Past service recovery research is dyadic in nature, focusing most prominently on bilateral relationships between the firm and the individual customer. However, customers rarely make use of services alone, thus creating a failure context where another customer (i.e., bystander) is socially present, and switching from a dyadic to a triadic perspective. A phenomenon, idiosyncratic to triadic relationships, is coalition forming, which refers to the temporary alliance among individuals to pursue a specific goal. In service recovery, customers may form coalitions to increase compensation likelihoods and service employees may form coalitions to strengthen their position against roaring complainants. Prior research suggests that relationship strength between people drives coalition forming, which can be explained by emotional contagion processes. Therefore, we seek to answer: (1) Are coalitions formed in service failure situations, and (2) do they follow predictable patterns? To answer our research questions we conducted a 2x2x2 experiment (N=1242), using video and photographic material to manipulate relationship strengths (high vs. low) between the complainant, the bystander, and the service employee. The trilateral dimension was incorporated as another between-factor by illustrating the eight different relationship-strengths versions in the perspectives of the three parties. Manipulations and variable measurement were checked successfully. T-tests confirm that the likelihoods of the complainant-bystander and the service employee-bystander coalition are significantly larger than zero (p’s < .01). Specifically, the likelihood-values ranged from 4.5% to 33.1%. For the complainant-bystander relationship, results reveal a significant and positive indirect effect of relationship strength on the complainant’s coalition forming intentions with the bystander, mediated by his emotional contagion with the bystander. The same holds true vice versa for the bystander’s coalition forming intentions with the complainant. Accordingly, data reveal a positive indirect effect of relationship strength on the likelihood of the complainant-bystander coalition, mediated by both party’s emotional synchrony (i.e., measured as the product of both party’s emotional contagion with each other). Likewise results were obtained for the service employee-bystander relationship, where relationship strength eventually increased the likelihood of the service employee-bystander coalition, mediated by both party’s emotional synchrony. Our results show that there is value in considering service failure and recovery situations from a triadic perspective, which so far has received considerable research attention outside service recovery, for instance in social-psychological domains such as family therapies.