It has been thirty years since the Grim Reaper advertisement haunted our television screens and shocked Australia. This controversial campaign was Australia’s hard-hitting response to the emerging threat of HIV.
Since then, mass media campaigns consisting of television, radio and print media have addressed diverse public health issues such as smoking, obesity, vaccination and road injury. Studies show that personal testimonies and negative (including fear), graphic and emotive health messages tend to be the most effective for inspiring behavioural change.
However, as the media environment changes, primarily through the increasing dominance of social media, health promotion and public health campaigns also need to change.
One key challenge is the need for public health organisations to increase their knowledge and practical experience of social media in order to develop, implement and evaluate the social media component of public health campaigns.
This article from a team at the Cancer Council Western Australia makes an important contribution to this area.  It provides some valuable insights into how to measure key outcomes from social marketing campaigns that include Facebook as a platform. Using examples from two CCWA campaigns, the authors outline how to obtain the relevant data from Facebook accounts and how to use this data to inform campaign evaluations and future communications activities.
Sarah Beasley, Justine Leavy and Simone Pettigrew write:
Mass media campaigns have become increasingly sophisticated and social media is now an essential component. Today, Facebook is the most popular platform in Australia, with approximately 15 million active monthly users. The targeting capability of Facebook advertising provides an excellent opportunity for public health organisations to reach their target audiences. But as Facebook continues to grow, so too does the content that is competing for space in users’ newsfeeds.
This forces public health organisations to move away from the familiar one-way communication style of traditional mass media and better adapt to the interactive style of social media. There has never been a more pressing need for public health organisations to better understand what makes Facebook content successful.
Through our work at CCWA, we have administrator-level access to the Facebook pages of Make Smoking History (MSH) and LiveLighter (LL). MSH and LL are comprehensive State-wide social marketing campaigns that include Facebook as a secondary media channel. Last year, CCWA and Curtin University worked together to determine which Facebook posts were the most ‘effective’. We chose reach and engagement rate as the key indicators of effectiveness.
Measuring reach and engagement
Being an administrator for a Facebook page means being able to access data directly from Facebook Insights for that page, which overcomes the weakness present in previous studies that use publicly available information to determine reach and engagement rate for Facebook posts.
For example, calculating an engagement rate without access to Facebook Insights often means taking the total number of visible post interactions (likes, comments, shares) and dividing by total page fans (likes). This means interactions such as link clicks which are only visible via Facebook Insights – and may actually be the end goal of a post – are not included. Reach is often estimated in a similar way, by adding up the total number of publically visible interactions on a post.
This will be an underestimate because the majority of people actually reached by a post will not interact with it in a visible way. Another weakness is that without access to Facebook Insights, it is also difficult to determine if a post has been boosted by advertising dollars. Reach figures for organic posts (those without advertising dollars) have been falling steadily.
Paid promotion of posts is becoming increasingly essential to get content in front of the people who follow your page and beyond. Paid posts have substantially increased numbers of post interactions so comparing them to organic posts is unfair.
Reach calculated by Facebook Insights
Using Facebook Insights data, you’re given an actual, discrete number of all people who saw your post, fans and non-fans, rather than an estimate.
Engagement rate calculated by Facebook Insights
Facebook Insights automatically calculates the engagement rate for individual posts as the number of people who liked, commented, shared or clicked on the post as a proportion of the total number of people the post reached. This means actions such as ‘link clicks’ are captured in addition to publicly visible interactions such as likes, comments and shares.
It is important to consider reach alongside a post’s engagement rate to put it into perspective. In some instances, a post may get a couple of quick likes to begin with from enthusiastic followers (often our own colleagues) then taper off quickly. The reach will remain low so the engagement rate will look impressive, even though the post was not seen by many people at all.
Measuring effectiveness
For our research into the MSH and LL Facebook pages, we downloaded descriptive data including engagement rates and reach for organic posts directly from Facebook Insights into Microsoft Excel. We randomly selected at least one post per week from each page uploaded in 2015 for inclusion in the study.
We used NVivo 10 to code the posts into tone and content categories and produced a table for each campaign showing the mean engagement rate and absolute reach for each category.
| Number of posts [n]1 | Engagement [%] Range (mean) |
Mean reach [n] | ||||
| LL | MSH | LL | MSH | LL | MSH | |
| Post content | ||||||
| Blog | 7 | 2 | 1-5 (3.0) | 1-4 (2.5) | 5588.7 | 2145.5 |
| Health Advice/Tips | 37 | 28 | 0-10 (4.1) | 1-13 (5.6) | 3462.4 | 2120.0 |
| Question | 2 | 2 | 2-6 (4.0) | 3-4 (3.5) | 2103.0 | 1695.0 |
| Competition | 4 | 0 | 2-7 (6.0) | – | 1051.8 | – |
| Testimonial | 2 | 3 | 4-7 (5.5) | 2-11 (6.7) | 5153.5 | 2876.3 |
| News Article | 7 | 12 | 2-21 (7.7) | 0-12 (4.2) | 1700.9 | 2171.7 |
| Campaign Information | 3 | 0 | 4-10 (6.0) | – | 3153.3 | – |
| Miscellaneous | 2 | 8 | 1-3 (2.0) | 0-11 (4.8) | 2403.0 | 2029.0 |
| Mean of means (μ x̅ ) |  |  | 4.8 | 4.6 | 3077.1 | 2172.9 |
| Â | ||||||
| Post tone | ||||||
| Informative | 39 | 28 | 0-21 (4.9) | 0-13 (5.1) | 3037.1 | 2076.2 |
| Emotive | 0 | 4 | – | 1-11 (5.5) | – | 1550.5 |
| Motivational/Supportive | 18 | 20 | 1-7 (3.7) | 2-13 (6.6) | 3666.9 | 2608.6 |
| Negative | 5 | 14 | 0-7 (4.5) | 1-11 (5.0) | 1500.4 | 2192.0 |
| Humorous | 1 | 12 | 3 (3.0) | 2-7 (4.6) | 2929.0 | 1884.6 |
| Other | 4 | 1 | 2-8 (6.0) | 3 (3.0) | 1051.8 | 1222.0 |
| Mean of means (μ x̅ ) |  |  | 4.4 | 5.0 | 2437.0 | 1922.3 |
1Posts can be coded to multiple categories
Key advice for optimising Facebook posts
Our research indicates that in order to drive higher levels of engagement and reach on Facebook, health organisations should prioritise:
- Testimonials: Posts with testimonial content were the most effective for both the MSH and LL campaigns.
- Posts that adopt a Motivational/Supportive/Encouraging tone: Posts with this tone were the most effective for MSH and also performed well for LL.
- Informative posts: Posts that contained new information that related to the LL campaign (Campaign Information), or any post delivered with an Informative tone tended to achieve above average engagement and reach.
 Competition and News Article posts gained high levels of engagement but did not perform as well in terms of reach. News article posts always included a link to the original article, which encourages engagement. Reach for competition posts may be hindered by the fact that sharing a post is likely to increase the number of entrants and reduce a person’s chances of winning.
Health Advice/Tips was the most common post content on both Facebook pages. Shocking, I know. The varying success of these posts was likely influenced by the strength of the categories they were coded with.
Some differences between the use of Facebook by MSH and LL were observed. Negative posts were more effective for MSH than for LL in terms of reach. This could be related to the subject matter of each campaign – smoking for MSH and overweight/obesity for LiveLighter.
The LiveLighter blog posts had a much greater reach than for MSH. However, overall, the Blog category had low engagement despite achieving average to high reach. This suggests that blog posts should focus on high-yielding content (such as testimonials) and encourage engagement by provoking online discussions and including a call to action.
The study indicates that although humorous posts need not be avoided altogether, they should be used with caution. Comedic posts are subjective and potentially controversial.
Conclusion
Although published studies of public health-related Facebook use are scarce, posts containing educational information, news articles or humorous content have tended to perform well. In contrast to these studies, our research has shown that testimonial posts have been the most effective across the MSH and LL Facebook pages, which is consistent with mass media advertising. Health advice/tips and informative posts are the most common, but could be strengthened by adopting a motivational tone wherever possible.
In many ways, these results highlight that Facebook is a social platform and posts should adopt an interactive and communicative style. The most effective posts are those that are personal or inspirational in nature and are therefore more likely to be shared amongst users. Consideration of post tone and content can maximise the impact public health organisations’ Facebook pages have in reaching and facilitating engagement with their target audiences.
