« A Tale of Two Talkers | Main | Digital Sustainability »

07/07/2010 How Good is Tweet Level at Predicting Influence?


MTV is apparently using the Edelman Digital tool Tweet Level to choose one of its next DJs. Their competition is using Tweet Level (TweetLevel) to gauge which contestant should have the greatest impact if chosen. While the tool is one of the elements used to gauge popularity, it's interesting that MTV is trying to get beyond things like number of followers (even if Tweet Level itself uses that measurement as one of its criteria).

TweetLevel isn't something new; there have been a number of different online tools to gauge Twitter power, such as Twitter Grader.  Actually, I'm not entirely sure why one is better than the other since in the tests I've done the results are actually quite close.

In any case, I decided to put Tweet Level to the test and compare its results with the voting on Social Media Day in Burlington. And no, I'm not doing this because I didn't win (as you'll see there was an unanticipated sleeper that was missed), but because I wanted to see if TweetLevel's automated system matched up with what people in Burlington actually said.

Here was the Burlington voting for Social Media Day:

King - Shay Totten
Throne pretenders - Rich Nadworny, Joe Mescher

Queen - Nicole Ravlin
Throne pretenders - Bridget Butler, Dana Freeman

Business - Small Dog Electronics
Throne pretenders - August First, Handy's Lunch

Non Profit - VPR
Throne pretenders - Echo Center, Cots

Here's who Tweet Level says has the most juice (higher is better)

King
@shaytotten 57
@rnadworny 54
@joemescher 47

Queen
@PMGNicole 55
@missmagpiefgs 51
@birddiva 49

Business
@hellosmalldog 50
@augustfirst 49
@handyslunch 43

NPO
@vprnet 51
@voicesvt 45
@cotsvt 32

If you look at that, Tweet Level predicted who should have won, and they actually did! So far so good. It worked.

What did Burlington Tweeps miss? On the King side, both Michael Martine and Mitch Lieberman scored 60, higher than any of the King finalists. On the Queen side Lara Dickson matched Nicole with a score of 55, a tie. On the business side, Magic Hat scored a Burlington high of 66 (!) but that may not be fair since they're a big brand, at least by Vermont standards.

In looking at this, maybe TweetLevel is a good indicator of influence. It's certainly something I'll keep my eye on and look to use in various clients' social media marketing

TrackBack

TrackBack URL for this entry:
https://www.typepad.com/services/trackback/6a00e5538e53f988340133f2209b94970b

Listed below are links to weblogs that reference How Good is Tweet Level at Predicting Influence?:

Comments

My Web Sites

Categories