The Alexandria, Virginia mayoral primary, slated for June 12 this year, was going to be hot. Everyone who followed politics here in the Port City — home of Silverback Strategies since its inception in 2007 — knew that.
Unfortunately, not many people follow politics here in the Port City.
"I’ve lived in Alexandria for 12 years and, like many people around Washington, have enjoyed paying roughly zero attention to local politics," Andrew Beaujon wrote with admirable honesty in the May 6 edition of The Washingtonian. "When you live in a place where the schools are decent and the trash gets picked up, it’s just so easy to justify being a low-information voter — heck, it’s practically a local right."
Only 17 percent of Alexandria residents cast a vote in the 2015 primary. Turnout would obviously be a critical factor in the 2018 primary.
So what do you do, then, when beating the incumbent likely depends on increasing voter turnout — and not just any random voter off the street, but the people more likely to cast a ballot for you?
That’s the question Silverback faced when we worked with the challenger this spring.
Getting Out The Vote
Support for our client skewed towards the 18-to-45-year-old demographic, so we made the decision to run the digital campaign solely on Facebook.
Why? Facebook's powerful suite of awareness KPIs, the ability to run interactive ad experiences within the ecosystem and the overall reach of the platform for this particular audience (79.5 percent penetration among 18-49 in the US) made it the ideal choice.
Facebook, however, did pose some real challenges — particularly in light of the controversies surrounding political advertisers, data misuse, and foreign interference in another recent election you may have read about (oh,and this, too).
This was what the Facebook ad environment looked like as we rolled out these campaigns:
- Ads with political content were being automatically detected by Facebook and paused.
- Political advertisers were required to have their identity verified by uploading copies of their identification, sharing personal information like Social Security numbers and uploading a code sent to a physical mailing address in order to be verified as "real."
- A “Paid for” disclaimer was automatically appended to ads.
- Estimated reach for custom audiences is no longer available.
We were able to work around these issues — getting our team members verified and taking delays into consideration — but Facebook's rollout of its own guidelines was a little wonky, the review process was painfully slow at times, and we had to factor this uncertainty into our work while the clock on the campaign kept on ticking.
We weren't the only advertisers to experience these issues, but that didn't make us feel better when we were literally waiting for the mail carrier to arrive so we could run our digital ads.
The Value of Traditional Research: Precise Insights Inform a Personalized Campaign
Through polling provided by the challenger’s camp conducted prior to the campaign, we knew the following:
- The incumbent mayor was slightly favored (46 percent of voting intent vs 43 percent).
- Although the challenger was vice mayor and active in Alexandria government and politics, the incumbent mayor had name-recognition advantage in the city.
We also knew that the most important issues for residents were:
- Schools/education/teacher pay (21 percent)
- Affordable housing/low income housing (17 percent)
- Traffic/transportation/parking (16 percent)
- City development/urban planning/growth (13 percent)
We understood that the primary would be polarizing — 54 percent of the individuals surveyed were already certain of the candidate they would vote for.
However, when presented with information about the challenger’s positions on local policies and his City Council voting record, his favorability significantly increased — especially on issues related to affordable housing or education.
Finally, given the limited turnout for local elections, we knew that we'd have to do two things:
- Teach potential voters who the challenger was;
- Get them to come out to vote on Primary Day.
The best Facebook strategy, therefore, would be to increase awareness of the candidate first, and then convince residents of the importance of the election by focusing on the local issues that matter most to each one of them.
Generating Awareness with Facebook’s Innovative Ad Experiences
Creating Initial Contact Through Video Ads
Six weeks prior to the election, we launched the first salvo in our video ad campaign — a short, fast-paced, one-minute video with excerpts from the candidate’s campaign speech.
Over the course of this phase of the campaign, we reached approximately 76,000 locals — that's over 65 percent of Alexandria's voting population — over three times each while driving an estimated 14 percent ad recall.
In other words, roughly 10,000 people could remember the candidate for several days post-video view. That’s nearly the same amount of people who voted in the 2015 primary.
Despite consistent increases in Facebook CPM driven by growing competition for the inventory among advertisers, video still proved to be an efficient way to generate awareness. Our cost per 10 seconds view was generally maintained in $0.15 to $0.20 range.
A Comparison Ad Tailored for Digital: The Power of Canvas
Comparison ads are an essential part of traditional political campaigns and present an opportunity to showcase the candidate’s view on local policies compared to an opponent’s.
In phase two, we used Facebook Canvas ads to lay out the stark differences in public policy between our client and the incumbent. We used direct quotes from meeting minutes and major local publications like The Washington Post to clearly show how the two candidates were approaching issues important to voters from different angles.
- 59 percent view percentage
- 30 seconds spent in the ad unit(!)
- Between $2.80 to $3 per open
We were very happy with the ad performance and, anecdotally, people began referencing the Canvas ad in conversations with the candidate and his team.
Getting Specific Through Hyperlocal Targeting
As local debates were heating up and candidates getting more media exposure, we switched to the third phase of the campaign: hyperlocal messaging.
We interviewed thought leaders such as previous mayors and the chairman of a local preschool on key topics to produce short endorsements of the candidate. Specific audiences were then targeted with the custom messaging, while Facebook delivery was optimizing for engagement.
We saw our cost per engagement sink as low as $0.12 — an investment the grassroots campaign could comfortably afford.
Election Day: Mobilizing the Base & Getting Voters to the Polls
Two days prior to the election, we began the fourth phase of our strategy and started running a message around the importance of getting out the vote. Each resident could make a difference!
Four audiences were prioritized:
- Residents previously reached by the awareness/ hyperlocal campaigns.
- Residents who had either viewed the campaign video, read our Canvas ad or engaged by commenting on the candidate’s organic posts.
- Residents with favorable voting intent based on the canvassing data. These individuals indicated they were favorable to the candidate during his canvassing campaign.
- The candidate’s base on Facebook — the Alexandrians who'd Liked the challenger.
Reaching Voters From the Previous Primary
Using public voter data from past primary available to all candidates, and coupling that info with Facebook’s powerful audience tool, we were able to match nearly 70 percent of past voters on Facebook and specifically target these folks with relevant reminders to go to the polls.
The Final Result? A Come-From-Behind Win
By using favorability data from the canvassing campaign, we were able to estimate that our client would get between 53 and 54 percent of votes — if everything went according to plan.
By midday on primary day, preliminary results showed higher turnout in precincts with a broader share of the young families that our client needed in order to win.
The final result: our client won, and our projections were pretty close. The tally at the end of June 12: The Challenger, 52.87 percent; The Incumbent, 47.13 percent.