When pushing back is good: the effectiveness of brand responses to social media complaints

Although both orientations provide their respective type of benefits primarily, they do not do so exclusively. That is, hedonic and utilitarian benefits are distinct, rather than functioning as two poles on a single continuum (Babin et al. 1994; Cotte et al. 2006). When they co-occur, one type of benefit may dominate the other.

In addition, we identified 2608 positive, 50,788 neutral, and 990 non-identifiable comments. A neutral code indicated the comments contained no valuations, such as tagging other persons, responses to lotteries, or neutral questions (e.g., “Where can I find product X?”). Non-identifiable comments featured different languages, non-readable characters, or incomprehensible statements.

Respondents were recruited through a commercial online panel and randomly assigned to one of four conditions in a 2 (primed observer benefits: hedonic or utilitarian) × 2 (response strategy: accommodative versus defensive) between-subjects design. With an unconscious priming procedure, we prompted the respondents to seek hedonic or utilitarian benefits when evaluating the brand responses (see Study 2 for details). After reviewing a stimulus that contained a user complaint below a brand post and a corresponding brand response (accommodative or defensive). The respondents indicated if they wanted to click the “Like” button in response to the brand reply, write a comment, do both, or do nothing at all. We also measured their likelihood of clicking the “Like” button or posting a comment on two single-item, 11-point scales (likelihood between 0 and 100% in 10% increments). If respondents wanted to write a comment, we asked them what it would be. Next, we measured respondents’ perceived benefits using a pooled scale of either hedonic or utilitarian benefits (see Study 2; α_hed = .893, α_uti = .843). Finally, respondents indicated their general tendency to click the “Like” button for brand posts and comments on a single-item, six-point scale from “never” to “several times a day.” In total, 60 respondents indicated that they would click the “Like” button, 14 intended to write a positive comment (8 would do both), and 103 chose no positive reaction (17 indicated they would write a negative comment). A chi-square test to assess the association between perceived benefits and positive responses to the brand reply indicated a significant χ² value (Pearson) of 54.06 (p < .001) and a significant Cramer’s V (.566, p < .001). We also conducted a linear regression with the likelihood of clicking the “Like” button as a dependent variable. In a model without any controls, perceived benefits explained 33.02% of the variance in the probability of clicking the “Like” button (standardized β = .575, p < .001). When we included the experimental factors (priming condition and response strategy,) as well as respondents’ general tendency to click “Like”, perceived benefits still had a strong, significant effect on the probability of clicking the “Like” button (β = .505, p < .001; R² = 37.19%). This relationship between perceived benefits and clicking the “Like” button remained consistent in all subanalyses for each experimental group (all β > .50, p < .001).

As an additional robustness check, we conducted the analyses with only “Likes” as the dependent variable (without positive comments; Relling et al. 2016) and found no substantial differences. The dependent variable that combined both “Likes” and positive comments was more comprehensive, so we used it in our main model analyses.

The following criteria were used: word count (N_acc = 177, N_def = 178), words per sentence (M_acc = 14.8, M_def = 14.8), use of social words (Share_acc = 15.2%, Share_def = 13.5%), use of emotional words (Share_acc = 6.2%, Share_def = 8.4%), and pronouns (Share_acc = 11.9%, Share_def = 10.7%).

A person’s mood can affect how he or she processes information and motivation to engage in resisting counterarguments (Wegener et al. 1995). If the priming procedure were to induce a more positive mood in the hedonic condition than in the utilitarian condition, the effect of the response strategy could be confounded by a mood effect. However, observers also might seek hedonic benefits when they are in good or bad moods, perhaps as an instrument for mood management (Babin et al. 1994).