Simulating a Tweetstorm

A "tweetstorm" is a high volume of tweets about or around a certain topic. We often want to simulate a tweetstorm to see how crisis teams will deal with the information overload. 
The key to creating an effective tweetstorm is to simulate the correct timing, volume and mix of tweets.

The chart below shows an example profile of how a tweetstorm might build from an initial incident. Note that there could be some delay between the "event" and the reaction to it depending on how many witnesses there are likely to be. If it's an internal problem, will the public be immediately aware of it?


The building of the storm can be considered in four phases:

  • Phase 1 - Early witnesses. How and from where does the incident first become known to us? Call from security? Broken website? Tweet on social media? Details might be sketchy a this stage. We know something has happened but maybe that's all. Obviously depends on the scale of the incident. A bomb at a shopping centre is going to have more early witnesses than a coach crash in the mountains.
  • Phase 2 - Breaking news. The press pick up on the story and start tweeting "BREAKING" type tweets. Details could be scarce or contradictory. 
  • Phase 3 - Floodgates open. Because more people will be following the press than some random bystander (assuming they're not an "influencer" of course) the trickle of tweets becomes a flood as the world wades in with opinion, conjecture, emotion and lots more. See the table below for the types of content to consider.

  • Phase 4 - Morphing. The incident becomes an opportunity for people to express other concerns and dig up old grievances. What started as one thing becomes a lightning rod for other issues.
With Conducttr TeamXp, these phases of content can be built using the Pattern of Life feature.
The diagram below shows how our Pattern of Life content wraps around key events that are usually represented by a master events list (MEL). 
In the case of simulating a tweetstorm, the phase 1 and phase 2 content might typically be scripted into the MEL as this is key information pertinent to the scenario. However, we can then allow exercise control the freedom to decide how much of a tweetstorm should build and how it should look; and this can vary on a team-by-team/syndicate-by-syndicate basis.



The image below shows a pattern of life "deck" for a recent client exercise. Each button on the deck represents a stack of content that can be published "on demand" with the timing between content varied button-by-button.
The buttons marked with "i" represent the background hum of twitter - celebrity nonsense combined with more scenario-relevant tweets. So far though we're just making the world feel alive, the storm has yet to build.
In this exercise the official press reaction is all on the MEL and so we have buttons for initial public reaction to the incident ("1a") followed by disinformation (3) about the event. We then have an influencer (4) hijack the incident by raising his pet issue and this adds additional tweets around that topic.
In this exercise there are 4 syndicates and the deck has been configured to allow exercise control the ability to load teams separately with additional retweets of key messages and by introducing additional tweets from known industry "commentators".
By mixing the content stacks with our pattern of life deck, players get to experience a very rich and realistic simulation - and this rich information environment will be potentially be different for each team.


Check out the video below for Pattern of Life in operation