009). National destination marketing organisations and Web 2.0. Mercati e Competitività, 2009(1), 71–94.

29.

Hanjalic

Reginald

L. L.

Biemond

(2001). Recent advances in video content analysis: From visual features to semantic video segments. International Journal of Image & Graphics, 1(1), 63–81.

30.

Hanna

Rohm

Crittenden

V. L.

(2011). We’re all connected: The power of the social media ecosystem. Business Horizons, 54(3), 265–273.

31.

Haridakis

Hanson

(2009). Social interaction and co-viewing with YouTube: Blending mass communication reception and social connection. Journal of Broadcasting & Electronic Media, 53(2), 317–335.

32.

Hawkins

(2009). Twitter, Facebook and beyond: How to manage your company’s online presence. Office Pro, 69(4), 20–23.

33.

Heere

James

J. D.

(2007). Sports teams and their communities: Examining the influence of external group identities on team identity. Journal of Sport Management, 21(3), 319–337.

34.

Holt

D. B.

(2004). How brands become icons. Boston, MA: Harvard Business School Press.

35.

Hong

I. B.

(2007). A survey of web site success metrics used by Internet-dependent organizations in Korea. Internet Research, 17(3), 272–290.

36.

Huang

Puri

Liu

(2000). Multimedia search and retrieval: New concepts, system implementation, and application. IEEE Transactions on Circuits & Systems for Video Technology, 10(5), 679–692.

37.

Huberman

B. A.

Romero

D. M.

(2009). Crowd sourcing, attention and productivity. Journal of Information Science, 35(6), 758–765.

38.

Hwang

Mcmillan

S. J.

Lee

(2003). Corporate web sites as advertising: Analysis of function, audience, and message strategy. Journal of Interactive Advertising, 3(2), 10–23.

39.

Jenkins

(2006). Convergence culture: Where old and new media collide. New York/London: NYU Press.

40.

Kaplan

A. M.

Haenlein

(2010). Users of the world, unite! The challenges and opportunities of Social Media. Business Horizons, 53(1), 59–68.

41.

Kozinets

R. V.

(1999). E-tribalized marketing? The strategic implications of virtual communities of consumption. European Management Journal, 17(3), 252–264.

42.

Krishnamurthy

Dou

(2008). Advertising with user-generated content: A framework and research agenda. Journal of Interactive Advertising, 8(2), 1–7.

43.

Kruitbosch

Nack

(2008). Broadcast yourself on YouTube. Proceeding of the 3rd ACM international workshop on Human-centered computing—HCC ’08, 7–10. New York, New York, USA: ACM Press.

44.

Kucuk

S. U.

Krishnamurthy

(2007). An analysis of consumer power on the Internet. Technovation, 27(1–2), 47–56.

45.

Laroche

Habibi

M. R.

Richard

M.-O.

(2013). To be or not to be in social media: How brand loyalty is affected by social media? International Journal of Information Management, 33(1), 76–82.

46.

Lasswell

(1948). The structure and function of communication in society. In Bryson

Lyman

(ed.), The communication of ideas (pp. 37–51). New York: Institute for Religious and Social Studies.

47.

McKelvey

Zavoina

(1975). A statistical model for the analysis of ordinal level dependent variables. Journal of Mathematical Sociology, 4(1), 103–120.

48.

McMillan

S. J.

(2000). The microscope and the moving target: The challenge of applying content analysis to the world wide web. Journalism and Mass Communication Quarterly, 77(1), 80–98.

49.

McWilliam

(2000). Building stronger brands through online communities. Sloan Management Review, 41(3), 43–54.

50.

Nielsen (2012). State of the media: The social media report. Retrieved 30 November 2014, from: http://www.nielsen.com/us.en/insights/reports/2012/state-of-the-meida-the-social-media-report-2012.html

51.

Novak

T. P.

Hoffman

D. L.

(1996). New metrics for new media: toward the development of web measurement standard. Retrieved 2 March 2010, from http://elabresearch.ucr.edu/blog/uploads/papers/NewMetricsforNewMedia_TowardtheDevelopmentofWebMeasurementStandar.pdf

52.

Pace

(2008). YouTube: An opportunity for consumer narrative analysis? Qualitative Market Research: An International Journal, 11(2), 213–226.

53.

Palmer

J. W.

(2001). Web site usability, design, and performance metrics. Information Systems Research, 13(2), 151–167.

54.

Pehlivan

Sarican

Berthon

(2011). Mining messages: Exploring consumer response to consumer-vs. firm-generated ads. Journal of Consumer Behaviour, 10(6), 313–321.

55.

Rheingold

(1993). The virtual community: Homesteading on the electronic frontier. New York: Addison-Wesley.

56.

Riegner

(2007). Word of mouth on the web: The impact of web 2.0 on consumer purchase decisions. Journal of Advertising Research, 47(4), 436–447.

57.

Rui

Gupta

Alex

(2000). Automatically extracting highlights for TV baseball programs. In Proceedings of the eighth ACM international conference on Multimedia (MULTIMEDIA ’00). ACM, New York, USA, 105–115.

58.

Sahouria

Zakhor

(1999). Content analysis of video using principal components. IEEE Transactions on Circuits & Systems for Video Technology, 9(8), 1290–1298.

59.

Smith

A. N.

Fischer

Yongjian

(2012). How does brand-related user-generated content differ across YouTube, Facebook, and Twitter? Journal of Interactive Marketing, 26(2), 102–113.

60.

Šnuderl

(2008). Tagging: Can user-generated content improve our services? Statistical Journal of the IAOS, 25(3/4), 125–132.

61.

Sussan

Gould

Weisfeld-Spolter

(2006). Location, location, location: The relative roles of virtual location, online word-of-mouth (eWOM) and advertising in the new-product adoption process. Advances in Consumer Research, 33(1), 649–650.

62.

Tuten

T. L.

(2008). Advertising 2.0: Social media marketing in a Web 2.0 world. Westport: Praeger.

63.

Vescovi

Gazzola

Checchinato

(2010). Invading customers: New markets relationships. Journal of Marketing Trends, 1(1), 41–57.

64.

Waldfogel

(2009). Lost on the web: Does web distribution stimulate or depress television viewing? Information Economics and Policy, 21(2), 158–168.

65.

Waters

R. D.

Burke

K. A.

Zachary

H. J.

Buning

J. D.

(2011). Using stewardship to cultivate fandom online: Comparing how NFL teams use their websites and Facebooks to engage their fans. International Journal of Sports Communication, 4(2), 163–177.

66.

Wells

Hackanen

E. A.

(1997). Mass media & society. London: Ablex Publishing Corporation.

67.

White

(1980). A heteroscedasticity-consistent covariance matrix estimator and a direct test for heteroscedasticity. Econometrica, 48(4), 817–838.

68.

Wilson

E. J.

Sherrell

D. L.

(1993). Source effects in communication and persuasion research: A meta-analysis of effect size. Journal of the Academy of Marketing Science, 21(2), 101–112.