Fundraising and the Obama Campaign

We basically found out our guts were worthless. From Patrick Ruffini at #roots12.

  1. RT @nathanhjb: Zombie Myth #1: Email is dead. #2: Winning FB means winning at the polls. #zombieroots #roots12
  2. OFA Digital team started out as 12 people, grew to 200. #roots12
  3. Outbound Team: 18 people full time on email. 5 people full time to content for the website/blog. 4 people dedicated to social. #roots12
  4. "Smart data and randomized experiments gave our side a tool for victory the other side didn't have." #roots12
  5. Persuasion-1, moderate support high turnout. GOTV strong support, low turnout. Unknown: new registrants #roots12
  6. DCCC did its own modeling because DNC went to OFA. Catalist did. #roots12
  7. DCCC would do microtargeting calls (around 1500 voters) to build models. #roots12
  8. Catalist had an I/O tool to translate polls into data store without a human #roots12
  9. We rescored 45 models in a day to capture new registrants. These were our most important votes. #roots12
  10. Average national email featured 4 different messages with 3 different subject lines each. 18 segments per message, on average #roots12
  11. "I will be outspent" email raised $2.6 million for the campaign [Best email I ever sent had a similar theme.] #roots12
  12. "Our North Star was optimization" Sent to "more than 10,000 segments. <- WOW #roots12
  13. 3-to-1 increase in conversions from QuickDonate members. We always sent them the optimized version. #roots12
  14. We never tested to the West Coast or non-donors because they weren't awake or they didn't respond #roots12
  15. Media reports they were testing email on different demographic segments were false. They were testing on performance/tactics. #roots12
  16. Crossposting email to the blog was incredibly effective. #roots12
  17. The website was too fundraising focused, entirely by design and entirely intentional. #roots12
  18. "Can we get away with sending less email?" "Mixed results." Sending more email? Ka-ching. #roots12
  19. Note: The $3 and $5 asks were ONLY sent to non-donors. So they weren't leaving money on the table w/ donors. #roots12
  20. We could get valid data on segments of 17,000 to 20,000 people, but much bigger for a less responsive list. #roots12
  21. Very rarely did the top of the campaign interfere in the email program. Messina didn't say, you must send this. #roots12
  22. Emailer was not necessarily built to scale at this level. #roots12
  23. Histogram showing distribution of model changing. We could see Republicans moving but not Democrats. #roots12 http://pic.twitter.com/EoiI0o3Q

Did you find this story interesting? Be the first to or comment.

Liked!

Ben Domenech

Editor, policy fellow, throwback. You can tell a man’s vices by his friends, and his virtues by his enemies.

Total views
303

Storify

@Storify