Keywords

1 Introduction

Social media has become an integral part of consumers’ daily lives. On average, global internet users spend 144 min per day browsing through social networks (Clement 2020). In particular social networks that offer online videos and streaming services have gained popularity. YouTube is the most successful social media platform in this realm. For instance, it was found that 90% of internet users in the United States accessed YouTube to watch online video, thus, ranking as the number one online video platform (Tankosvka 2021). Perhaps not surprisingly, this prompts worldwide brands and their marketers to use YouTube to disseminate their brands messages. The delivery of brand messages is often carried out by so-called social media influencers. Influencers are usually social media users who are at the center of a sufficiently large, engaged and trusting community of mostly young consumers and attract the interest of relevant brand marketers, who need to engage with these communities (Lou and Yuan 2019; Yuan and Lou 2020). There is a growing stream of research that investigates the effectivity of paid influencer marketing campaigns (for a review: Hudders et al. 2020). However, not much is known about how influencers and brands can reach their audience more organically. With people spending much of their screen time on YouTube, often just randomly browsing through content, there is a high potential to organically reach new customers, viewers, or subscribers. In fact, it is the only way to reach people on YouTube who are not already subscribers to one’s YouTube channel, thus, being of high importance to influencers and marketers. But what makes social media users click a video when they seemingly have an unlimited number of videos to pick from? YouTube offers a range of features that may influence people’s likelihood to click a video. Video thumbnails, which are snapshots of videos that people see as they’re browsing YouTube, are one of the most powerful tools to make people click a video. Similar important is the video title, which aims to get people’s attention through buzz words or capital letters. Once a video is clicked and consumed, research has shown that advertising messages on YouTube, in particular when they come from influencers, can increase brand awareness, brand attitudes and purchase intentions (Dehghani et al. 2016; Lee and Watkins 2016; Sokolova and Kefi 2020).

Against this backdrop, we aim to address two research questions: (1) Which thumbnail characteristics lead to increased views of videos from social media influencers? (2) Which other visual cues (e.g., video title, verification badge, display of video duration) lead to increased views of videos from social media influencers? Upon answering these research questions, we propose six hypotheses. First, drawing on color psychology, we hypothesize that a thumbnail with higher colorfulness increases video clicks. Second, drawing on the phenomenon of parasocial interaction, we expect the presence of the influencer in the thumbnail to increase video clicks. Third, underpinned by empirical work on consumer responses to emotions, we propose that emotional face expressions (i.e., surprise) will more likely gain people’s attention, thus leading to increased video clicks. Fourth, drawing on the elaboration likelihood model and negativity bias, we propose that textual cues that increase people’s attention such as capital letters or negative emotions will lead to increased video clicks. Fifth, drawing on expectancy theory, we expect an inverted U-shape between the displayed video duration and video clicks. Finally, using source credibility theory, we hypothesize that a verification badge leads to increased video clicks.

To test our hypotheses, we collect data from the most successful YouTube accounts and analyze the data with a range of different tools including image processing and language processing. First, we compile a list of 1,000 influencers with the most successful YouTube accounts (in terms of number of followers) in 10 different categories. These categories include gaming, comedy, beauty, film and others. The social media analytics website Socialblade is used to collect relevant data (i.e., name and channel ID) of the Top 100 accounts in each category. Second, we use YouTube’s Application Programming Interface (API) to collect information (e.g., thumbnail, number of views, and video title) on the latest 50 videos of each account. Third, we use the Image Colour Summarizer API to analyze the colorfulness, dominant colors and lightness of each of the thumbnails. Fourth, we utilize the Microsoft Face API to check whether a person is present in the thumbnail and analyze facial expressions (i.e., emotions) in each of the thumbnails. Finally, we use textual analysis program Language Inquiry Word Count to analyze language characteristics (i.e., emotions, capital letters, negations or pronouns) of each of the video titles.

Our study aims to make three main contributions. First, we use big data with a total of 50,000 video thumbnails and video titles from the most 1000 most popular YouTube accounts. Prior research on YouTube influencers is dominated by surveys (Lee and Watkins 2016; Sokolova and Kefi 2020; Xiao et al. 2018). Here, we study real-world thumbnails, video titles and most importantly, engagement metrices such as video views (as opposed to intentions to click a video). Second, rather than focusing on how the video content may influence consumer attitudes and behaviors, we focus on what makes consumers click the video in the first place. Thus, being able to investigate how brands and social media influencers can increase their reach beyond their usual following. In so doing, we provide specific advice on how changes in the thumbnail and the video title for a specific video can increase reach. Third, we draw on source credibility theory, elaboration likelihood model, color psychology, parasocial interaction, negativity bias and expectancy theory to provide an overview with multiple theoretical perspectives of the mechanisms involved when people browse through YouTube and make the decision to click a video or not.

2 Influencer Marketing

Influencer marketing involves the identification and use of social media influencers. In line with Lou and Yuan (2019, p. 2), we define a social media influencer as “first and foremost a content generator: one who has a status of expertise in a specific area, who has cultivated a sizable number of captive followers-who are of marketing value to brands- by regularly producing valuable content via social media”. An influencer creates and shares content related to niche areas of interest or expertise (e.g., politics, sports, gaming) on social media platforms such as YouTube, Twitter and Instagram. Individuals or consumers with an interest in the particular area can follow and interact with influencers to keep up with the latest news and trends.

Research has investigated influencer marketing’s impact in various contexts (e.g., De Veirman et al. 2017; Djafarova and Trofimenko 2019; Lou and Yuan 2019; Weismueller et al. 2020). For example, Lou and Yuan (2019) investigated how message value and credibility affect consumer trust, which ultimately influences brand awareness and purchase intention. Research into which content and influencer characteristics leads to increased value for brands - in form of brand awareness and purchase intention - is important, because it can showcase the effectivity of influencer campaigns. However, to the best of our knowledge, no research has investigated how influencer campaigns can increase their reach. This includes the reach within an influencer’s following (as not every follower necessarily clicks the video) and the reach beyond the influencer’s following. Investigating on how social media influencers can increase their reach is important as it can boost the performance of an influencer campaign.

3 YouTube’s Browsing Environment

This study focuses on YouTube, an online video-sharing platform, that allows users to create and upload videos that are viewed and shared by hundreds of millions of viewers. More specifically, over 2 billion logged-in users visit YouTube each month, people watch over a billion hours of video a day and those videos receive billions of views (YouTube 2021). YouTube’s browsing environment consist of a range of features, including a thumbnail, a video title, a verified badge and a label that displays the duration of the video. Figure 1 displays an overview of how the YouTube browsing environment is designed. YouTube has algorithms that usually detect user watching patterns and recommend content that consumers like based on their previous watch history. Thus, the YouTube browsing environment is heavily personalized. However, even when a video is suggested by the YouTube algorithm, it is not guaranteed that the user clicks on the video. Whether a user clicks on the video or not may depend on various features that were previously discussed. In the following, we examine those features in more detail.

Fig. 1.
figure 1

YouTube browsing environment

Full size image

4 The Impact of Thumbnails and Other Visual Cues on Video Clicks

4.1 Thumbnails

The most prominent feature in the YouTube browsing environment is the thumbnail. A thumbnail is either a randomly chosen frame of the video or a custom-made image from the video owner. Thumbnails act like billboards to help viewers decide to watch a video and often consist of a combination of photo, text and icons (see Fig. 2).

Fig. 2.
figure 2

YouTube thumbnail

Full size image

One of the most dominant characteristics in a thumbnail is color. Color theory has been widely applied in the marketing literature, including the context of brand recognition (Amsteus et al. 2015), product packaging (Garaus and Halkias 2020) and social media marketing (Yu et al. 2020). For instance, Yu et al. (2020) found that generally, individuals appeared more likely to respond to brighter and more saturated pictures on Instagram. Further, research on ad impressions found that more saturated images lead to an increase in click-through rates (Sokolik et al. 2014). More saturation in images appear to evoke greater levels of attention and arousal in people, which in turn increases their engagement with content that contain images with high saturation. Based on the above, we propose that increased colorfulness (saturation) leads to increased video clicks.

Another dominant characteristic is whether a person is present in the thumbnail, or not. Research has found that one of the main drivers of why people engage with influencer content is the para-social interaction (PSI) the audience creates with the influencer (Lou and Kim 2019; Sokolava and Kefi 2020). PSI creates a feeling of personally knowing the influencer. While PSI is usually related to how openly the influencer engages with the audience, we argue that the presence of the influencer in the thumbnail may increase the feeling of closeness to an influencer, therefore, leading people to click the video. Hence, we hypothesize that the presence of an influencer in the video increases video clicks.

If the influencer is present, it is of interest which emotions are expressed through the influencer’s face. The importance of emotions in affecting people’s attitudes and behaviors have been widely discussed in a marketing context (for a review: Gaur et al. 2014). Research on the use of emotions on YouTube found that emotional videos ads positively impact sharing of the ad. Moreover, it was found that an element of surprise can increase shares of YouTube ads (Tellis et al. 2019). Beyond emotional content, previous studies have shown that emotional facial expressions capture visual attention (Sawada and Sato 2020). In sum, emotional and surprising content and emotional facial expression can increase people’s attention and influence subsequent behavior. Based on the above, we suggest that surprising and emotional face expressions will lead to increased video clicks.

4.2 Video Title and Other Visual Cues

Perhaps, the second most dominant feature is the video title. A video title is restricted to 100 characters and the video owner can freely choose the title. Although, to the best of our knowledge, there is no research on video titles on YouTube and how they attract video views, there is a large number of studies on the importance of text in marketing (for an overview: Berger et al. 2020). In this study, we expect that text can increase people’s attention through capital letters and emotional words. Capital letters are used as an expression of emphasis, thus, making people more alerted (Posner and Boies 1971). In professional and personal settings capital letters are also used to signal strong emotions. Drawing on the elaboration likelihood model (Petty and Cacioppo 1986), we propose that people who are browsing through YouTube are in a low-effort mode, thus, being easily persuaded to click the video by visual cues that get their attention such as capital letters. Consequently, we expect the use of capital letters can increase video clicks.

Beyond capital letters, emotions are often expressed through positive and negative words. In our study, we can expect negative words to increase people’s attention, hence, increase the likelihood of people clicking the video. This is underpinned by people’s negativity bias (Rozin and Royzmann 2001; Vaish 2008), which refers to people’s tendency to attend more to negative stimuli rather than positive stimuli, which in turn has a greater effect on people’s attitudes and behaviors. Therefore, we propose that negative words will lead to increased video clicks.

Other visual cues include the video duration and the verification badge. Practitioners and marketers alike have tried to find the optimal duration for videos on YouTube (Chi 2018; Tellis et al. 2019). Consumers live in a time-poor environment, which indicates that a video should be long enough to convey critical information, but short enough to suit the time constrained environment of consumers.

Thus, we can expect users to already take the video duration into account before even clicking a video. This is line with expectancy theory, which proposes that an individual selects a specific behavior based on the expected outcome (Vroom 1964). In our context, it is the time investment in relation to the benefit the video provides. While a very short or a very long video can likely provide benefits to the viewer, we expect videos with a mid-range length to show the highest benefit in relation to viewer’s time investment. Thus, we argue an inverted U-shape between video duration and video clicks.

The verification badge next to the YouTube channel name is displayed when a channel has been verified by YouTube. To be eligible for verification, a channel needs to have at least 100,000 subscribers, a channel banner, description, profile picture, and content, and be active on YouTube (Google 2021). Consequently, the verification badge indicates credibility. Prior research on influencer credibility found that credibility can increase brand awareness and purchase intention (Lou and Yuan 2019; Weismueller et al. 2020). Drawing on source credibility theory, we hypothesize that the presence of a verification badge increases video clicks.

5 Conclusion

Influencer marketing has become an important marketing communications tool as it offers engagement with a large number of potential buyers in a short time period at lower cost than traditional advertising (Hudders et al. 2020). YouTube, which is one of the most successful social media platforms with more than 2 billion logged-in users each month, is the second most used platform for influencer campaigns (Mediakix 2019; YouTube 2021). Consequently, brands are regularly engaging those influencers to promote sponsored content on their YouTube channels. While a large number of studies focused on content characteristics and influencer characteristics to investigate how these characteristics influence consumer attitudes and behaviors, research on what makes people click YouTube videos from influencers in the first place remains largely unexplored. Our study intends to close this gap and examine which features (i.e., thumbnail characteristics and video title) increase video clicks. In so doing, we provide specific advice on how changes in the thumbnail and the video title for a specific video can increase reach. An increased reach enables marketers to expose their advertising message to a larger audience beyond the followers of the influencer. More specifically, the advertising message may reach people who have similar interests as the followers of the influencer, but are not yet subscribed to the influencer’s YouTube channel. What follows is a boost in performance for the influencer marketing campaign. A boost in performance is beneficial as influencer campaigns can increase brand awareness, brand attitudes and purchase intentions (Dehghani et al. 2016; Lee and Watkins 2016; Sokolova and Kefi 2020).