Out of the three keynote speakers attending SES Chicago 2009, I was looking forward to Dan Siroker’s keynote the most.  Siroker is the former deputy new media director for the Obama presidential transition team and considering the location of Search Engine Strategies, I could only imagine the kind of information and footage that would be expected.

Simply put, I did not want Siroker’s keynote to end.  Great information and amazing footage from the Obama campaign.

This SES Chicago recap will cover the keynote address by Dan Siroker.

My recaps will include some useful tidbits that I picked up during day 1 of SES Chicago.  Tweets were made from @marketwire as well as my personal handle @shinng using the hashtag #seschi and #rt2eat.  So what is the meaning of the latter hashtag?  Well, the creative marketing team @marketwire wanted to give back for the holiday season so we decided to take advantage of the Marketwire presence.  For every ReTweet of our tweets that include the hashtags #seschi #rt2eat, Marketwire will be donating $1.00 per RT for a max contribution of $500 to benefit the Greater Chicago Food Depository.

Opening Keynote with Dan Siroker, former Deputy New Media Director, Obama Presidential Transition

The title of Siroker’s presentation is “How We Used Data to Win the Presidential Election”

Dan Siroker started the keynote with his background and how he got involved with the Obama Transition Team.  Siroker previously worked at Google as a product manager for Google Chrome.  What inspired him to jump ship was when Obama visited Google in 2007.  A couple weeks after the visit to Mountain View, CA, Siroker took a leave to come to Chicago to help Obama in any way.  After going back to Google for a couple months, Siroker came back to Chicago to lead the analytics team.  This position eventually led to working on the presidential transition.

Below is the Obama organization chart.  Dan was part of the analytics team under the New Media division.

Next up, the stats.  We all remember how Obama whooped McCain in virtually every single category right?

• Popular vote
  • Obama=53%
  • McCain=46%
• Electoral vote
  • Obama=365
  • McCain=173
• Facebook friends
  • Obama=2.4 million friends
  • McCain=0.6 million friends
• YouTube views
  • Obama=just under 100 million views
  • McCain=20 million views
• Unique website visitors
  • Obama=just under 150 million unique visitors
  • McCain=30 million unique visitors
  • Obama had 2x the number of unique website visitors than people voted for McCain
• Money raised – money is easy to measure so this metric was used.  Also, most of the money was raised online.
  • Obama=$656 million
  • McCain=$201 million
  • The money Obama raised offline ($156 MM offline) was almost equivalent to the money McCain raised altogether ($201 MM)

Lessons Learned

1. Define Success –>Define quantifiable success metrics
   • Defining success was getting Obama elected as president
   • (cost per click) –> website –> (sign up rate) –> email sign up ($ per recipient) –>raise money
   • Success can be broken down by CPC, sign up rate, dollar per recipient
   • Siroker noticed that they were doing a good job getting website visitors at a favorable cost, but a poor job of getting those visitors to sign up for email updates.  However, they were doing a good job of getting the email recipients to give money, so there was an opportunity to improve email sign ups.
2. Question Assumptions – this is where it got really interesting.  Siroker showed a multivariate splash landing page experiment where they tested the call to action button and the media to see which worked best.  They experimented with lots of attributes, buttons, content, and media on the page.
   • 
   • 
   • So which button and which media won?  Button winner=”learn more”.  Media winner=”family image”
   • They used Google Website Optimizer for the multivariate testing.  By using the winning combination, conversion went up from 8.2% to 11.6%.
   • 
   • 
3. Divide and Conquer – another multivariate test was done for the donation button.  Below is the chart for the test.  The actual results were scattered and none were 100% right.  So what was the point?  It all depends on where you came from and who you were.  It is all about the metrics.
   • 
4. Take Advantage of Circumstances
   • During a Clinton speech at the campaign headquarters in August 2008, Dan began updating their splash image to mention Clinton.
   • Another example was when Sarah Palin mocked community organizers.  The team sent an email to their list using the mocking as a way to raise money.  They asked for $100 donations to prove that community organizing is not something to be ashamed of; they raised over $10 million within 24 hours.
   • Below is an email sent within the organization to Siroker
   • 
5. Turn Your Customers into Evangelists – they did this everyday on blogs, forums, and videos.  The team focused on supporters, not the candidate.  This motivates others to get involved.
   • Siroker proceeded to show what he thought was the most effective video of the campaign.  I swear I heard some sniffles for those watching it for the first time.  The reason he showed this was not because it was so moving, but to show that users typically rally around a cause.  This motivates them to volunteer, donate, and vote for a specific campaign.
   • The idea of social media marketing as it relates to the Obama campaign was that users were connecting and rallying with each other, bringing other people to the cause, bringing in donations, volunteers, and getting them to vote for Obama.
   • Siroker compares social media marketing with affiliate marketing and proceeds to explain “social affiliate marketing”.  The model is to rally people to support a cause and give money based on that cause to a business.  It is a transparent relationship because the affiliate (or cause) gets money from the business and users interact with the cause.
   • 
   • 

Q&A

Q: How did you use analytics to integrate your marketing campaigns across channels?

A: They didn’t have time for it.  Siroker wishes they could have done better, but time was a factor.

Q: Did you do demographic segmentation?

A:  They never did experiments to segment.  To clarify, Siroker says that the challenge was to figure out what worked before they signed up, not after.  So in other words, they didn’t look at the information after the user donated.

Q: Obama had a lot of visitors.  Was the number of visitors directly proportional to the amount of fresh content provided?

A: No, Siroker does not think so.  He thinks that the difference was the persuasiveness of its content and that Obama’s policies attracted tech-savvy individuals.

Q: What was the most important factor in deciding what would appear on the campaign home page and landing pages?

A:  There wasn’t one.  That’s why defining goals is so important.  They asked themselves what were the things people cared about in a campaign.  Messaging was key.