A Night at the TwitterBowl: Successful but Unwieldy

Update: The project wasn’t as unwieldy as first thought, super analyst Josh Bernoff did the heavy lifting and has analyzed the results.

Well that was an amazing game, I’m sure my many colleagues at Forrester (Cambridge, near Boston) are throwing fits, but you’ve got to hand it to all the players that was seriously a great game.

A recap and you can view all responses
The social media experiment went very well, there are over 2500 responses to the superbowlads account. I spent over an hour hand copying all the replies on that account to this spreadsheet of all the responses. This is a read only spreadsheet where you can do a search (let it load, it takes time) and see how people are talking about the different brands and the ads.

The twitter application held up ok, although many of the replies did not show up on the replies page in real time, you could use the search tools to quickly see what folks were saying.

Massive Volume
It was pretty amazing, every time I refreshed the search tools (terraminds or twittersearch), new responses would appear in rapid order. There were so many responses coming in, (about 625 per hour, or 10 every minute) it was really hard to keep track. I tried to summarize key findings (such as many folks liking X commercial or hating Y commercial), but it became difficult to track.

Track your brand, or commercial
If you work for a company (or you have a client) that advertised on the superbowl, you should be doing searches using the twitter search tools, and add “superbowlads + brandname” to find out what people thought in real time. For example, I know the Dell blogger team is keen to knowing what we all thought, you can check this query of superbowlads + dell to see what folks rated it, it probally wasn’t as positive as they would have hoped.

For Example, I’ve scored the Dell advertisement

  1. I found all the responses using a query
  2. I tallied the responses on a spreadsheet: 27 of them
  3. I added them all up for a total of 70
  4. Divided them
  5. Dell Ads received an average of 2.59 as the average response
  6. The savvy advertiser will read all the responses and try to underhand why it scored that way

You could even query by user name and see how they rated content, Josh Bernoff was very consistent in his ratings, you can see all his scores.

Please analyze
I hope some smart advertising, interactive, or measurement company exports the data and does some formal analysis. I’m tapped out now, and hope someone else takes the lead

New tools needed
Next year, I hope there’s a real time polling toll that we can use, that will auto create data and reports, so I don’t have to do it manually. If anyone has a better solution for future, please let me know. In any case, it was clearly a success, as we were all chatting, laughing, and critiquing the experience in real time –location not being a factor.

The web is now a live opt-in focus group.

Who’s talking about the impacts of Twitterbowl:

I’ll update this over time, leave a comment if you see others talking about it

  • Essential keystrokes participated, but felt the ads were overall very weak this year.
  • ShinyRed Recaps the impacts of the game, and twiterbowl on society.
  • Some say this was a landmark media experiment
  • Chris Heuer interviewed me on why I did this experiment: