By Sysomos Inc.
When a tweet generates a reply (aka @) or a retweet (aka RT), it suggests the tweet has resonated enough with someone that it sparks a conversation or encourages someone to share it with their followers.
Given the power of the reply and the RT, we thought it would be interesting to explore how many tweets actually generate a reaction. We also wanted to collect more insight into the characteristics of these reactions. For example, what is the timeframe for a retweet to happen?
To carry out the research, Sysomos examined 1.2 billion tweets posted in the last two months.
We found that 29% of all tweets produced a reaction - a reply or a retweet. Of this group of tweets, 19.3% were retweets and the rest replies. This means that of the 1.2 billion tweets we examined, 6%, (or 72 million) were retweets.
We discovered that 92.4% of all retweets happen within the first hour of the original tweet being published, while an additional 1.63% of retweets happen in the second hour, and 0.94% take place in the third hour.
This means that if a tweet is not retweeted in the first hour, it is very likely that it will not be retweeted.
The graph below shows the fraction of tweets from the second hour onwards - the x-axis shows the time in hours since the original tweet, while the vertical axis shows the fraction of retweets within a particular hour. The 92.4% of all retweets, which happen within the first hour, are not displayed in the chart. 1.63% of retweets happen in the second hour, and 0.94% take place in the third hour.
When examining replies, there is a similar pattern to how retweets behave. We found that 96.9% of @ replies happen within the first hour of the original tweet being published, while an additional 0.88% of replies happen in the second hour. After that, reply activity dramatically declines.
We also examined the distance between an original tweet and the replies it attracts. Of all tweets that generated a reply, 85% have only one reply. Another 10.7% attracted a reply to the original reply - the conversation was two levels deep.
Only 1.53% of Twitter conversations are three levels deep - after the original tweet, there is a reply, reply to the reply, and reply to the reply of reply.
The chart below shows how the fraction of Twitter conversations (on vertical axis) with respect to how long the conversation lasted (on horizontal axis). It shows that only a small number of users actually have the ability to engage on Twitter in a significant way.
The table below shows the fraction of retweets as number of hours since the original tweet.
The table below shows the fraction of @ replies as number of hours since the original tweet.
The video below shows a sample visualization of tweets and their retweets over time. The spiral represents the time axis. Each blue dot is a tweet, with the size of the blue dot representing the number of retweets and replies to that tweet. Each green line shows a retweet to the original tweet and each orange line shows @reply to the original tweet.
From insights to business results.
Sysomos helps you get the true social intelligence you need to make better business decisions. Research without limitation, stay on top of what your community is saying, and advertise to people looking for what you’re selling.
Sysomos' Products Include
Sysomos MAP: MAP is like the search engine for social media. Listening tools give you snapshots of social activity, based on complex queries you must build carefully in advance. With Sysomos MAP, simply type in your keyword, and instantly see results, including reach and context for your topic today and historically. Unlimited searches, unlimited results, no surprise costs.
Sysomos Heartbeat: You wouldn’t run a website without web analytics, so your social strategy needs strong analytics, too. Know what your customers are saying, when a crisis is brewing, how healthy your reputation is, what products matter to your community and who is influential in your space. And it’s all real-time, because the social web doesn’t wait.