Background:

In February 2020, AccLab South Africa launched an experiment  that sought to increase UNDP SA’s online engagement after drawing inspiration from the sizeable following of popular South African media houses.  At the time UNDP SA had 3000 Twitter and 7500 Facebook followers in a country of ~59 million people (59% of whom have internet access). Partnering with popular media houses was seen as an opportunity to boost UNDP SA’s online following, and for the AccLab this meant we could reach more people to source unusual solutions, and additionally, assist the CO to raise public awareness and participation in its activities. With the onset of COVID-19 in March 2020, little did we realise that such online interactions would soon become the new normal - accentuating the need for improved digital interaction more than ever before.

The Experiment:

In February 2020, two experimental posts were launched to encourage UNDP SA’s existing online followers to interact with us digitally. As part of the experiment, popular media houses and innovation partners were requested to share these posts with their networks, thereby allowing UNDP SA to tap into their audiences and increase its digital engagement. The experimental hypothesis was: If SA media and innovation partners shared/advertised UNDP SA posts, then UNDP SA will receive greater online engagement.

Figure 1: Experimental posts on UNDP SA social media platforms and shared/advertised by popular media houses and innovation partners

The performance of the experimental post on UNDP SA’s Facebook page was examined and reported on in Part I of this Blog. The findings on Facebook demonstrated that having media and other UNDP partners share and advertise UNDP SA posts on their own social media, radio and online news platforms results in an increased online engagement on UNDP SA’s posts. While the number of Facebook followers increased only slightly during the experimental period, it was a rise nonetheless, and was higher than that observed over the last one year. The rise indicated that should UNDP SA continue to request partners/media to share and advertise our posts/articles, over time the gradual increase in followers will result in a sustained increase in online participation and interaction in the long term. In this blog, we report on the experimental results observed on UNDP SA’s Twitter Account and its website.

The performance of the experimental posts on UNDP SA’s Twitter account were examined. Their online reach and engagement were compared to that of other UNDP SA tweets between February 2019 and April 2020. Figure 2a shows that UNDP SA tweets received the greatest interaction (impressions, retweets, likes, etc.) between 21 Jan 2020 and April 2020. However, the experimental posts (tweeted on 28 Feb 2020) were not solely responsible for this, as other tweets during this period garnered greater online interaction. The two experimental tweets received 2677 and 2561 impressions each. This is well above the average impressions received for all tweets between Feb 2019 and 20^th Jan 2020. The total impressions on each experimental tweet were below the maximum impressions on tweets during at least 3 analysis periods (see Figure 2a). After assessing the tweets that obtained more impressions than the experimental posts, the following trends were identified.  

Between Jan 2020 and April 2020, tweets promoting events, and which tagged the organising parties, received more impressions that the experimental posts. Such an example were the tweets made during the Independent Electoral Committee (IEC) Conference, where UNDP SA served as a co-host. Between June 2019 and Aug 2019 tweets calling on citizens to highlight their Mandela Day initiatives received significant interaction on twitter. Additionally, a UNDP workshop in Dakar to kick off the Accelerator Labs also received much attention. From March 2019 to May 2019, an event hosting the RBA director and posts welcoming new staff members elicited a high number of impressions. This analysis suggests that on Twitter, tweets promoting events receive the greatest online interaction. In addition, tagging co-hosts/partners/donors, etc in tweets, and having these partners re-tweet the posts, boosts the exposure of UNDP SA. In fact, this trend, of having media and innovation partners share and boost UNDP SA’s posts, provide further (and unexpected) evidence proving the experimental hypothesis true.

Figure 2: (a) Detailed analysis of UNDP SA tweets between Feb 2019 and April 2020, (b) Examples of UNDP SA tweets covering events and tagging partners that received the greatest online interaction.

UNDP SA’s website analytics indicated that the experiment resulted in increased traffic on the UNDP SA website. Figure 3a illustrates that the total number of users on the website increased to 4883 in February 2020. This was higher than the total number of users per month during the previous year. Slightly higher traffic was seen only in Oct and Nov 2019 (where users reached 5036 and 5308 respectively).

During October 2019 two major job adverts for UNDP were launched on the website, and during November 2019, a call for SMMEs to participate in the SA Entrepreneurial Summit was launched. These events/activities increased the number of website users and aligns to the findings on Facebook, where job adverts featured in the Top 10 Facebook posts. It also aligns to the findings on Twitter where events received the greatest online interaction. The website analytics also demonstrated that the total number of sessions and pageviews peaked, reaching its highest level during the experimental period (Feb 2020 to April 2020) when compared to the last one year. However, based on the twitter analytics, the IEC Conference held in March 2020 would have also contributed to the increase in website traffic during the experimental period. 

Figure 3: UNDP SA Website Analytics showing (a) number of website users, (b) number of web sessions, and (c) number of pageviews between February 2019 and April 2020; including the experimental period between Feb 2020 and April 2020. 

Summary of Findings:

On Facebook, the experiment resulted in a significant increase in the extent of online engagement and post views. While the total number of Facebook followers increased during the experimental period, it was not a major rise when compared to the previous year. The experimental post was the only call to action type post that featured in UNDP SA’s top 10 posts between Feb 2019 and April 2020. These results prove the experimental hypothesis true.

On Twitter, tweets posted during live events such as workshops and conferences, and which tagged UNDP SA’s co-hosting/donor/participating partners received the greatest online interaction. These posts received more online interaction than the experimental posts, however, also prove the hypothesis true; that when partners/media share UNDP SA’s posts, UNDP SA receives greater online engagement by leveraging the audiences of their partners. 

On the UNDP SA website, a higher number of sessions and pageviews were observed during the experimental period. The number of users on the website during the experimental period rose but were not as high as that observed during Nov and Oct 2019 - when major job adverts and events were hosted by UNDP.  This further confirmed the Twitter and Facebook analysis findings that events, and job adverts receive greater online responses.

While a high engagement on job adverts and events is desired, there is as much need to increase user engagement on posts requiring action, such as citizens contributing to reports, policies, innovation projects, etc. Since the experimental hypothesis proved true; UNDP SA can increase its online engagement and citizen participation on call to action type posts by leveraging the online audiences of media and partners. Moving forward, UNDP SA could tag their partners and media, and directly request retweets, shares and adverts by them to boost online engagement. Over time, this will gradually result in an increased number of UNDP SA page/account followers, sustaining an increased online engagement in future. 

For more information and to share ideas feel free to contact me, at klariska.moodley@undp.org.