What factors affect the rate of audience growth on Twitter? To answer this, we tracked 500 Twitter users over 15 months. We examined factors related to three categories:
- social behaviors (e.g., interactional communication choices that a user makes)
- message content (e.g., writing style and linguistic cues)
- social network structure.
Our data revealed the following eight habits of highly successful Twitter users:
- Don’t whine online. This means tweeting content that is more positive in nature, rather than negative (including swear words). Negatively-oriented content will often be a turn-off to a potential new follower who is assessing whether to make a connection with you (exceptions for when it is used in conjunction with humor, inspiration/education, or controversy).
- Talk *to* people, rather than talking *at* people. Employing directed communication strategies (e.g., mentioning other users in your tweets, retweeting others (with attribution), and replying to & favoriting others’ tweets), rather than broadcast communication strategies (which do not target anyone in particular) will help make you more visible and personable – both of which will help you attract new followers. Having engaging interactions with your existing followers also helps you leverage your extended social network in order to become visible to (and hopefully appealing to) the followers of your followers, as well as those people who have friends or followers in common with you.
- Be informative, rather than “meformative”. The overwhelming majority of connections on Twitter fall into what we call a “very weak social tie” category. In other words, for many Twitter users, Twitter is a social network made up mostly of connections between virtual strangers and weak acquaintances rather than very good friends. For these kinds of ties, details about the mundane minutiae of your everyday personal life (like what you ate for breakfast, the outcome of your daughter’s soccer game, etc.) are much less attractive than timely or novel bits of news.
- Don’t abuse hashtags. Hashtags serve a very useful function; when used as intended, hashtags help to signal keywords within tweets that are related to a broader public topic, conversation, or group or to express humor, excitement, sarcasm or other contextual content. On the other hand, hashtags are more difficult to read, especially when the tag contains more than a single word (e.g., #occupywallstreet, #libyafeb17). So when you combine the readability issue with the fact that it is sometimes tempting to #spam #with #hashtags (i.e., over-tagging a single Tweet), then it is no wonder that many micro-bloggers feel that excessive #hastagsareannoying.
- Use more sophisticated writing. People rely on linguistic cues like spelling and vocabulary to compensate for the lack of traditional contextual cues available in face-to-face settings. When deciding whether or not to follow a virtual stranger, Twitter users seek out well-written content over poorly written content.
- Be clear about who you are, and what you’re about. Completely fill in all the parts of your user profile. Again, in the absence of face-to-face interactions, it’s about sending the signals that indicate you are a real person with real interests. Having a personalized photo, something about your geographic location, and listing a website are helpful. Your profile description should also indicate what it is you will likely be tweeting about – the richer the details in your description, the better the results for attracting new followers.
- Tweet more, and don’t go too long between updates. It’s about visibility and engagement! The more you tweet, the more visible you are. Most of the users in our data set tweeted less than 8 times per hour, but some went days and weeks between tweets. Accounts with long periods of stagnation are less attractive than those with up-to-date content.
- Follow-back. Reciprocity – paying back a new follower by following them in kind – is a useful strategy because it reduces the likelihood that an existing follower will un-follow you.
For more, see our full paper, A Longitudinal Study of Follow Predictors on Twitter.
C.J. Hutto (@cjhutto), Georgia Institute of Technology
Sarita Yardi-Schoenebeck, University of Michigan
Eric Gilbert, Georgia Institute of Technology