The use of social networking platforms for sexual health promotion: identifying key strategies for successful user engagement

This study was a prospective descriptive analysis of Facebook and Twitter profiles drawn from our previous systematic examination of SNP for sexual health promotion carried out in 2010 [23]. The search strategy (described in detail previously) [23] used databases of published scientific literature, SNP search functions, and electronic sources of grey literature to identify SNP involved in sexual health promotion.

Search results and inclusion criteria for SNP in this analysis are described in Figure 1. Of the 178 SNP identified in the previous review, we included only Facebook and Twitter profiles (n = 130) given that they are the most popular platforms globally [31] and the most enduring over the follow-up period. Profiles were excluded if they were no longer available online or were ‘inactive’; that is, if the host organisation had not posted in the 30 days prior to data collection (September 1, 2011). We performed quantitative and qualitative analyses on the selected 100 Facebook and Twitter profiles that remained active and accessible.

We examined one month of content from the 60 Facebook profiles (August 6 to September 5, 2011) and 40 Twitter profiles (September 4 to October 4, 2011). Two researchers extracted and recorded relevant quantitative metrics (see Table 1). We generated PDF images of web content covering the period of interest to enable researchers to revisit content. Data were analysed in Stata 11 (StataCorp., Texas, 2009).

We defined ‘reach’ as the size of the user base, that is, the number of Twitter followers or Facebook ‘page likes’. (‘Liking’ an organisation’s Facebook page or following their Twitter account is equivalent to subscribing to the organisation’s status updates or tweets).

‘Interaction’ was used to indicate user engagement, and was defined as any interaction that occurred between users or between a user and the host of the profile. To create a measure of overall interaction, we summed all interaction metrics (Table 1) for each profile to get a ‘total interaction score’. We also calculated mean interactions per user for each profile by dividing the interaction score by the number of users.

Our measure of ‘engagement’ combined the size of the user base (reach) and the level of interaction of the users. We considered profiles with the largest user bases and highest levels of interaction to have the highest levels of online engagement. To measure engagement, we ranked all profiles within each category of reach and interaction, with profiles that had the greatest reach and highest interaction scores being ranked highest. We then summed the reach and interaction ranks for each profile and ranked these sums again to calculate an overall engagement success ranking.

We used quantitative and qualitative methods to compare top-ranked and low-ranked profiles to identify factors and strategies associated with engagement.

Ethics approval was received from The Alfred Health Human Ethics Committee (Project 6/15).

Eight host organisations had both Facebook and Twitter profiles in the top ten ranking for engagement, demonstrating that some organisations were able to achieve success across both platforms.

Table 2 illustrates reach and interaction metrics for top ten ranked profiles and all other profiles. For the top ten Facebook profiles, the median number of users was 15,156 compared with 560 for all other profiles; and for Twitter, 8,558 compared with 852. The median total number of user interactions for top ten Facebook profiles was 1325 versus 16 for all other Facebook profiles, and for Twitter, 937 versus 37. The most common types of interaction in the top ten profiles were Facebook likes of posts, and Twitter follower retweets, however across all Twitter profiles, ‘mentions’ were the most common type of interaction.

Interaction per user and activity metrics are shown in Table 3 for top ten ranked profiles and all other profiles. The level of interaction per user was significantly greater for top ten ranked Facebook profiles than all other profiles (p = 0.05), but no significant difference (p = 0.57) existed between top ten Twitter profiles and all other profiles. Top-ranked profiles were more active than other profiles: they had a median of 46 posts per month compared with nine for Facebook profiles (p < 0.01), and 124 tweets per month versus 29 for Twitter profiles (p < 0.01).

Previous work in this area has identified common key metrics, including reach and interaction, which can be used to help measure, monitor and inform the success of such social media interventions [8,25,27]. However, relatively few have actually gone beyond using these metrics to measure success, to identify features or strategies that could enhance the health promotion potential of web information systems and online applications [28-30]. We chose reach and interaction as our key metrics to measure online engagement by SNP-based sexual health promotion campaigns. Through quantitative methods, we found that higher levels of host activity (posts and tweets) and greater levels of interaction per user characterised profiles with high levels of engagement. We then explored associations between SNP metrics to gain more insight into achieving increased reach and interaction. We found a strong positive correlation between host activity and levels of interaction per user, especially for top ten Twitter profiles. This suggests that it is important to maintain host-initiated activity on SNP to encourage meaningful interactions both between the host and users, but also between users, in order to ensure ongoing engagement of the target audience. However, it is unlikely that high levels of activity alone are enough to ensure interaction, and consideration should be given to the type of activity that is most likely to promote interaction with and between users, and what type of activity generates certain kinds of interaction. For example, organisations may want to go beyond simply reaching people with messages, and focus more on drawing users into conversations and debate on a topic, that is, considering the content value of posts and tweets [34].

For top-ranked Facebook profiles, there was no association between the reach and the level of interaction per user. That is, the likelihood of a Facebook user interacting is not related to the number of users/friends of the profile. Interaction per user may be more related to host activity (as above) or the target audience or purpose of the profile, e.g. forming an interactive online community versus broad dissemination of health information. For top-ranked Twitter profiles, there was a strong negative correlation between reach and interaction per user: as the number of users increased, the level of interaction per user decreased. This may be reflective of the types of users and how and why people use Twitter. It suggests that the top-ranked Twitter profiles are broadcasting to large numbers of users, with many users being passive receivers of the information rather than interacting and engaging in conversation. Twitter is commonly used to post messages to many users, to follow live updates during major events (sporting, cultural, news), or to receive regular brief updates from large organisations or celebrities. Conversation on Twitter is generally limited in comparison to Facebook, which is designed to support social networks of friends or individuals with common interests, allowing them to share content by posting photos, links or creating stories and conversations. Organisations should reflect on their health promotion goals and target audience when selecting an appropriate SNP platform; for some reaching a large audience with brief messages is more important, but others seek primarily to engage a particular group or community and generate discussion.

Through the simplified qualitative content analysis of top-ranked and low-ranked profiles, we identified key strategies that characterised successful online engagement: making regular posts; directly engaging with users through individual responses and acknowledgment; encouraging interaction and conversation by posing questions; utilising multimedia uploads and relevant links; and highlighting celebrity involvement. Other researchers have identified similar strategies [27-30]. Preece and Shneiderman (2009)analysed online social activities and developed a framework of ‘usability’ and ‘sociability’ factors that they believed motivated online social participation [27]. They described how organisations could harness usability factors such as frequent updates and posting interesting, relevant and well-presented content to ensure users could easily engage with their content. Similarly, they identified sociability factors such as using charismatic leaders or respected authorities, and enabling development of user ‘visibility’ (an online social presence), which increased the chances of user engagement [27]. Others have found that acknowledging users’ contributions enables them to stand out or gain ‘digital status’ and is a key motivator for online contribution or interaction [35,36]. Utilising these key strategies effectively requires dedicated resources and careful planning, design and monitoring [24,37]. In our research, top ten profiles appeared to be part of long-term ventures – organisations’ overall communication strategies – rather than short-term campaigns or initiatives, highlighting the importance of maintaining an online presence and working towards building online communities engaged with a cause/issue. It is clear from both our review and the work of others that in order to deliver successful interventions through SNP, hosts use various key strategies to take advantage of the different functionalities of the platforms, as well as improving the quality and utility of content being delivered to their target audience.