Careful, Companies! Pay Attention to Your Tone of Voice, You Are not SalesmenFIRM GENERATED CONTENT IS A PRECURSOR TO FUTURE PROFITABILITY. BUT RESEARCHERS SAY TO BEWARE OF SOUNDING TOO POSITIVE, AND USE A NEUTRAL APPROACH. IN THE END, IT'S WHO YOU ARE IN REALITY THAT COUNTS
by Anatoli Colicev, Bocconi Department of Marketing
Social media marketing spending by American companies has hit a staggering $13.7 billion in 2017. Yet, in a recent survey of chief marketing officers (CMOs) conducted by Duke University, only one in six CMOs agreed that social media marketing has a quantitatively proven impact on business performance. The disconnect between social media marketing expenditures and their perceived effectiveness is alarming. Thus, it becomes imperative to investigate how and under which conditions social media marketing can lead to better performance. While this can be addressed by analyzing the effect on sales, a more long-term oriented approach is to investigate how social media affects customer metrics that eventually translate into profitability. One key challenge is that consumers mostly use social media to connect to each other, making it yet unclear whether active marketing through this media helps or hurts customer metrics.
In a study with Ashish Kumar and Peter O’Connor, we address these issues by studying how firm generated content (FGC) and user generated content (UGC) affect brand awareness, consideration, purchase intention, and customer satisfaction. For this research, we used a sample of 19 US brands and analyzed their daily social media engagement on Facebook over a two-year span. We collected daily volume, valence and richness (e.g. image, video) of social media buzz associated with the brand. The volume of buzz is given by the total amount of user and firm generated conversations around the brand and is relatively straightforward to measure. For example, brands can directly fetch the number of likes, comments, shares and overall impression from Facebook Insights. However, measuring the valence of social media posts is more challenging. Given the vast amount of social media data, brand managers cannot read every single comment, post and re-share. A more pragmatic approach, which we employ and recommend, is to use machine learning algorithms that allow us to automatically and systematically extract the tonality and other features of posts from a vast amount of textual data. We then related social media measures to consumer metrics collected from a 5 million panel of customers from YouGov group which is a marketing research company specialized in survey data.
Using the estimates from our model, we find that user generated content has a strong relationship with brand awareness. In other words, due to higher virality and reach, user generated content is the most effective way of spreading brand identifiers and increasing awareness. We also find that firm generated content is more effective in driving consideration and purchase intent. Thus, firm generated content seems to be effective in driving future profitability. However, it’s not that simple. Indeed, in order for firm generated content to be effective, it should not have a strong positive tone. Why? A positive tone can be perceived as being overly commercial and pushy. It implies that one of the hallmarks of marketing communications in traditional settings - adopting positive language – may not be effective in the social media context. So, what firms should do? They should use more objective, neutral tone in their social media posts and try to employ more videos and images.
Importantly, we also find that brands with good corporate reputation have more leverage in terms of how they use social media. These findings echo the paradigm that being an ethically responsible business provides a wide range of positive externalities. We find that even positive language in firm generated content is not detrimental to brand perceptions for brands with higher corporate reputation. Thus, keeping the house in order seems to be an overall effective strategy in the digital age.