If you have been watching our activity closely, you probably know that our specialty is identifying employees (of any company) on Twitter, and analyzing their activity on Twitter from the employee advocacy perspective. You can see this in some of our previous analyses covering UK companies or tech companies, as well as in our employee advocacy leaderboard.
However, in the last few months we have been working on a new layer of even deeper analytics. The scope of the new analytics is to better understand the content that is shared by the employees: where it is shared from, how often it mentions the brand, and how it drives engagement.
To illustrate the new developments, we decided to take a break from our habit of analyzing multiple companies, and to focus on a single one instead: Adobe. Why Adobe? Because it consistently ranks high in our tech leaderboard, it has a lot of employees on Twitter (almost 2,000). They also launched a company-wide program for employee training and activation called Social Shift.
To better understand why it is successful in employee advocacy, we’ve asked Lauren Friedman, Head of Social Business Enablement at Adobe, to provide us an inside perspective.
For this post, we analyzed the activity of 1,967 Adobe employees for 2 months, from December 1st to January 31st. We analyzed more than 65,000 tweets from the employees in this period. The timeframe includes the Christmas holidays which naturally translate in lower brand-related Twitter activity.
The first thing that we looked at is what platforms were the most used by employees to post on Twitter, and the table below provides the answers. Unsurprisingly, the Twitter web client and the mobile apps for iOS and Android are the most used, but we see other platforms like Instagram, EveryoneSocial (an employee advocacy platform), and LinkedIn. One important observation is that there is no single platform that dominates employee activity. Only Twitter for iOS and the web client have a share of activity larger than 20%, and none others have more than 10%.
If we look at how many tweets were posted by each user on a platform, TweetBot and EveryoneSocial are clear leaders, followed by TweetDeck. No platform reaches a frequency of 1 tweet per active user per day.
|Platform||Users||Tweets||Brand Tweets||Tweets / User|
|Twitter Web Client||846||16484||2001||19.5|
|Twitter for iPhone||679||19230||2071||28.3|
|Twitter for Android||180||4519||335||25.1|
|Twitter for iPad||126||2323||175||18.4|
|Twitter for Mac||56||1500||83||26.8|
|Mobile Web (M5)||45||125||12||2.8|
|TweetBot for i?S||31||1405||57||45.3|
When it comes to tweets about the brand (i.e. those that mention ‘Adobe’), things are changing. Out of the most used platforms, LinkedIn has the highest percentage of tweets about the brand (26.4%). This is in line with Adobe’s Social Shift social media training program, where “we focus our training efforts on Twitter and LinkedIn as those are used more for professional and company news than Facebook, etc”, says Lauren. Other platforms with high percentages of tweets about the brand are EveryoneSocial (22.6%) and TweetDeck (15.2%) You can see the complete list below.
|Platform||Users||Tweets||Brand Tweets||Brand Tweets %|
|Twitter Web Client||846||16484||2001||12.14%|
|Twitter for iPhone||679||19230||2071||10.77%|
|Mobile Web (M5)||45||125||12||9.60%|
|Twitter for iPad||126||2323||175||7.53%|
|Twitter for Android||180||4519||335||7.41%|
|Twitter for Mac||56||1500||83||5.53%|
|TweetBot for i?S||31||1405||57||4.06%|
Finally, we wanted to look at how much engagement is generated by each platform, in terms of likes and retweets. Below, you can see the results, ordered by retweets per tweet (you can sort by other columns, too). The platforms that lead the pack are Buffer, Twitter for Mac, and TweetDeck, with at least 1 retweet for every 2 tweets and 1 like per tweet. The platforms with the lowest number of retweets are EveryoneSocial, LinkedIn, and Instagram. The ranking is roughly similar if we look at the number of likes per tweet.
One possible explanation for the low engagement on EveryoneSocial and LinkedIn is the low number of followers per tweet, meaning that employees that tweet on these platforms are less savvy than those using other platforms. This hypothesis is confirmed by Lauren: “Many Adobe employees using the EveryoneSocial tool and participating in our Employee Advocacy program are, in fact, social media “beginners.” We’re working closely with them to help enable and empower them to be more active on social media and grow their audiences (resulting in more engagement).”
In the case of Instagram, the reason for low engagement could be that the majority of interactions are happening on Instagram itself rather than on Twitter.
|Source||Tweets||Accounts||Likes||Retweets||Avg Likes||Avg Retweets||Avg Followers|
|Twitter for Mac||1290||55||1584||781||1.23||0.61||33393|
|Twitter for iPad||1494||97||2300||660||1.54||0.44||15198|
|Twitter for Android||2482||150||4391||943||1.77||0.38||3816|
|Twitter Web Client||11782||776||11739||3223||1.00||0.27||2293|
|Twitter for iPhone||10893||565||11842||2758||1.09||0.25||3908|
The other main focus of our analysis was looking at the content of the tweets and trying to find out what makes them engaging. We went about this by studying four “entity” types for each tweet: URLs, hashtags, user mentions or pictures. Then, we categorized the tweets into 3 categories of engagement: those that had no retweets, those that had between 1 and 4 retweets, and those that had at least 5 retweets. We excluded tweets that are themselves retweets for this analysis. The table below will show you the results.
The most often used entities where URLs, increasing from 49% for tweets with no retweets to 63% for tweets that have at least one. There seems to be a correlation between usage of URLs and the number of retweets. In similar fashion (but with a more visible trend), usage of media (pictures, images) generates more engagement. Among tweets with no engagement, 16% had a hashtag and 20% had a picture. If we look at tweets 1 to 4 retweets the number increases to 28%, and then to 39% for at least 5 retweets. This is a significant increase. For user mentions and hashtags, the impact varies. Tweets that have between 1 and 4 retweets have more of them than tweets with no engagement. However, the values decrease again for tweets with at least 5 retweets. so no trend can be identified. Also, surprisingly, tweets with the highest engagement contain the largest percentage of tweets with no entities (14%).
In the table below, you’ll see similar stats, but this time only for tweets that mention the brand. Among brand-mentioning posts, URLs appear in 74% to 76% of them. But for hashtags, user mentions and pictures the dynamic is changing. Tweets with 1-4 retweets have the most of each, but their frequency decreases for tweets with 5+ retweets. Even so, pictures and hashtags seem to be stronger indicators of higher engagement than user mentions.
After taking a close look at the tweets posted by Adobe employees on Twitter, we were able to draw several important conclusions. First, the activity of employees is spread along multiple Twitter platforms (client), with none of them having a majority share of activity, and only 2 having between 20% and 30% share.
Second, the frequency, topics (brand-related or not) and engagement of tweets vary significantly for each platform.
Third, while entities like URLs, hashtags, user mentions or pictures are not consistent indicators of a better engagement, some of them, like pictures and hashtags generate more retweets in almost any case. URLs have a neutral effect, while user mentions can help generate a few retweets, even if they are not that used in the most popular tweets.
In addition to what we learned in this post from analyzing the platforms and content of Adobe employees in detail, looking at the wider picture shows that their social efforts are clearly paying off. They have more than 15% of their employees identifying themselves in their Twitter bio, more than any company that we analyzed so far. The 2000 accounts that we analyzed have almost 1.5 million followers collectively, and posted almost 4000 tweets about the brand in January.
I’d like to end this post with a crisp explanation from Lauren Friedman regarding the objectives of their program: “Our Employee Advocacy program is heavily focused on building a personal brand. We are not working to build an “army of Adobe Bots.” Instead, we want to encourage our employees to be well-rounded thought leaders in their areas of expertise.” It looks like they are well on their way to achieving this!