Journal of Marketing Communications, volume 30, issue 3, pages 301–317, 2024 (online 13 August 2022)
- 564 Facebook users
- Online experiment
- Apology vs corrective action
- Moderation analysis
- Portugal
Free full text: the authors’ accepted manuscript, identical in content to the published article. Please cite the published version.
- crisis management
- online firestorm
- SCCT
- situational crisis communication theory
- social media
- online complaints
- negative word of mouth (e-WOM)
- virality
- brand attitude
- crisis response strategies
- apology
- corrective action
- Facebook brand pages
- brand reputation
- complaint handling
- unethical behaviour
- customer service failure
- online experiment
- Facebook look-alike posts
- fictitious brand
- paired t-test
- multiple regression
- moderation analysis (PROCESS)
- SPSS
- Portugal
- social media crisis response
- community management
When an angry complaint about a brand starts spreading on Facebook, the brand's reply changes how people feel about it, and better feelings make them less willing to share the complaint. In an online experiment with 564 Facebook users in Portugal, an apology worked best overall, but for service failures and unprofessional communication a corrective action did just as well.
Key findings
- A reply to the complaint lifts brand attitude. After reading a complaint on the brand’s Facebook page, participants’ attitude towards the brand averaged 2.21 out of 7; after the brand’s reply it rose to 3.54 (p < .001), for all three types of incident (Langaro et al., 2024).
- A better brand attitude reduces the urge to share the complaint. Brand attitude after the reply lowered intentions to forward the negative post to others (β = −.178, p < .001), controlling for gender and age (Langaro et al., 2024).
- An apology works best in general. The type of reply moderated this effect (interaction β = .262, p < .05): after an apology, better brand attitudes cut forwarding intentions far more steeply than after a corrective action (Langaro et al., 2024).
- For service and communication failures, both replies work. For complaints about core business problems (poor customer service) or unprofessional communication (deleted comments), the moderation was not significant: apologies and corrective actions performed comparably (Langaro et al., 2024).
At a glance
- Design
- Online experiment on Qualtrics: each participant saw one of three simulated Facebook complaints about a fictitious retailer (Lusoretail), rated the brand, saw a randomly assigned reply (apology or corrective action) and rated the brand again
- Participants
- 564 Facebook users in Portugal recruited through Facebook (survey in Portuguese, back-translated); 58.3% men; ages 18 to 65 and over; 282 saw each reply
- Incidents
- Unethical behaviour (dumping waste in a river), a core business problem (unreachable customer service) and unprofessional communication (deleting critical comments), all high-arousal and with the firm in control (preventable crises)
- Measures
- Brand attitude: five 7-point semantic differential items (Spears & Singh, 2004), α = 0.97. Intention to forward the negative post: two items (Chiu et al., 2007), α = 0.92
- Analysis
- Paired t-test (H1), multiple regression with gender and age as controls (H2) and moderation with PROCESS 3.5 (H3), in SPSS 25
Figure 1. Moderation effects of the type of response strategy
Brand attitude before and after the brand's reply
Tables
Table 1. Detailed characterization of the sample
| Sample characterization | Total sample |
|---|---|
| N | 564 |
| Gender: male | female | 329 (58.3%) | 235 (41.7%) |
| Age | |
| 18–24 | 124 (22.0%) |
| 25–34 | 125 (22.2%) |
| 35–44 | 139 (24.6%) |
| 45–54 | 98 (17.4%) |
| 55–64 | 58 (10.3%) |
| 65 and over | 20 (3.5%) |
| Education | |
| Basic education (9th grade) | 34 (6.0%) |
| High school degree (12th grade) | 142 (25.2%) |
| Bachelor’s degree | 216 (38.3%) |
| Master’s degree | 143 (25.4%) |
| PhD | 29 (5.1%) |
| Profession | |
| Unemployed | 38 (6.7%) |
| Worker | 335 (59.4%) |
| Student | 120 (21.3%) |
| Student worker | 43 (7.6%) |
| Retired | 28 (5.0%) |
| Facebook usage | |
| Less than once per week | 33 (5.9%) |
| Once per week | 8 (1.4%) |
| Two to three times per week | 57 (10.1%) |
| Once per day | 141 (25.0%) |
| Several times per day | 325 (57.6%) |
| Average daily time spent on Facebook | |
| Less than an hour | 147 (31.5%) |
| 1–2 hours | 116 (24.9%) |
| 2–3 hours | 111 (23.8%) |
| 3–4 hours | 58 (12.4%) |
| More than 4 hours | 34 (7.3%) |
Values as printed in the article (decimal commas changed to points). The daily-time categories add up to 466, not 564.
Table 2. Descriptive statistics and intercorrelations across variables
| Variable | α | M | SD | 1 | 2 | 3 |
|---|---|---|---|---|---|---|
| 1. Brand attitude before the response (BA PRE) | 0.97 | 2.21 | 1.27 | – | .318** | −.147** |
| 2. Brand attitude after the response (BA POST) | 0.97 | 3.54 | 1.29 | .318** | – | −.179** |
| 3. Intentions to forward negative e-WOM | 0.92 | 3.69 | 1.84 | −.147** | −.179** | – |
** p < 0.001. Values as printed in the article.
The question
Online firestorms, sudden floods of complaints about a brand on social media, threaten reputation and sales, and companies reply hoping to stop people from passing the complaint on. Situational crisis communication theory says that when a crisis is preventable and the firm is to blame, it should accept responsibility, but it does not say whether an apology or a corrective action works better, nor how the reply curbs sharing. We asked whether a reply changes brand attitude, whether that brand attitude reduces intentions to forward the complaint, and whether apology and corrective action differ in this effect.
Implications
Practical implications
Brands should always reply to complaints that could escalate: almost a third of negative posts on social media get no response at all. Because brand attitude carries the effect, community managers should craft replies that are likeable, agreeable and appealing in wording and structure. In general an apology is the safer choice; when the complaint is about a core business problem or unprofessional communication, an apology or a corrective action can be chosen depending on what is operationally and legally convenient, which calls for careful protocols and training.
Theoretical implications
The study extends situational crisis communication theory to online firestorms on social media and shows that it fits this context. It identifies brand attitude as the mechanism through which matching response strategies curb intentions to forward negative content, and it separates the effects of two accommodative strategies, apology and corrective action, which the theory had treated as equivalent, showing that the type of incident matters.
Limitations and future research
- Participants volunteered, so the sample consists mostly of frequent Facebook users.
- The brand was fictitious, and the wording of the three complaints varied somewhat.
- Future work could add crisis history, prior reputation and the volume of complaints, and test forgiveness and personal identification with the incident as mechanisms.
Questions this paper answers
Should a brand apologise or fix the problem during an online firestorm?
In general an apology worked best: in an experiment with 564 Facebook users, better brand attitude after an apology reduced intentions to share the complaint more than after a corrective action. For service failures and unprofessional communication both worked about as well (Langaro et al., 2024).
Does replying to a complaint on Facebook help a brand?
Yes. Brand attitude rose from 2.21 to 3.54 out of 7 after the brand replied, and better brand attitude lowered intentions to forward the complaint (Langaro et al., 2024).
What is an online firestorm?
A sudden flood of negative comments and complaints about a brand or person on social media, often triggered by a single complaint and amplified as others join in and share it.
What is situational crisis communication theory (SCCT)?
A theory that matches a firm’s crisis response to how responsible the public thinks it is: the more preventable the crisis, the more the reply should accept responsibility, for example by apologising or taking corrective action (Coombs, 2007).
Key concepts
- Online firestorm
- A sudden discharge of large numbers of negative messages about a brand on social media (Pfeffer, Zorbach & Carley, 2014).
- Situational crisis communication theory (SCCT)
- A theory that matches crisis response strategies to the responsibility the public attributes to the organisation (Coombs, 2007).
- Apology
- A victim-centred response in which the organisation expresses remorse and accepts responsibility.
- Corrective action
- A victim-centred response in which the organisation fixes the problem and commits to preventing it from happening again.
- Brand attitude
- A consumer’s overall evaluation of a brand, from bad to good and unappealing to appealing.
- Negative e-WOM
- Negative electronic word of mouth: sharing or forwarding unfavourable brand-related content to others online.
Open science and materials
- Accepted manuscript (PDF) Free full text, including the tables and figure
- Published version Journal of Marketing Communications (Taylor & Francis)
Study materials
- Complaints (stimuli)
- Unethical behaviour: ‘How are you capable of throwing waste – that you created – in the river that flows through the city? You are polluting the environment and affecting public health? DEPLORABLE and SHAMEFUL unethical practice!’ Core business problem: ‘I have been trying to contact your customer service for over 2 weeks without success. There are other clients in the same situation! UNACCEPTABLE and OUTRAGEOUS business practice!’ Unprofessional communication: ‘How were you capable of deleting my comment without responding to me? And apparently, I am not the only one complaining! INCOMPREHENSIBLE and UNFORGIVABLE communication practice!’
- Replies (stimuli)
- Apology: ‘We are very sorry for what happened, and we assume responsibility. On behalf of Lusoretail, please accept our sincere apologies’. Corrective action: ‘We are already taking corrective actions to solve the problem immediately. In the future, actions will be taken to prevent an incident like this from happening again at Lusoretail’.
- Scale items
- Brand attitude (7-point semantic differentials): unappealing/appealing, bad/good, unpleasant/pleasant, unfavorable/favorable, unlikable/likable. Intention to forward the negative post (1 = completely disagree, 7 = completely agree): ‘This Facebook post is worth sharing with others’; ‘I will forward this Facebook post to others’.
How to cite
Langaro, D., Loureiro, S., Schivinski, B., & Neves, H. (2024). In the eye of the (fire)storm: better safe or sorry? Crisis communication strategies for managing virality of online negative brand-related content. Journal of Marketing Communications, 30(3), 301–317. https://doi.org/10.1080/13527266.2022.2109056