Let’s Get Real (Time)! The potential of real-time marketing to catalyze the sharing of brand messages

Abstract

Social media are increasingly populated with brand messages that are linked to timely events, a practice that is also known as real-time marketing (RTM). In this study, we examine whether RTM is an effective strategy to boost sharing behaviour, and if so, what moment- and content-related characteristics contribute to its effectiveness. A content analysis of brand tweets from Nielsen’s top-100 advertisers (n = 1500) shows that RTM positively affects word of mouth. RTM is especially a more effective strategy when brand messages are linked with unpredictable events (vs. predictable). This can be explained by the meaningfulness dimension of creativity; brands make a more meaningful connection to timely moments in unpredictable RTM than in predictable RTM. Furthermore, we found support for the beneficial effects of moment-driven visuals; RTM messages are shared more when public events are visually integrated with those messages. No such effect was found for moment-driven hashtags.

Keywords:

• Brand content
• creativity
• news jacking
• real-time marketing
• social media
• word of mouth

Introduction

The introduction of social media platforms was heralded for its potential to reach the public at large. Consumers organized themselves as networks of fans and followers around branded social media profiles. Without allocating enormous advertising budgets, messages from brands could be easily pushed through these communities and beyond, due to the networked nature of social media and the ability to pass messages in real time (Fournier and Avery 2011; Kaplan and Haenlein 2010).

Nowadays, we have learned that the unpaid distribution of social media content (i.e. organic reach) is not a known fact. As brand activities migrated to social media, the timelines of Facebook and Twitter became cluttered (Marshall 2014). In response to this development, social media such as Facebook, Twitter and recently Instagram introduced algorithms to function as gatekeepers. Messages only pop up on one’s timeline when these algorithms identify a message as potentially relevant for that person (Bucher 2012). As these algorithms attribute less relevance to brand messages than to messages from friends and family (Isaac and Ember 2016), only a small percentage of the brand’s fan base will have an opportunity to see its messages and share them with others. While Facebook restricted the organic reach of brand content to only 2% in 2014, it can be expected to head towards zero (Marshall 2014; Mosseri 2012). To open the gates of Facebook or Twitter’s algorithms, brands (again) have to pay. Or, to harness the potential of social media, brands need to offer the right content at the right moment, so that social media users are motivated to share with a larger audience through friends and followers.

So far, research has explored the effects of various content and audience characteristics on consumers’ sharing behaviour. This body of literature has enhanced our understanding of social media advertising (for a review, see: Knoll 2016). Posting content that sparks message sharing is important as it enables brands to obtain visibility for their messages and to enhance their persuasiveness. Peers are consistently rated the most trusted source of influence (Nielsen 2015). When peers share brand messages on their social media timelines, advertisers thus can create a halo effect of trust for their messages (cf. Boerman, Willemsen, and Van der Aa 2017). This explains why marketers rely on shares as a KPI for advertising success (Peters et al. 2013).

Although various audience and content characteristics have been examined to gain insight into the effectiveness of brand messages on social media, less is known about moment-related characteristics. As social media is defined by its real-time nature, what is the right moment to post content on social media in order to stimulate word of mouth? This study extends previous research by examining this question. More specifically, we examine whether brands can boost sharing behaviour when they align their messages with public events happening in the moment – a practice that is also known as real-time marketing (RTM). By doing so, we follow up on Weingarten and Berger 2017’s call (2017) to gain more insight into time’s role in influencing word of mouth.

Second, this study contributes to the literature by examining what moments are most valuable for RTM to generate message sharing. We propose and test two different types of moments to connect brand content with timely events: (1) predictable moments: moments that are known to happen, like holidays, seasonal happenings (e.g. start of summer), and public events that are announced far in advance (e.g. Super Bowl) and (2) unpredictable moments: moments that unexpectedly happen, like trending topics, social media content that obtains viral status (e.g. #thedress), and breaking news stories (e.g. the death of a well-known artist). Moreover, based on the literature, we examine whether unpredictable RTM messages differ from predictable RTM in terms of originality and meaningfulness, and subsequently, in their potential to stimulate sharing behaviour. As such, we make an initial attempt to shed light on the question of whether RTM is successful in enhancing word of mouth, and also why and under what circumstances.

Finally, this study examines what content-characteristics in RTM messages affect sharing behaviour. Under the premise that RTM is successful when making a link to publicly discussed events happening in the moment, we expect that RTM will stimulate word of mouth when it makes this link explicit by means of moment-driven hashtags or visuals. As such, with this study, we attempt to gain a deeper understanding of RTM, and the role of content characteristics to leverage its success.

RTM as a social media advertising strategy

More than 10 years ago, Kumar, Jacob, and Sriskandarajah (2006) pointed out the importance of timing to increasing the revenue of advertising. Posting time is known to be one of the main factors taken into account in the algorithm of social media like Twitter (Oremus 2017; Wachaspati 2014). However, only a handful of studies have examined timing as a driver of brand post success (for a review, see Sabate et al. 2014). For example, Cvijikj and Michahelles (2013) showed that posting during workdays increases word of mouth, especially when making use of so-called ‘low hours’ (between 4 am and 4 pm).

To gain momentum, practitioners do not only think in terms of timing but also in terms of moments. Brands increasingly link content with moments or events that are publicly discussed on social media in real time. Joining conversations by aligning or associating brand messages with timely events is also known as RTM. By injecting an RTM message in real-time conversations about public events, brands believe that they can free ride on the attention that these events garner from social media users, become part of the ongoing conversation, and obtain otherwise unattainable visibility for their brands (Kerns 2014; Scott 2011). The dozens of brands creating social media content about #thedress illustrates this strategy, such as Heineken who claimed that their beer was ‘definitely white and gold’. The message associated their product with a photo of a dress that had gone viral when people disagreed over its colours. Some saw black and blue while others saw white and gold (see Figure 1). The message was a sudden success, leading towards an uplift in the number of shares per follower (Kerns 2014).

Figure 1. RTM message linking beer brand Heineken with unpredictable moment #thedress.

Targeting moments that attract captive audiences is not a new strategy in itself. Even before the introduction of social media, brands linked traditional media messages with public events. A classic example is a Christmas ad that was created for whisky brand J&B in 1990; a Billboard ad that simply stated ‘ingle ells, ingle ells—The holidays aren’t the same without JB’. Already in 1997, Sutherland noted that such RTM messages create ‘[…] opportunities to hitch a ride and harness the brand to something that will help move it more effortlessly and drive its budget further’. Indeed, various studies demonstrate that it is beneficial for brands to target events that garner public attention, such as sports events, as these responses can ‘spill over’ to brand messages that are shown in connection to these events (Moorman, Neijens and Smit 2007; Moorman et al. 2012).

The introduction of social media, however, has sparked a renewed interest in RTM from practitioners. RTM messages are, by definition, focused on public and timely moments. As such, RTM capitalizes on the properties of social media – that is, their networked infrastructure and real-time nature – and their affordances to users. This is demonstrated by studies within the realm of Uses and Gratifications, showing that media use is driven by motivations such as desired experiences and ‘gratifications sought’ (for an overview, see Ruggiero 2000). According to Hermida et al. (2012), people make use of social media such as Facebook or Twitter to find out what is happening in the lives of others and the world. The results of a representative survey show that social media use is mostly motivated by the need to keep up with social or community events (mentioned by 76% of the respondents), news and views (71%), and the ability to obtain first-hand information about important happenings (63%). Similar findings were reported by Voorveld (2016) who showed that consumers mostly use social media to interact with others, and keep up with current events and developments.

These findings are not surprising, as social media are designed to elicit public conversations about what is happening in the moment (Park et al. 2017). When logging in, Facebook and Twitter prompt their users to share their current status with other people, by asking questions such as ‘What’s on your mind’ and ‘What’s happening’. This is reflected in online conversations, as holidays, sports events, or other public events are highly discussed on Facebook and Twitter (Kerns 2014; Scott 2011). Especially noteworthy in this respect are various studies that have been conducted on the Super Bowl. These studies show that the Super Bowl is highly discussed on social media, along with the brand messages that are shown during the sports event (Allagui and Breslow 2016; Nail 2007). For example, Spotts, Purvis, and Patnaik (2014) demonstrated that almost one third of spectators tweeted about Super Bowl advertising. Moreover, social conversations about the game, as well as advertisements during the game, were positively related to people’s inclination to promote brands to other people.

Altogether, these studies suggest that RTM is an effective strategy to increase the impact of brand messages in a social media context, as its focus on public conversations and up-to-date events seems to cater to the needs of social media users, thereby making brand messages more relevant. This, in turn, can fuel sharing behaviour, as research finds a positive link between perceived relevance and people’s willingness to share brand content (van Noort, Antheunis and van Reijmersdal 2012). In line with the arguments outlined above, we formulated the following hypothesis:

H1: Brand messages that make use of RTM elicit more sharing behaviour than messages that do not make use of RTM.

Moment characteristics: comparing predictable versus unpredictable moments

Social media offer a battery of new opportunities for brands to chime in with timely events. Some can be planned well ahead, while others require brands to be quick to seize the moment. The former strategy is often applied in response to predictable moments, and the latter in response to unpredictable moments.

The most important difference between predictable and unpredictable moments is that the predictable moments can be foreseen as they return on a regular basis (e.g. holidays, sports events), and/or are announced on so-called content calendars: lists of events and happenings that are anticipated to garner public attention so brands can plan the creation and publication process of RTM (Walters 2017). Examples include Christmas, Blue Monday, and the Super Bowl. Unpredictable moments are not announced on such content calendars, as these moments simply happen and cannot be predicted in advance (Kerns 2014). Examples include a surprisingly cold summer, and an unexpected ‘dab’ (dance move) by a world leader that suddenly become topics of discussion.

Brands show a growing interest in unpredictable moments, which are more feasible than ever to target. Social media allow brands to monitor public discussions, in real time, and to connect with social media users, in real time (Allagui and Breslow 2016). As a result, brands can take up trending topics, breaking news stories, or other unpredictable events that garner sudden attention (e.g. #thedress, #pokemongo, #RIPPrince), and use it to their advantage to obtain viral status. A classic example is the RTM message of Oreo, who chimed in with the power outage during the Super Bowl of 2013 by tweeting ‘Power out? No problem’, accompanied by a visual of an Oreo cookie in the dark with the text ‘You can still dunk in the dark’. More recently, LEGO, Nestlé and IKEA jumped on the bandwagon and released their own versions of the viral video #mannequinchallenge, in which people remain frozen in action while playing the song ‘Black Beatles’ by Rae Sremmurd.

In parallel to the growing interest in targeting unpredictable moments, the field is increasingly adopting a newsroom approach to marketing. This is a strategy that embraces core journalistic principles to produce relevant content in a shorter time frame (Lavecchia 2013). These two developments are not unrelated. One of these journalistic principles is to create content with news value to break through the clutter, gain public attention, and become the story of the day. Unpredictability is considered to provide such news value (for a review, see: Harcup and O’Neill 2001). This is supported by the journalism literature, showing that unpredictability, and related concepts such as unexpectedness, determine people’s decision to share particular content. Unpredictable events are believed to be more relevant or ‘meaningful to a large audience’ (Rudat and Buder 2015, 76) than predictable moments, and thus are more often covered in the news, and shared on social media (e.g. Galtung and Ruge 1965; Harcup and O’Neill 2001; Rudat and Buder 2015).

Although the value of unpredictable events has not been tested for RTM specifically, there is anecdotal evidence, showing that unpredictable events yield more word of mouth than predictable events. This is provided by Kerns in his book ‘Trendology’ (2014), in which he reports a retweet uplift of 400% when brands targeted predictable events versus an uplift of 1,200% when brands targeted unpredictable events. This finding is in line with the word of mouth literature, which shows that emotions that are triggered by unexpected events, such as surprise and arousal, drive sharing behaviour (Berger and Milkman 2012). Both emotions are characterized by an activated state of mind, which is known for enhancing consumers’ attention to brand messages, and for stimulating word of mouth (Berger and Milkman 2012; Derbaix and Vanhamme 2003; Schamari and Schaefers 2015).

In line with these studies, we formulated the following hypothesis:

H2: RTM messages that are associated with unpredictable moments will elicit more sharing behaviour than RTM messages that are associated with predictable moments.

Creative crafting of RTM messages

To seize the opportunities provided by unpredictable moments, brands need ‘to live in the moment’ and ‘actively shape [brand messages] in real-time’ (Mitchel 2014). How do time constraints, as imposed by unpredictable moments happening in real-time (vs. predictable moments), affect the creative crafting of RTM messages? Is it beneficial for the creative quality of content to target unforeseen moments and create content on the fly? These are important questions to address, as creativity is ‘thought to be one of the most important components of advertising effectiveness’ (Lehnert, Till, and Ospina 2014, 274). The advertising literature shows that messages are more effective in generating favourable consumer responses, including word of mouth, when these messages score highly on creativity (Moldovan and Lehman 2010), or its defining concepts (cf. Rudat and Buder 2015; Stathopoulou et al. 2014; Tafesse 2015).

To address these explorative questions, we compare predictable RTM with unpredictable RTM on originality and meaningfulness; two concepts that in tandem determine the creativity of a message (Ang, Lee and Leong 2007, 220).

Originality, sometimes also referred to as novelty or divergence, is the extent to which a message is unexpected and deviates from the norm (Ang and Low 2000). Along these lines, White, Shen, and Smith (2002) gauge originality by adjectives such as over-used-fresh and predictable-novel. In a similar vein, Kilgour and Koslow’s (2009) operationalize originality as original, different, and unexpected.

Meaningfulness, also referred to as appropriateness, is the extent to which an idea is on-strategy (i.e. right for the target) (Lehnert, Till, and Ospina 2014). This could mean different things for different advertising strategies, but in general, it ‘concerns whether the elements in an ad are relevant to the message conveyed and the intent of the ad’ (Ang and Low 2000, 836). In line with this conceptualization, White Shen, and Smith (2002) operationalize a meaningful ad as relevant (vs. irrelevant), logical (vs. illogical), and making sense (vs. senseless). A meaningful message is thus one that is relevant to the purpose and goal of the message. Meaningfulness is therefore considered to be a precondition or qualifier of creativity. Even when a brand message scores highly an originality, it will not be considered creative unless it is also meaningful (For a discussion see Lehnert et al. 2014; White et al. 2002).

Although RTM is considered to be a creative marketing technique, especially when brands target unpredictable moments (Kerns 2014), it remains unclear how originality and meaningfulness vary for the type of moment being targeted, and how this affects shaing behaviour

In line with the journalism literature, unpredictable moments (vs. predictable moments) offer brands the opportunities to resonate with the audience with more meaningful content. By chiming in with unpredictable moments, brands can create a halo effect of meaningfulness onto their messages (cf. Neijens, Smit, and Moorman 2009), as unpredictable moments have stronger news value and thus are considered to be more relevant or meaningful to a wide audience (Rudat and Buder 2015, 76). Yet, this does not necessarily make the RTM message itself meaningful. For RTM messages to be meaningful, it must also make a logical and relevant connection between the brand and the moment that is targeted.

The same holds for originality. Unpredictable moments enable brands to demonstrate their creative abilities. Unlike predictable moments that evolve around culturally familiar events, that are commonly targeted for RTM purposed (e.g. holidays, sports events), unpredictable moments often revolve around events or topics that suddenly become a topic of public interest. These events are predisposed to elicit associations previously unconnected to the brand. The deliberate combination of concepts that have never previously been linked, is a creativity technique that is found to trigger original ideas (Kilgour and Koslow 2009, 300). Yet, it remains unknown whether this will also result in more original RTM content.

Given the circumstances under which unpredictable RTM messages are created, unpredictability is more likely to hamper than to facilitate the creative quality of RTM messages. Unlike predictable moments that allow brands to schedule and plan the creation of content well in advance, unpredictable moments are subject to time constraints. Time constraints can negatively affect the originality of content as original ideas require idea incubation which is a time consuming endeavor (Moreau and Dahl 2005). Koslow et al. (2006) also found that time pressure is negatively related to creativity. They, however, do not specify for what dimensions of creativity this negative effect holds.

Given the exploratory nature of this study, we formulated research questions to explore the dynamics involved with the creative crafting of unpredictable and predictable RTM messages, and its outcome on message sharing:

RQ1: How, if at all, is moment predictability related to (a) the originality, and (b) the meaningfulness of RTM message?

RQ2: How is (a) the originality, and (b) the meaningfulness of RTM messages related to message sharing?

Content characteristics: moment-driven hashtags and visuals

Under the premise that RTM is successful by making a link to timely events, we expect that RTM will stimulate forwarding behaviour when it makes this link explicit. This can be done in at least two ways.

First, RTM can make use of moment-driven hashtags (cf. Kerns 2014), that is, hashtags that make a reference to timely events. Hashtags are implemented in social media to assign topics to posts, and, as such, make content searchable (Boyd, Golder, and Lotan 2010). As a result, content becomes accessible to a larger public of social media users than those that belong to a brand’s fan base. Thus, hashtags function as ‘conversational tagging’ (Huang, Lin, and Lin 2009), thereby increasing ‘the ability to find what other people are talking about in real-time’ (Boyd, Golder, and Lotan 2010).

So far, the literature reports ambiguous results for hashtag effectiveness. While some studies reported a significant and positive relation between hashtag presence and sharing (Suh et al. 2010), others did not (Araujo, Neijens, and Vliegenthart 2015). Hashtags, however, may have a stronger effect on sharing behaviour when inserted in RTM messages, as in these circumstances, there is greater need for conversational tagging. RTM messages are only able to free ride on the attention that public events garner, when labelled as a topic people are talking about in real-time. Thus, we expect the following:

H3: The presence of moment-driven hashtags in RTM messages (vs. absence) yields more shares.

Second, RTM messages can make use of moment-driven visuals (cf. Kerns 2014). When RTM messages contain moment-driven visuals, the message is accompanied by material that makes some kind of visual connection to the public event. For example, in the visual depicted by Figure 2, the World Wildlife Fund makes a clear connection with Pokémon GO, a location-based augmented reality game that became a sudden hit in summer 2016 (i.e. an unpredictable moment). The visual shows an endangered animal, held under shot by a ‘Poké Ball’, together with the text ‘Don’t catch ’em’ which is a direct reference to Pokémon’s payoff. As such, the moment becomes integrated in the visualization of the brand message (see Figure 2).

Figure 2. Example of an RTM message that contains a moment-driven visual and moment-driven hashtag.

From the literature on word of mouth, it is known that visuals positively affect sharing behaviour (Araujo, Neijens, and Vliegenthart 2015; Cvijikj and Michahelles 2013; Sabate et al. 2014; Tafesse 2015). Visual imagery makes brand messages ‘vivid’. Vividness involves the representational richness of media (content) and can be defined as ‘the ability of content to depict a situation in ways that approximate reality’ (Tafesse 2015, 930). Thus, textual content is perceived less vivid than visual content (for a review, see Tafesse 2015).

The presence of visuals may make RTM messages more vivid, but the presence of moment-driven visuals allows RTM messages to make a vivid connection with the public. By integrating the moment into the visual, the message makes a connection with events that, at that moment, are happening in people’s lives. This is supported by research showing that brand messages need to connect with the audience to be effective. Connectedness is a strong driver of a variety of desirable consumer responses, including recall, positive attitudes, and warm feelings towards the ad (Ang, Lee, and Leong 2007). If a message does not fit with what is going on in the lives of consumers, ads are likely to be rejected or discounted. Thus, based on the literature, we propose:

H4: The presence of moment-driven visuals in RTM messages (vs. absence) yields more shares.

Method

To test the hypotheses and the research questions, we performed a content analysis of brand messages, as posted on Twitter by the top-100 Dutch advertisers, compiled by Nielsen in 2016. Twitter was chosen as a research context for this study, given its focus on real-time content production and content sharing.

We selected brands from Nielsen’s top-100 advertisers as this ranking (1) covers brands that score highly on gross media spending to minimize variations in brand familiarity, and (2) covers brands from 10 different market segments to increase the generalizability of the results (cf. Araujo, Neijens, and Vliegenthart 2015). We randomly selected three brands per market segment, resulting in 30 brands in total. For each of these brands, we collected the official Twitter profiles that were used for marketing communication purposes. Brand messages were obtained by using a social media monitor, OBI4wan, which collected all tweets from the selected brands as posted between 1 June 2015 and 1 December 2016. To ensure equal group sizes, tweets were subjected to a stratified random sampling method, with brand name as stratum. This procedure resulted in a sample of 1500 unique tweets equally distributed over the brands (n = 50 per brand). Of these tweets, 97 were no longer accessible for coding, as they had been removed by their profile owners. Also, five brand messages were targeted at specific individuals, rather than the general public. After removing these tweets, the final sample consisted of 1,398 tweets.

Procedure

The data was manually coded by two coders. The coders were trained over the course of two weeks to apply a coding instrument that was developed based on literature and pilot tests. The coding instrument included instructions to identify the presence of RTM techniques in brand Twitter posts. As part of their training, the coders double-coded a subsample of tweets to determine inter-coder reliability (10% of the sample).

Measures

Sharing behaviour

Sharing behaviour was operationalized as the number of retweets that each brand tweet obtained. To guarantee a normal distribution of the residuals, we used natural logarithms, being calculated as LN (retweets +1) (cf. Sabate et al. 2014) (M = 3.63, SD = 14.87; min =0; max = 308).

Presence of RTM

RTM messages are brand messages that are associated with public events that are temporary in nature (cf. Kerns 2014; Moorman, Neijens, and Smit 2007). For brand messages to be categorized as RTM, they had to meet two criteria. First, the coders determined whether a brand message was aligned or associated with a temporary event, happening, trend or moment (0 = ‘no’; 1 = ‘yes’). In case this question was answered with ‘yes’, the coders additionally determined whether this event, happening, or trend was public in nature (0 = ‘no’; 1 = ‘yes’). Examples include holidays, (inter)national sports games, elections, news stories targeted at or discussed by a general audience, or popular topics such as those identified as ‘trending’ on Twitter (Average Krippendorf’s α = .85). Analyses showed that RTM was a common strategy; 18.7% of all brand messages were associated with public and timely events.

Unpredictability of the moment

In case a brand message could be identified as RTM, the coders coded whether the moment was predictable (0), or unpredictable (1). Predictable moments are events that return on a regular basis and/or are expected to take place. These events are often announced in advance. Examples include public holidays, season related events, and public events that are usually mentioned on a content calendar. Unpredictable moments are events that simply happen and could not be predicted in advance (Rudat and Buder 2015). Examples include breaking news stories, trending topics, or other unexpected events; thus content that could not be planned in advance such as #thedress or Pokémon GO (Krippendorf’s α = .84). Of all RTM messages, 77.1% could be marked as an RTM message in response to a predictable moment, and 22.9% could be marked as an RTM message in response to an unpredictable moment.

Originality

To assess originality, we used the Creative Product Semantic Differential Scale (CPSS) (White, Shen, and Smith 2002)¹. On a 5-point scale, the coders judged the originality of each RTM message, based on three bipolar adjectives: over-used-fresh, ordinary-unique, and expected-novel (Krippendorf’s α = .81, M = 3.29, SD = .87).

Meaningfulness

Based on White, Shen, and Smith CPSS instrument (2002), the coders determined on a 5-point scale to what degree the brand connected meaningfully with the moment in the RTM message, based on three bipolar adjectives: illogical-logical, senseless-made sense, weakly connected-strongly connected (Krippendorf’s α = .93, M = 2.97, SD = 1.38)¹.

Presence of moment-driven hashtag

The coders determined whether the RTM messages were linked to timely moment through moment-driven hashtags (0 = ‘no’; 1 = ‘yes’) that is: hashtags that make a direct or indirect reference to the moment (e.g. #Christmas; #XMass #Santa; Krippendorf’s α = 1.00). A moment-driven hashtag was present in 60.8% of the RTM messages.

Presence of moment-driven visual

The coders determined whether timely moments were visually depicted and integrated in the brand message (0 = ‘no’; 1 = ‘yes’; Krippendorf’s α = .93); that is: visual material that makes some kind of visual connection to a public event (e.g. a visual in which a Christmas tree is depicted). A moment-driven visual was present in 78.4% of the RTM messages.

Covariates

RTM in brand tweets often comes with visual imagery such as photos or videos. The presence of such material is found to be a strong predictor of sharing behaviour, and was therefore included as a covariate (cf. Araujo, Neijens, and Vliegenthart 2015). For similar reasons, we controlled for the presence of hashtags (i.e. hashtags that do not link a brand tweet to specific events or moments) (Suh et al. 2010).

Results

To test the hypotheses, we employed multilevel regression analyses, using a random intercept-fixed slope model. Allowing random effects on intercepts is recommended to reduce the variance caused by variables on the second level (Cohen et al. 2013), which in this study was the brand level. This procedure was selected as certain brands may prompt more retweets than others because of differences in popularity or the size of the fan base. Controlling for such brand-level effects is desirable. An intercept-only model gives an intraclass correlation of .24 (also indicated as rho), meaning that 24% of the variation in sharing behaviour can be explained by differences between brands.

The first multi-level regression analysis aimed to test the value of RTM, and included the presence of RTM as an independent variable, sharing behaviour as a dependent variable, and the presence of visuals as a control variable. The results provided support for H1. Brand messages that make use of RTM elicit more sharing behaviour than brand messages that do not make use of RTM (b = .18, SE = .06, p = .002).

The second multilevel model that was conducted aimed to test the relation between moment- and content-related characteristics on one side and sharing behaviour on the other side. In this multilevel regression model, sharing behaviour was included as a dependent variable, moment unpredictability, presence of moment-driven hashtags, and presence of moment-driven visuals were included as independent variables. Furthermore, the presence of hashtags and visuals were included as covariates. This allowed us to examine the contribution of moment-driven hashtags and moment-driven visuals, above and beyond the presence of hashtags and visuals in general. The results are reported in Table 1.

Table 1. Multilevel regression analysis examining the effects of moment- and content characteristics on the sharing of RTM messages.

	b	SE B	p
Fixed effects
Intercept	.62	.20	.003***
Visual	.07	.19	.711
Hashtags	.12	.22	.581
Unpredictable versus predictable moment	.42	.18	.025*
Moment-driven hashtag	.20	.20	.327
Moment-driven visual	.44	.17	.010**
Random parameters
Variance of intercept	.28	.12	.018*
Variance of residual	.91	.08	.000***
−2 Restricted log likelihood	750.48

Note. *p ≤ .05, **p ≤ .01, *** p ≤ .001

The results provided support for H2, which predicted that RTM messages associated with unpredictable moments elicit more sharing behaviour than RTM messages associated with predictable moments, as demonstrated by a positive and significant relation between moment unpredictability and sharing behaviour (b = .42, SE = .18, p = .03).

The aim of RQ1 and RQ2 was to explore how, if at all, moment predictability is related to originality and meaningfulness, and subsequently, message sharing. For this purpose, we performed a mediation analysis using PROCESS macro (Hayes 2013, model 4), with 5,000 bootstrap samples to estimate the 95% bias-corrected bootstrap confidence intervals (BCIs) for inferences regarding indirect effects. Unpredictable RTM (vs. predictable RTM) functioned as the independent variable, originality and meaningfulness as the mediator, and sharing behaviour as the dependent variable. The presence of visual imagery was included as a covariate.

As demonstrated by Figure 3, unpredictable RTM yields more shares than predictable RTM, as indicated by a positive and significant direct relation between moment unpredictability and sharing behaviour (b = .36, p = .05). As such, the results replicate the previous findings as reported for H2.

Figure 3. Mediation model: Effect of moment unpredictability (unpredictable vs. predictable RTM) on sharing behaviour via originality and meaningfulness. Only for RTM messages (n = 258).

Note. *p ≤ .05, **p ≤ .01, *** p ≤ .001

Furthermore, the results show that this effect is mediated by the two dimensions of creativity. Interestingly, unpredictable RTM (vs. predictable RTM) was negatively related to originality (b = –.34, p = .01; RQ1a), which, in turn was positively related to sharing behaviour (b = .25, p = .00; RQ2a). Indirect effects confirmed that the negative indirect path between unpredictable RTM (vs. predictable) and sharing behaviour through originality was significant (indirect effect = –.09, boot SE = .05, BCI [–.21, –.02]. Moreover, unpredictable RTM (vs. predictable RTM) was positively related to meaningfulness (b = 1.46, p < .00; RQ1b), which, in turn, was positively related to sharing behaviour (b = .12, p = .03; RQ2b). This a positive indirect path between unpredictable RTM (vs. predictable) and sharing behaviour through meaningfulness was significant (indirect effect = .18, boot SE = .09, BCI [.02, .35]. Examination of the pairwise contrasts of the indirect effects further shows that the positive indirect effect through meaningfulness was statistically larger than the negative indirect effect through originality, with BCA 95% CI of .08–.47.

Finally, H3 and H4 predicted that moment-driven hashtags and moment-driven visuals drive sharing behaviour. As demonstrated by the multilevel regression analyses reported in Table 1, we found no significant relation between moment-driven hashtag presence and sharing behaviour (b = .20, SE = .20, p = .32). H3, stating that the presence of moment-driven hashtags in predictable RTM messages (vs. absence) yields more sharing behaviour, was thus rejected. We did, however, find support for H4. In line with our expectations, RTM messages containing moment-driven visuals yielded more shares than RTM messages containing no moment-driven visuals (b = .44, SE = .17, p = .01).

Discussion

Social media timelines are increasingly populated with brand messages that are linked to timely events; a practice that is also known as RTM. By linking their messages to public and real-time moments, brands are becoming part of ongoing discussions, thereby hoping to engage consumers while promoting their brand at the same time (Kerns 2014; Scott 2011). The underlying rationale is that this allows brands to fuel word of mouth, and to break through the gates of social media algorithms to enhance reach and persuasion.

Although RTM is believed to be an effective strategy to boost word of mouth, no research has been conducted to test its effectiveness. The aim of this study was, therefore, to shed more light on the value of RTM. More specifically, it sought to gain more insight into the question whether RTM can be an effective strategy to boost sharing behaviour, and if so, what moment- and content-related characteristics are likely to contribute to its success. To address this aim, we performed a content analysis of brand tweets from Nielsen’s top-100 advertisers (n = 1,500).

Findings and implications

The findings contribute to the literature in four important ways. First, we found that brand messages that make use of RTM elicit more shares than messages that do not make use of RTM (H1). A theoretical implication of this finding is that the success of brand messages is determined by time; a factor that has been largely neglected in prior research. An exception is the work of Weingarten and Berger (2017), who examined the effects of temporal location – that is, whether something happens in the future or the past – on word of mouth. The authors find that future events evoke more sharing than past events, as they elicit more arousal. The present study extends the work of Weingarten and Berger, as it shows that something as simple as a reference to current happenings (vs. no reference) also increases word of mouth.

A second contribution lies in the identification of moment-related characteristics as drivers of RTM success. The results show that RTM is an even more effective strategy when brand messages are linked with unpredictable events (vs. predictable events). This finding corroborates the journalism literature, that has a long research tradition in examining what content gets shared (e.g. Galtung and Ruge 1965; Harcup and O’Neill 2001; Rudat and Buder 2015). In line with this stream of research, we found that RTM messages are more retweeted when targeting unpredictable events than when targeting predictable events. This suggests that unpredictability does not only determine the relevance of content created by journalists and publishers, but also the relevance of content created by brands. This finding attests to the idea that, in today’s media landscape, brands should think more as journalists or publishers (Lieb 2011). As consumers grow more sceptical of commercial messages, brands are urged to publish newsworthy content that is worth sharing by itself (cf. Boerman, Willemsen, and Van Der Aa 2017).

Third, the current study sheds light on the effects of moment-related characteristics and its underlying mechanisms. Shares increase as a function of meaningfulness and originality. These dimensions of creativity, however, are differently affected by predictable and unpredictable moments, thereby leading to opposing indirect effects. Unpredictable RTM yielded less original content, and hence less shares (RQ1a and RQ2a). These findings can be explained, by the circumstances under which unpredictable RTM messages are created. RTM messages targeting unpredictable moments cannot be planned in advance and, hence, are created on an impromptu basis under tight deadlines. Indeed, prior research shows that constraining the time available to complete a given task leads to less original ideas, as original thinking requires time (Moreau and Dahl 2005). Similar findings are reported in a study on advertising creativity (Koslow et al. 2009), although this study does not specify for what dimension(s) of creativity this negative effect holds. The present study extends this work by showing that time constraints, as imposed by unpredictable moments, negatively affected originality but not meaningfulness.

In contrast, unpredictable RTM yielded more meaningful content (RQ1b) and, hence, more shares (RQ2b). This positive indirect effect through meaningfulness was stronger than the negative indirect effect through originality. As a result, the net effect of moment unpredictability on sharing behaviour was still positive. The positive indirect effect through meaningfulness is in contrast with the originality bias hypothesis, which posits that creatives have a stronger focus on creating original content than meaningful content (Kilgour et al. 2012). This bias is stimulated by industry awards, that weigh originality more heavily as a criterion for creativity than meaningfulness. The present study adds to the literature by showing that the presence of such a bias is context dependent. When engaging in unpredictable RTM on social media, brands seem to favour meaningfulness over originality.

Finally, it was shown that the presence of moment-driven visuals drive sharing behaviour, above and beyond the presence of visuals in general (H4). This finding contributes to the literature by challenging the finding that visual materials, such as photos and videos, by themselves are sufficient to drive word of mouth. It is well-known that the presence of visual imagery can enhance the sharing of brand message (Sabate et al. 2014; Tafesse 2015). Less is known about the various strategies that can be used by brands to make visual content more appealing. This is an important gap in the literature, considering the fact that more and more marketers engage in visual storytelling (Yuki 2015). The present study fills this gap, by showing that brands can benefit from the presence of moment-driven visuals, that is: content that makes a visual connection to a public event happening in the moment.

Against expectations, the presence of moment-driven hashtags did not lead to more shares (H3). While this finding is consistent with some studies reporting no significant effects for hashtag presence (Araujo, Neijens, and Vliegenthart 2015), it is inconsistent with others that did find a positive effect (Suh et al. 2010). This suggests that hashtags may not always serve as searchable content (Araujo, Neijens, and Vliegenthart 2015; Boyed, Golder, and Lotan 2010), but more as signals that mark experiential topics (Zappavigna 2015). This seems plausible for RTM messages, as moment-driven hashtags make reference to events people are experiencing at that particular moment.

Limitations and directions for further research

Although this study provides valuable new insights into the effects of RTM, there are some limitations that provide directions for future research. First, this study adopted a naturalistic study design. Natural settings are considered suitable, especially in those circumstances in which context-induced attention is expected to spill-over to brand messages (for a discussion, see Moorman, Neijens, and Smit 2007; Moorman et al. 2012). In cluttered media landscapes, brand messages have to compete for attention. Brands, therefore, can benefit from linking their messages with an engaging context that people willingly pay attention to. Such processes are difficult to capture in experimental settings, as such settings often come with forced media exposure.

However, natural settings do not allow the same internal control as experimental settings. Although this study applied multilevel regression modelling to control for brand effects, it is possible that other types of variables affected the relationships found. For example, it remains unclear whether all consumers are equally susceptible to RTM effects. As mentioned earlier, when engaging in word of mouth, people may be driven by different types of motives. While some are motivated by social bonding, others are motivated by impression management concerns. Prior research shows that people are more likely to share information about future events (vs. past events), except when sharing would make people look bad (Weingarten and Berger 2013). Further research is thus recommended to examine the interplay of moment- and audience-related characteristics.

Further research is also recommended to examine how RTM affects various forms of word of mouth. Social media users can engage in word of mouth by sharing a brand post, but also by ‘liking’ it. When users ‘like’ post, they indicate an interest in the brand post. When users share a brand post, they want others to see the brand post as well. Hence, they represent different forms of engagement. This is why social media algorithms give shares the most weight, and likes the least (cf. Cvijikj and Michahelles 2013; Facebook, n.d.). As this study was focused on the question how brands can break through the gates of social media algorithms, it concentrated on the former metric. Future research is needed to test whether similar effects can be reported for liking.

Finally, this study did not control for the exact time a brand message was posted on Twitter. When it comes to RTM, timing seems crucial. According to practitioners, the best moment to post a RTM message is right before an event is picked up by a broad public (Kerns 2014; Scott 2011). When brands do not jump on the bandwagon quickly enough, they may miss the opportunity to free ride on the attention that public events garner. What is trending at one particular moment may be old news on the next. Future research is thus recommended, to examine posting time as a factor that could affect the life cycle of RTM messages.

Despite these limitations, this study provides several managerially relevant insights. Most importantly, it provides evidence-based support for RTM as an advertising strategy to boost shareability of brand messages on social media, especially when it targets unpredictable moments. A valuable implication deriving from this conclusion is that advertisers should shift their focus from planning content in advance towards being responsive to what is happening in real time, and making meaningful ad hoc content, without compromising on its originality. Creativity will be more important than ever, as well as the ability to truly listen to one’s audience. This involves monitoring conversations on social media, discovering topics people actually care about, and using these insights to actively shape content in real time. This study shows that RTM is real business, and provides actionable insights on how to effectively make use of this advertising strategy. This is imperative, as marketing is predicted to become more and more unpredictable in the future (Kerns 2014).

Note on contributors

Lotte M. Willemsen (PhD), is professor in Crossmedia Business at HU University of Applied Sciences, Utrecht. As per September 2019, she will work as a professor at Rotterdam University of Applied Science, where she will hold the chair ?Communication in a Networked Society?. Her recent research interests focuses on digital transformation, and what this means for brands and consumers, as well as the communication between them. Her work has been recognized with awards and grants from the European Advertising Academy (EAA), the Marketing Science Institute (MSI), and the Dutch National Science Foundation (NWO).

Komala Mazerant (MSc.) is a PhD. candidate and senior lecturer at the Rotterdam University of Applied Sciences. She works at the department of Communication Studies where she lectures, coordinates and conducts research on content marketing and influencer marketing. Her focus in research is on the effectiveness of branded content on social media.

Anne-Lise Kamphuis (MSc.) is a self-employed (marketing) researcher, currently working as a research fellow at the Cross media Business lab, HU University of Applied Sciences, Utrecht. In addition, she works at Odion, as a researcher and policy consultant. Over the course of her career, she worked on various research projects, qualitative and quantitative, for a wide range of organizations, including Synovate (now Ipsos). Although diverse in scope, they all focus on gaining a better understanding of the customer. She is especially interested in consumer insights, privacy, and social media marketing.

Gerrita van der Veen (PhD.) is professor in Marketing, Market Research & Innovation, and managing director of the Research Centre for Innovation in Business and Communication at HU University of Applied Sciences, Utrecht. Her topics of interest are Brands & Communication. Gerrita was trained as a social psychologist. She has previously worked in various research disciplines and positions at ? among others - Synovate (and its predecessors), SWOKA, the Institute for Strategic Consumer Research and VU University Amsterdam.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research is funded by a SIA RAAK grant from the Dutch National Science Foundation, awarded to Gerrita van der Veen.

Notes

1 The literature on advertising creativity measurements takes two different directions. The first questions whether creativity can be systematically assessed. This school of thought assesses ad creativity indirectly either by examining how an ad is perceived by a target group (professionals, audience), by examining the creative styles used in an ad, or by examining proxy’s such as industry awards (Ashley and Tuten 2014; Kilgour et al. 2012). The second relies on the product creativity literature that considers creativity as an identifiable and measurable concept. This school of thought acknowledges that evaluations of creativity can differ from person to person, yet at the same time, demonstrate systematic patterns that can be reliably assessed (Mercanti-Guérin 2008; White et al. 2006). Indeed studies demonstrate that human coders can reach agreement over what is considered creative, original or meaningful (e.g. Kilgour and Koslow 2005). The current study belongs to the second school and applies the Creative Product Semantic Scale to gauge originality and meaningfulness. This scale was selected as it yielded similar scores on these dimensions of creativity amongst various professionals, college students, and the general public (White, Shen, and Smith 2002). When using this measure, White, Shen, and Smith found no significant differences in the judgments of the three groups regarding the originality and meaningfulness of the ads.