Dec 10 Posted 7 years ago Thanks for the information. Very useful and thought provoking. It does seem to confirm a truth in communications that we have learned in the non-profit sector that is becoming even more relevant today with the prevalence of social media. Studies show that while negative communications, built on fear for example, may generate some initial awareness and get noticed they generally do not produce sustainable changes in behavior. Positive communications build on attainable rewards and benefits work best in this regard all the time. The twitter analytics you discussed seem to indicate that on an initial surface level people would rather engage with positive proactive peeps and tweets...something that asset-based thinkers know as the "laws of attraction". Look forward to more of your posts. @hankwasiak
Dec 7 Posted 7 years ago @Carl Vervisch
There are many problems with Twitter accounts : Spam accounts, celebrities and politicians that have many thousands of followers because of who they are to name a few. These users were not part for the analysis although in some cases it is hard to identify which Twitter users have auto-follow bots. Most users tend to un-follow users that do not give valuable information. You could also analyze this information based on the follower-to-following ratio and not strictly follower count (which is also discussed in one of my other posts).
Note also that i do not imply that by just using these words you are going to draw many followers. To do this you need many other things and this is explicitly stated towards the end of this post.
Dec 7 Posted 7 years ago @Ari Herzog
Good point. Although this text mining exercise did not take into account the fact that 70% of Twitter users quit shortly after they create an account this problem could possibly be solved by checking the date of the last Tweet for each of the followers.
Dec 7 Posted 7 years ago @joel foner
Any causal relationship is by no means implied (notice how many times i use the word appears to the text). Please also notice the phrase "Use as possible clues only"
What you say is correct..this is merely an example of what kind of insights could possibly be extracted from Twitter. Please check lifeanalytics for more examples.
Dec 6 Posted 7 years ago Great research, introducing me to some concepts I hadn't considered, but isn't there a fallacy to correlate popularity with followers at a time when 70% of new Twitter users quit in the first 30 days? Therefore, who cares how many followers one has if x% are inactives?
The better metric is to look at number of @ replies and/or retweets within a certain time frame.
Webinars On Demand
April 05, 2017In the ever-changing world of digital marketing, operational efficiency, quick turn-around times, testing and adapting to change are crucial to...
January 25, 2017While we’ve had access to the Internet since the early 1990s, we’ve only started to experience the full effects of its disruption on public rela...
Video is expected to account for three-quarters of all mobile traffic by 2020. But, creating powerful, effective video requires a significa...
Marketers are constantly seeking to engage with their buyers and drive actions that help buyers move rapidly through their customer lifecycl...