The effect of interventions on Twitter in four target groups using different measures of influence

In this paper, the influence of interventions on Twitter users is studied. We define influence in (a) number of participants, (b) size of the audience, (c) amount of activity, and (d) reach. Influence is studied for four different target groups: (a) politicians, (b) journalists, (c) employees and (d) the general public. Furthermore, two types of interventions are studied: (a) by all Twitter users (i.e., uncontrolled interventions), and (b) those tweeted by an organization that benefits from any resulting influence (i.e., controlled interventions). As a case study, tweets about a large Dutch governmental organization are used. Results show a relation between the number of uncontrolled interventions and influence in all four target groups, for each of the defined types of influence. Controlled interventions show less influence: significant influence was found for the general public, but influence for politicians and employees was only mildly significant, and no influence was found for journalists. The effect found for uncontrolled interventions, however, suggests that this influence is indeed reachable for some target groups, even when the number of interventions is small, and very well reachable for all target groups, provided the number of interventions is large enough. In addition to this, we found that interventions influence groups to a different extent. Own employees were influenced strongest, differing significantly from the other groups. © 2014, Springer-Verlag Wien