Social Media Analytics, Part 1
By: Dr. Kathy Kelley
In last month’s blog, Social Media Strategies for Tasting Rooms (http://bit.ly/1UUG2EM), I provided information to help you determine which social media networks you might choose and why; strategies on using some of the more common social networks (and a few new ones, too) to connect with audiences; when to post, how often, when to use #hashtags; and how to find content that you could include in your posts.
Over the next couple of posts, I will provide information on how to find and decipher the analytics for each of the more common tools, and some tips on how to use the data to make more of an impact. The information in this post focuses on Twitter.
What is the Twitter Analytics Dashboard and What Information does it Provide?
The Twitter Analytics dashboard (www.analytics.twitter.com) is free and can be accessed by anyone who has had a Twitter account for 14 days. It provides users with a summary of their Twitter activity for the calendar month and a 28-day period. I have included an image of Denise Gardner’s (@DeniseMGardner) “Home” tab on her dashboard to illustrate what one looks like.
You can find the following on the Home tab:
1) Number of followers. Denise had 393 followers, as of March 24, an increase of eight from the previous 28-day period.
2) Number of tweets that she sent in March 2016, as of the 24th (16 tweets; see image below), and for the previous 28 days (18 Tweets). Denise published four more tweets between February 26th and March 24th than she did the previous 28 days, which is indicated by the 28.6% increase in tweets.
3) Number of Impressions, which are “tweets sent that actually generate interaction or replies from others online” (http://bit.ly/1UOlQEs).
Denise’s 18 tweets had a total of 4,114 impressions, which is the number of followers (393) who viewed the tweet on Twitter’s Android and iOS apps or on Twitter.com (http://bit.ly/1nlIItD) and the number of replies she may have received during this period. This number does not include those who viewed her tweet on other platforms, like Hootsuite or Tweetdeck (http://bit.ly/1RDXey4).
A great visual that shows the number of impressions per day can be found on the “Tweet” tab.
While the daily average number of impressions that Denise received was 147, there were four days when the number of impressions were above 400, as demonstrated by the “height” of the light blue bars.
Notice that on the days when she had a greater number of impressions she did not necessarily post the most tweets. Rather, the days with a higher level of impressions are Fridays, the day that she tweets about a new Wine & Grape U. blog post, and the Monday and Tuesday of the following week.
Just below this graph is a list of her tweets for the 28-day period with their respective:
- Number of impressions.
- Total interactions with a tweet: number of retweets, tweets that mention her Twitter handle, users who “favorited” her tweets, new users who followed her, (http://bit.ly/1ncyZE0).
- Engagement rate. Although the engagement rate is already calculated, it is the number of engagements for a post divided by the total number of impressions. For example, Denise’s post on March 15 had 90 impressions and four engagements resulting in a 4.4% engagement rate (4/90 x 100 = 4.4%).
While the tweets are arranged in reverse chronological order, as seen in the image above, we could order them by:
- Top tweet, tweet with the most impressions/engagement would appear at the top of the list;
- tweets and replies, which would show how many followers replied to the tweet (see below); and
- promoted tweets, which are impressions and engagements generated for paid tweet that would “reach a winder group of users or to spark engagement from existing followers” (http://bit.ly/1KvffHH).
If I click on an individual tweet I can get even more data pertaining to the action that users took.
Some of the data you can you learn about each tweet (http://bit.ly/1lfPeU0):
- Detail expands. Clicking on the tweet provides additional detail (time and date the tweet was posted), all pieces of a multi-part conversation, and features (a follow/unfollow button, a textbox that can be used to reply, a menu with even more options).
- Number of users who liked the tweet by clicking on the heart icon.
- Link clicks, embedded media clicks, hashtag clicks, and profile clicks.
- Retweets and replies.
- When “your tweet resonated with someone else, and they wanted to give a virtual high five” (http://bit.ly/1EHhenw).
What was Denise’s top tweet based on number of impressions?
We can also learn which tweet generated the most impressions in March. As you can see in the image below, the tweet she published on March 11 that promoted her blog post, “What Penn State Extension Means to Me: From a Non-Agricultural Student to Today’s Extension Enologist” (http://bit.ly/1pcvTGm) resulted in 701 impressions.
Not only can she see the data for the 28-day period, but she can view data for just the past seven days, a specific calendar month, or up to a 91-day period. She can also use the CSV export tool to download data for up to 3,200 tweets going back to October 2013 (http://bit.ly/1nlIItD).
4) Number of Mentions: another Twitter user’s tweet that includes a @username (Twitter handle). Below is an example of a tweet that “mentions” her handle.
5) Number of users who visited her profile page. By accessing her profile page, users can learn how many total tweets she has published, a brief bio, a link to her LinkedIn account, followers in common with Denise, when she joined Twitter, thumbnails of her photos and videos, and others’ tweets that she “likes.”
With all this information available, what should you focus on?
While seeing “large numbers” of tweet impressions, mentions, followers, etc. on your dashboard is encouraging, the “percentage next to these numbers” is what you should be focusing on (http://bit.ly/1PvrhQJ), as it compares activity to that of the previous 28 days.
Although there was an increase in profile visits (an increase of 30.5%), it is not apparent who was looking at her profile and what action they took (followed Denise, retweeted her, etc.). Hence, this statistic might not provide you with information needed to make decisions. The increase in Denise’s impressions (an increase of 81.2%), mentions (an increase of 83.3%), and followers (an increase of 8%); however, are all indicators that her information was of value. Had the number of followers been lower than the previous 28-days, it would be necessary to look at her tweets for the two periods and try to detect what might have caused users to “unfollow” her.
Some reasons why Twitter users tend to unfollow others, according to Kissmetric.com (http://bit.ly/1ygP1X4):
- Tweeting about issues that are “off topic” and not what would be expected based on the user’s profile;
- not providing valuable content and information (helpful links, informative studies, quick tips, industry-specific news);
- being too personal and sharing too many personal details and experiences;
- posting offensive tweets;
- not tweeting; and
- publishing a whole bunch of tweets at once on more than one occasion.
When should Denise post her tweets?
Although your followers have different schedules and you will never know when they are actively using Twitter, there are some tools that can help determine when to post based on likelihood that your followers will see them. As you would expect, there are a variety of tools that charge a fee and have even more capabilities; however, for this blog I will focus on one of the free tools that provides some basic information. I took the image, below, from the free version of Followerwonk.com that shows when Denise’s followers are “most active” on Twitter.
Using this information, Denise could do a trial and see if her tweets get more impressions, etc., if she posted them during those times when her followers were “most active.” For example, she could begin by posting her tweets around 9 a.m., 11 a.m., or 1 p.m., then look to see if she did generate more of a response.
Other data available on Twitter analytics
By clicking on the “Audience” tab, you can learn a little bit about your followers: Gender, country of residence, U.S. state/region, and followers’ top 10 interests. According to a blog posted on Social Media Examiner, by viewing the list of interests, “you will know what areas your should focus on for both original content and the content you retweet” (http://bit.ly/1rQQVrq).
I have provided an image of my followers’ top 10 interests. I post about several different foods, business, and technology topics, which is reflected in the data below. If I posted specifically about wine my followers would be different and their top 10 interests would include more food and beverage related categories.
The case for posting your tweet more than once
If you follow some global brands on Twitter you probably see that they tweet about a product, event, and/or time in history more than once during a given time period. This is called recirculating a tweet and it is common practice. According to one blogger (http://bit.ly/1pKtwL9):
- Recirculating a “moderately successful post, one with about 2 initial retweets,” resulted in “about 2 more retweets.”
- Recirculating a “very successful post, one with an average 9 retweets,” resulted in it being retweeted four more times.
Reasons why you might want to recirculate a post:
- Some followers may be in a different time zone and miss your initial tweet.
- New followers may not see your tweet when you initially posted it (http://bit.ly/YOe7M8).
In addition, your original tweet could be missed if those who follow you also follow several other uses and/or users who post quite often. As an example, I follow 328 others on Twitter, many who post more than once each weekday morning between 8 a.m. and 10 a.m. I might see several hundred tweets and an additional handful of promoted tweets (ads). I’m certain that I miss some very interesting and important tweets.
While you may be concerned about disturbing your followers with these additional posts, consider this quote:
“Only sharing your content once on social media is a rookie mistake. Studies have shown that sharing your blog posts and other content several times will get you more retweets, views, and comments. As crowded as these networks are, it is unlikely you would bother anyone with multiple shares” (http://bit.ly/231zUML).
Finally, don’t recirculate every tweet you posts, but when you do change the words around, use different key words, include different #hashtags, change the image that you include in the tweet, compose your tweet as a question (for example, Did you know….), etc. (http://bit.ly/1RKIbjq). Look at your Twitter Analytics often, see if there are for any patterns regarding follower response, model some tweets after what you learn, and evaluate.
A special thank you to Denise Gardner for allowing me to show all her analytics in this post!
- Customer service
- Hard Cider
- PSU Wine & Grape Team members
- Sensory Training
- Students & Projects
- Team Members
- Wine Marketing
- Wine Marketing & Research Board