Karen Nelson-Field and Jennifer Taylor
Ehrenberg-Bass Institute for Marketing Science
As brands invest in building their Facebook profiles, rather than assuming that fans 'liking' them results in meaningful engagement, marketers should consider whether this level of interaction is a genuine measure of success.
If you enjoy this article from Admap, find out more about subscribing to Admap and Warc. This article may be shared or reproduced online, provided the contents are not altered and the source is acknowledged as:
Reproduced from Admap with permission. © Copyright Warc. www.warc.com/admap
'Engage or die' is the marketing catchphrase of the social media era, with Facebook the primary beneficiary of the shift in marketing thinking. Marketers are flocking to the Facebook platform (largely at the expense of other traditional media types) based on its touted ability to engage and support deep relationships with consumers. It is said that such relationships will lead to greater consumer loyalty. But with all the hype surrounding engagement and relationship building does Facebook really allow for any more genuine interaction with consumers than traditional forms of media?
The Advertising Research Foundation acknowledges there is no single theory that directly explains how engagement works, but discusses attention, visual perception, visual mental imagery, memory, emotions and social relations as elements that may contribute to a theory (Plummer, 2006). There are many competing definitions of engagement. Terms like 'involvement', 'experience' and 'connection' are used interchangeably, without agreement about how or whether they differ in meaning (Calder et al, 2009) and appropriate measurement.
Broadly, however, definitions take one of two forms: behavioural or attitudinal engagement. With the relatively recent rise of social media, there is renewed vigour in the search to understand what engagement is, how to measure it and how engagement can be beneficial for advertising. The lack of clear definition and application of the construct presents its own problems but if we work with the premise that in social media engagement takes the form of activity like posting a comment or sharing a story (or the attitude that underlies the subsequent interaction with the brand) how much engagement do we see?
One measure used to put the construct of engagement into operation is Facebook's new 'People Talking About This' number (PTAT). Introduced by Facebook in October 2011, the metric is said to measure fan engagement by counting 'stories' which are considered to be any type of direct interaction with a fan page such as: initial liking; liking specific content on a page, posting to a wall; commenting; sharing a post or other content from the page; answering a question; photo tagging; check-ins or RsVPing to an event. The PTAT number is publicly available on any brand page, sitting below the numbers of brand fans; fans who are likened to an army of brand advocates ready to engage on a moment's notice (Ed Keller, 2012).
But how many of these fans, after the initial 'like', are actually bothering to play their part in the proposed brand/consumer marriage? Are fans really engaged with the brands they choose at some point to 'like'? given the PTAT metric includes initial likes, we sought to understand the degree to which the PTAT number reflected natural fan growth compared with the degree to which fans actually interact with the brand after their first tryst.
On a weekly basis, over a six-week period between October and November 2011, we collected both the PTAT and overall fan numbers for the top 200 brands (by fan numbers) on Facebook. We then averaged each metric over this period to give us an 'average weekly PTAT' and an 'average weekly fan number' for each brand. This allowed us to subsequently calculate the proportion of PTAT to fan numbers. Finally for each brand we also calculated average weekly fan growth (%). It should be noted that we calculated an average week to iron out any timing effects given that the PTAT number (while updated daily) tracks unique users interacting with the brand in the trailing seven-day period, while likes are counted and displayed in real time.
First we considered the relationship between the average weekly fan growth and the PTAT proportion to total fans. The strong positive correlation (r = 0.93, p < .001) suggests that, in any given week, fan growth explains the largest amount of variation in the PTAT number. This suggests that the bulk of 'talking' is one-off 'joining' – a far cry from talking about, and engaging with, a brand. When we extract average weekly fan growth from the PTAT number it reveals the degree of interaction (after becoming a fan) – which is the core of engagement.
In any given week, less than 0.5% of Facebook fans engage with the brand they are fans of
Of the 200 brands studied, only one showed a level of engagement over 2%, and only 10% of brands reached the 1% mark. This number may even be conservative. The ambiguous nature of the PTAT number is such that the metric tracks unique user visits interacting with a page over a seven-day period, however no information regarding the distribution of unique users is revealed. Administrators cannot tell, week to week, whether many unique users interacted once or whether many interactions were from the same unique user. given that we know for Twitter "a very, very small group of 'elite' Twitter users (0.05%) is producing half of the content that gets consumed" (Econsultancy, March 2011), we think the latter likely.
It should also be noted that while the differences between these numbers are statistically significant, this doesn't really mean much. Due to the size of the samples (the average fan base size was approximately five million) even small differences will be rendered statistically significant. What is more important here is the managerial significance of the tiny rate of engagement across all brands in a big sample (Figure 1).
Figure 1: Engagement by product category type
Passion brands don't differ – Facebook fans are no more or less engaged
A common marketing mantra is that passion brands are different. Often a term incorrectly interchanged with niche, passion brands are those that reportedly transcend the typical consumer/brand relationship and demand the highest levels of love and respect from their consumers. These brands are considered to be inspirational, iconic and tap into dreams. We identified 11 brands from the sample that would likely be considered passion brands; however no notable differences in engagement were evident (Figure 2).
Figure 2: Engagement by passion brands
Brand popularity doesn't affect the degree of fan engagement
We also sought to understand whether any differences might be evident in the highest-ranking, versus the lowest-ranking brands: whether smaller brands (by Facebook ranking standards) attracted more or less interaction from fans than bigger brands. Our results show that the average engagement rate for the top ten brands on Facebook (ranging from Coke to Converse to Starbucks) was 0.36. The average engagement rate for the lowest ranking (of the Top 200) was 0.83 (ranging from Jeep to Tiffany & Co to Applebee's).
Altogether these findings suggest that few fans engage with a brand on Facebook on an on-going basis, regardless of category or brand status. While this may be dispiriting news for those brands with engagement-based objectives, the results are not altogether surprising given what we know about buyer behaviour and the reasonable questions that can be asked about the metrics of engagement.
The results, showing very limited levels of engagement, are predictable to those familiar with empirical patterns of buyer behaviour. Consumers don't love their brands; rather, they are polygamously loyal to a small group of offerings in the category (Goodhardt et al, 1984). In effect, we are loyal switchers rather than brand loyals (Uncles et al, 1995, Ehrenberg et al, 2004). If several brands in a category satisfy us then why should we be in a committed relationship with any one?
Different theoretical positions about the way advertising works can influence practitioners and academics to have a particular belief about the importance of high engagement. Advertising can be seen as either a strong or a weak force; the effects of it either being powerful – with the ability to change attitudes, manipulate customers and hence grow markets – or weak, with the power to grow markets by reminding consumers and nudging customers' propensity to buy (Jones 1990).
Despite the criticism that there is little evidence to support the idea that advertising is persuasive – that advertising leads to attitude change or that attitude has a direct impact on behaviour (Leighton 2009), the strong view is not an easy one to dismiss. It was often accepted by default rather than endorsement, though many agencies and advertisers now support the weak view of advertising (Jones 1990).
But, if advertising simply works by reminding people of the brand, leading to it "coming to mind, being familiar, safe, and satisficing (that is, being 'good enough')" (Ehrenberg et al, 2002), there may be little gain in doing anything more than reminding them of the brand. When focusing on achieving high levels of engagement we should question whether we are still trying to persuade consumers, even if our view of how advertising works is no longer aligned with this aim.
Without any clear definition, the term engagement has become commonplace in marketing practice, with many believing that engagement is of key importance in evoking a response (attitudinal or behavioural) from consumers, assuming it ultimately leads to purchases and brand loyalty.
At the very core of the social media mantra is the premise that brands need to engage their customers in order to grow but there is only a tenuous link between the effects of engagement and subsequent sales. Even if these top 200 brands achieved ten times their current level of engagement, what that ultimately means for the brand is uncertain. The push for engagement fails to explain what return, in real terms, a brand achieves by having highly engaging ads, on highly engaging vehicles or media. Behavioural engagement measures such as time spent looking at/reading/listening, sharing or repeated use of a media or vehicles tell us little about the end result that advertisers seek – sales. As McDonald (1995) points out, rarely would it be disputed that sales is the inevitable objective of 'for profit' organisational success; the final measure of advertising impact should be whether the viewer has a heightened propensity to purchase the product.
Overall, the results demonstrating low levels of interaction with brands, knowledge of repertoire rather than loyal brand buying behaviour and the fundamental uncertainty about what engagement is, or achieves, suggest that brands need to proceed with caution. They should focus instead on measurement of the sales effects of exposure to advertising with social media, tempering the expectation of finding high engagement that leads to sales.
Having a conversation with a brand seems not to be of top priority for consumers. This may come as a shock to Facebook disciples as it flies in the face of the social media buzz. But knowing that consumers aren't the marrying kind, more likely to date a small selection of brands than enter a serious relationship with just one, provides one of the reasons we should lower our expectations of brand engagement on social media. Before rushing to make more engaging social media content to seduce consumers to engage, we should also question whether it makes sense to focus on the ill-defined construct of engagement as a measure of success in the first place, when we ultimately want to know whether social media presents a better opportunity for sales, generated by our marketing dollar, than other media.
While some may look at the results of this study and point out that Brand A (at 1%) gets twice as much engagement as Brand B (at 0.5%), that's like saying 'you have 50 cents, I have one dollar, so I am twice as rich as you'. One dollar does not make you rich. The real question is how cost effective is it for a brand to attempt to drive engagement if the most they can reasonably expect is that 1% of fans will engage? And if they achieve higher engagement, will that translate into sales?
About the authors
Dr Karen Nelson-Field is senior research associate for the Ehrenberg-Bass Institute for Marketing Science, University of South Australia. Her research area focuses on social media and, in particular, how established marketing laws hold in the context of new media.
Dr Jennifer Taylor is senior research associate and lecturer at the Ehrenberg-Bass Institute for Marketing Science, University of South Australia. Her research focus is advertising, in particular, measuring the sales effects of advertising using single-source data.