Home | About | Log in | Get The Feed
eMetrics Twitter Backchannel Analysis

May

10

2008

eMetrics wrapped up earlier this week after 4 days of fun – of which I was only there for one day. Knowing that I would only be there for one day I set up a Twitter backchannel on Twemes.com. I’d seen Twitter used quite effectively at Media Re:Public and, of course, heard about its use at SXSW. Of course, since we’re analysts can’t do something new without some gratuitous analysis!

image

This chart shows the incremental contribution – on a percentage basis – of each Twitterer who participated during eMetrics. Note that I have removed myself ("Omomyid") from the data as I was the host and my behavior could be considered to skew the data.

As you can see, Bob Page was the top TwIt ;) with 25.6% (53) of all Tweets – way to go Bob!

There’s quite a large gap between Bob and the second rated TwIt – June Dershewitz who contributed 11.1% (23) of all Tweets. June was followed closely by Marshall Sponder at 10.1% (21).

There’s another small break in contribution and we get to Dave Rohrer and Eric T. Peterson who contributed 8.7 and 7.7% respectively.

Next, we have Vannesa Fox and Dean Burris adding 5.3% and 4.8%.

Finally, we have Rene Dechamps and Gradiva Couzin chipping in another 3.9% and 3.4% of Tweets and that gets us to 80.7% of all Tweets being driven by 9 individuals (36% of active #eMetrics twitterers.)

image

Here’s another take on the same data … basically just a chart of the raw counts. In this case it’s a little easier to see how more active Bob was than everyone else and the relative groups that were discussed above.

In addition a more striking grouping is visible in this chart:

1. Uber-Twitterers - From Bob Page to Eric T. Peterson, this group is responsible for 131 tweets covering 63.3% of all tweets. This group averaged 26 tweets per person, 17.9 per day and almost 3.3 tweets per person per day.

2. Engaged Twitterers – from Vanessa Fox to Jim Sterne, this group tweeted 43 times and contibuted 20.8% of the tweets. This group averaged 8.6 tweets per person 5.4 tweets per day and 1.1 per day

3. Casual Twitterers – this group, from Laura Forrest to Ian Thomas had 19 total tweets in 8 days. They averaged 3.8 per person, 2.4 tweets per day and 0.5 tweets per person per day. Casual Twitterers contributed 9.2% of tweets.

4. Sometime Twitterers – from u_m to Phil Sheard this group had a total of 8 tweets which works out to 2 per person and 1 per day. They contributed 3.9% to the twitter-stream.

5. One-Tweet Wonders – this is the largest group with 6 members each had just one tweet to #emetrics (2.9% of total) in the 8 days between May 1 and May 8, 2008.

So what’s all this telling us? Well, I’m not sure beyond the fact that there were discreet and obvious levels of engagement in the back channel.

I haven’t done a qualitative assessment of the tweets themselves – sure would to love to see that. But my general sense is that this backchannel was used more for status and social communication (e.g. ‘I’m at the lobby bar’) than as an idea space like I saw at Media Re:Public.

Bob Page image image image image image image image image image 

imageimage image imageimage image image image image image

image image image image image

John Peltier (See Comments) gives us this Zipf test chart on the distribution of tweets:

« Previous Post » Next Post

12 Responses to “eMetrics Twitter Backchannel Analysis”


“So what’s all this telling us?”

Your distribution of contributions vs. rank is described by Zipf’s Law (http://www.nslij-genetics.org/wli/zipf/index.html), which has been applied to web site traffic by Jakob Nielsen in Zipf Curves and Website Popularity. I’ve made a simple plot of this for you:


@Jon Peltier: Thanks John! I’ve added your chart to the body of the post.


Wow, analysis of anything! Seriously, thanks for setting up the backchannel. I noticed a lot of folks forgot (or chose not to) use it, especially when replying to tweets. But it was useful for me to get a sense of what was going on outside the realm of my own experience.

I had the advantage of spending Sunday attending the Industry Insights Day – where my cell phone had a reasonable network connection, and there was beaucoup twitter-worthiness. Most of my fellow Twitterers were only on duty Monday through Wednesday, many without a decent network/cell connection… a daypart analysis may bear out that hypothesis…

Can you talk more about the “idea space” you saw at Media Re:Public?


@bobpage: Bob, while your twittering definitely spiked on a couple of days, it was more distributed than you think. Because I wasn’t capturing the RSS in realtime it’s difficult to do the daypart analysis now.

One of my top recommendations to @jimsterne is better connectivity in the rooms.

As far as ‘idea space’ goes, check out Media Republic Tweetscan Basically, this idea space was lots of people tweeting what they heard during presos that they found interesting – e.g. note taking. It was a great way to see audience response to the sessions. I could have wished for more discussion (as was reportedly the case at SXSW) but still the many perspectives shared in Twitter were pretty darn cool.
I’ve never seen a more plugged in crowd which was pretty interesting since they were all journalists.

What’s interesting is if you go ahead and assume that Twitter is currently for early adopters and ‘influentials’ then the web analytics crowd must appear to be a very normal distribution of demographics. We had 26 (including me) contribute to the backchannel and even if there were double that actually tweeting about #emetrics that’s still somewhere south of the total population at #emetrics.

Just riffin’ here.


Bob brings up a good point – our ability to contribute was limited by the crappy network/cell reception at the hotel. This probably explains the “I’m at the lobby bar” reports – our phones actually worked there.

Anyhow, this was a fun experiment and I look forward to doing it again, hopefully with more of our web analytics comrades. I volunteer to host the backchannel at eMetrics DC this fall.

June “#2 Twit” Dershewitz


@June Dershewitz: Rock on! I’ll beat the horse again – @jimsterne needs to provide better connectivity in the conference rooms.

It’ll be nice to let someone who is actually there the whole time host.

Here’s a tip, if you plan to do analysis, use Excel to grab the RSS on a daily basis. I missed a lot of opportunities because I didn’t. Most importantly, time of day analysis and qualitative assessment. If I had been dumping the feed into excel on a daily basis I could have easily analyzed the content of the tweets – oh well!

I think the XChange format doesn’t really support this kind of communication, but what do you think?


To my visitor from Hawaii who has looked at this post 15 times since I posted it:

If there’s something I can help you with please leave a comment, email me (civy “at” instantcognition.com) or send me a message on twitter (@omomyid).
I’m happy to help — if I can.


[...] back-channel is always at work. In grade school it was a piece of paper that was surreptitiously circulated [...]


[...] back-channel is always at work. In grade school it was a piece of paper that was surreptitiously circulated [...]


[delicious] eMetrics Twitter Backchannel Analysis » Instant Cognition http://tinyurl.com/5f4fbm


[delicious] [from fjcapeletto] eMetrics Twitter Backchannel Analysis ” Instant Cognition http://tinyurl.com/5f4fbm

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>