Pathable : Event App Technology Blog

Data Analysis of Twitter at Events

I was puzzled. All the blog posts, webinars and tweets I was reading told me that Twitter and live events went together like peanut butter and jelly. I knew, in fact, that Twitter’s “big break” came at SXSW 2007, the annual Woodstock-Sundance of the high-tech / indie music / film worlds.

So how come when I actually read the Twitter history from various events, it seemed like mostly hiya chatter? In the Twitter echo chamber, this can sound like a raucous din, but in the cold light of day, are a couple dozen tweets from a handful of people really valuable?

I consider myself a Twitter skeptic and a data junkie. Don’t get me wrong: I use Twitter every day, I have three separate personae that I manage, and, in fact, I took a break to send a tweet while writing this sentence.

That said, I see a lot of questionable Twitterthusiasm out there, particularly as it relates to live events. I’ve attending more than my fair share of webinars that say they’ll discuss “5 Secrets for Using Social Media at Events” and read more than my fair share of blog posts purporting to describe the “innovative ways Twitter was used at X event”. Event managers (or technology vendors) will breathlesslly extol the value they saw in their wall-sized Twitter projections.

Sadly, so many of the “expert webinars” talk very abstractly about how valuable social media is, insisting that anyone who doesn’t use it is missing out, but fail to describe exactly what the measurable effects are. The blog posts and articles are similarly lacking in concrete facts or data: a few anecdotes about how people discussed the keynote in real-time and little else.

So I’m left asking myself: what exactly are people tweeting about as it relates to live events? How often? What value are they getting out of it?

Before I swerved into technology, I was involved in research psychology, and my motto was “show me the data”. So, hoping to answer my own questions, I took a stroll down data lane.

To better understand how Twitter is actually being used at a live event, I asked our intern, Jeff Losek, to analyze all 797 tweets associated with a particular event, Wordcamp SF. Wordcamp is a self-organized “unconference” centered around the use of WordPress, an open-source blogging platform (in fact, the platform on which this blog is built). The attendees (over 700 of them from 32 different countries) tended to be tech-savvy and relatively cutting edge. As such, this is a “best case” analysis of the use of Twitter.

Clearly, there are many, many events out there that would not see as much or as diverse use of Twitter as this event, but I thought it made more sense to analyze a “best case” event, based on the belief that where the cutting edge leads today, the rest of the world will follow tomorrow (or next year).

Methodology

We collected all tweets containing the event’s hash tag, #wordcampsf. Based on an initial analysis of the content of the messages, we categorized them into one of the following buckets:

  • I am, want to or can’t attend WordcampSF
  • Announcements or questions related to Wordcamp SF
  • I am preparing for or on my way to WordcampSF
  • I am at the event / a session
  • Talk to me or meet me if you are interested in a particular topic
  • Comments / quotes about a particular speaker
  • Here’s what I’m doing / feeling (not directly related to event)
  • Tweets directed at an individual
  • Retweets

Results

Here’s what we found:

Pie Chart of Tweets at WordcampSF event

A few interesting points jump out from this chart:

  • Tweets that are not directly relevant to the vast majority of event attendees (“Here’s what I’m doing / feeling”, “talking directly to someone else”) make up about 1/3 of the tweets sent.
  • Tweets that are useful to people who can’t physically be at the event (“Comments / Quotes about speakers”, “Announcements / Info / Questions related to event”) make up more than 1/3 of the tweets
  • Tweets that report people’s intended or actual location make up around 1/6 of the tweets (“Traveling to”, “At the event / session”)

This is actually reassuring. While there are a large number of tweets that I would consider “noise”, a signal to noise ratio of 1:3 is actually pretty good, especially considering the ease with which one can ignore “noise” in Twitter. Even the tweets that aren’t actionable or educational, such as people announcing their excitement around attending, have value: they advertise to non-attendees that the event is happening and to attendees that the sender is going to be a participant who would like to be included in discussions. We shouldn’t undervalue the impact of “buzz”.

Events have a lifecycle: attendees’ behaviors before the event is very different from their behavior while the event is going on is different from their post-event behavior. How do these differences reflect in Twitter use? In particular, is there value to the wall-projections of Twitter that we sometimes see at events?


These graphs shows how the different types of tweets changed over the event’s lifecycle (the event itself occurred on May 30th, 2009). The graph to the right shows the absolute volume of tweets, the graph above each category as a percentage of the total tweets for that day. For this analysis, we removed “Retweets” because the categorization doesn’t reflect the content of the tweets. Our take-aways?

  • The proportion of “noise” tweets are relatively low leading up the event, and only bloom after the event is complete (when the overall number of tweets decrease).
  • There is a steady stream of tweets with useful information or questions regarding the event that continue for several days post-event.
  • The types of tweets that it would be useful to see on a projection, comments about speakers or sessions, people’s location at the event, and annoucements or questions about the event, make up significant number of the overall tweets during the event itself.

I find this chart reassuring. It shows that although there is a background of chatter and noise, a significant amount of the communication going on over Twitter is useful and relevant, the type of information that I’d like to at least be peripherally aware of.

Our final question was “who is sending these tweets”. For this analysis, we wanted to see how the tweeting was distributed across people.

What do we learn from this chart?

  • While 258 total people sent at least one tweet, 20 people account for more than half of those. That’s consistent at a high-level with the “long-tail” notion of user-generated content (i.e., a large number of people contribute small amounts of content, but that content in aggregate accounts for a large proportion of the total content). The numbers, however, don’t fit cleanly in the 80/20 90/10 buckets that are often cited. Instead, it’s more like 50/50 (50% of the content is accounted for by a small number of high activity contributors, 50% by everybody else).
  • Most people tweet 6 times or less total about the event over the 9 days analyzed, less than a tweet / day.

On the whole, I find this data analysis curbs my skepticism. If I had to summarize:

Many people are tweeting about things relevant to an event. Prior to the event, they’re asking questions, requesting meet-ups and advertising their attendance and at the event itself, they are discussing the active talks and providing timely information about what’s happening at the event itself.

As noted early on, this is a “best case” example of how Twitter is used at an event, but it doesn’t seem overly optimistic to expect that these patterns will grow across the rest of the event industry over time. Given that Pathable has made some key investments in Twitter integration recently, including a “which of my Twitter friends are also at this event” feature and real-time Twitter updates from attendees, I’m encouraged that we’re on the right track.