The 2012 presidential election demonstrated better than any previous election just how important good email tactics are to a successful campaign. Some people have gone so far as to avow that it was email that won the presidency for Obama. Whether that is true or not, one incontestable fact that emerges from the campaign is that Obama’s team spent a lot more time and effort taking advantage of the most important aspects of any successful large-scale email campaign—testing and metrics.
The Obama campaign is a testament to the power of split testing and report analysis. They tested nearly every aspect of each email, from the subject line to the content, and even the formatting. In some cases, they tested as many as 18 variations before choosing one to send out to their subscribers. “When we saw something that really moved the dial, we would adopt it,” said the campaign’s email director, Toby Fallsgraff.
They tested everything from mild profanity (“Hell yeah, I like Obamacare”) to extremely terse subject lines (“Hey,” “Wow,” “High five!” “Me again”), to the downright weird (”Sometime soon can we meet for dinner?” “It doesn’t need to be this way,” “You must be at least a little curious”). They even tried “No subject,” which yielded strong results the first time they used it, but was largely ignored after that. Out of all of these, “Hey” yielded the best open rates. Perhaps ironically, or perhaps not, the subject line, “I will be out spent” was the most successful at generating donations, pulling in almost three million dollars.
Obama’s director of digital analytics, Amelia Showalter, was often amazed by which emails yielded the best results. The crew in the marketing department eventually started to bet on which ones would perform well and which would flop. Most of the time they were wrong. It became apparent to Ms. Showalter that trying to run a campaign without split testing would have been a study in futility.
When something worked, they would use it for a while, but usually the novelty wore off quickly and the crew would have to go back to testing to find the next winning Subject Line. The proof of the effectiveness of this technique shows up in the open rate results. The Obama emails had more than twice the open rate Romney’s emails.
The biggest complaint from nearly every quarter was the astounding amount of email sent by the Obama campaign, sometimes as many as ten emails per day. It shouldn’t surprise anyone, then, that more of the Obama campaign’s email got marked as spam—5% compared to Romney’s .08%. It undoubtedly helped Obama that most people recognized that this onslaught of email was a temporary thing, guaranteed to end after November 6th. A retail store trying the same tactic would certainly face a much higher complaint rate. Nonetheless, Toby Fallsgraff paid close attention to the metrics throughout the campaign, making sure that deliverability wasn’t affected. This was no “spray and pray” operation.
Another unexpected hit was profanity. Dropping in mild curse words such as “Hell yeah, I like Obamacare” got big clicks. But these triumphs were fleeting. There was no such thing as the perfect e-mail; every breakthrough had a shelf life. “Eventually the novelty wore off, and we had to go back and retest,” Amelia Showalter said.
Most of the news article focused on the Obama Team’s use of the A/B split testing of subject lines, but that was only part of the story. They were also using segmentation and dynamic content to make the emails more personal. “There were lots of interesting behavioral and personal markers that we used to make the content more dynamic and relevant for our supporters,” Ms. Showalter said. “[We added] language thanking people for previous volunteer service when putting out a wide call for future volunteer service, for instance. And there was a lot of personalization in the fund-raising emails, of course.”
When it comes to email marketing, Ms. Showalter acknowledged, it is best to leave no stone unturned. “We pretty much tested everything!” She said.
2. The Race for the Inbox, Return Path infographic.