To subscribe to the monthly C&E email newsletter and event announcements click here.

Low-propensity voters are an elusive target, but they’re often the margin for victory. Could sales techniques on social media be applied to increase LP voter turnout?

A team of inside players and outside groups experimented with a small budget on Facebook in the recent Virginia election. Here are some of the lessons learned by DemLabs and Local Majority, a Democratic PAC.

We discovered that it helps to use multiple targeting methods. We started with a list of LP voters from Catalist. This included people in four Virginia districts who had voted in 2016, but not in the prior gubernatorial cycle in 2013 election.

This list was supplemented with more names using Facebook’s “lookalike audiences” feature. We wanted to focus our limited funds on the neighborhoods that really mattered. To do this, we targeted our social media campaign to certain zip codes in the district that we found through the Statistical Atlas.

Here are some other tactics we used to get LP voters out on Election Day last month: 

Leverage local knowledge.

We needed to identify the hot-button issues that would resonate with voters in order to increase click-through rates. Local campaign advisors knew the terrain and proposed a few issues that would get people’s attention. This reduced the time and cost we needed to fine tune our social media campaign. The team decided to focus on the lack of infrastructure improvements causing long commutes and horrendous traffic jams.

Simplify the call to action.

Initially the plan was to have people donate to the campaign and request an absentee ballot. The gospel in e-commerce is to not distract a potential customer with options. The call to action has to be clear and measurable. We simplified our appeal to request an absentee ballot.

Optimize the ad design.

We designed over a dozen ads around the theme of requesting an absentee ballot would allow voters to send a message that something needed to be done. These ads had multiple combinations of images and copy and followed Facebook’s ad design recommendations. We wanted voters to see our ad at least five times in order to make a stronger impression, so having different flavors of the ad helped.

Experiment and refine.

We weren’t sure which would work the best in which district so we experimented extensively. Different ads were shown in different districts at different times of day to see which performed the best. The better performing ads got additional exposure while the laggards were dropped. Think of it almost as an online version of the Survivor reality series. This continuous experimentation paid off and we were able to reach over 150,000 unique individuals at a cost of under $11/CPM (cost per 1,000 impressions).

Use video-based messaging.

Videos perform extremely well on social media campaigns. We decided to leverage some of the excellent video content produced by the The People PAC to motivate LP voters to turn out. The combination of compelling video with the precise targeting of social media is a powerful combination. We measured how many people saw the video and how many watched it all the way to the end. It cost less than $0.05 per viewer to have over 87,000 people watch the video message.

We learned that a small budget can go a long way of social media, but it requires skills that small campaigns often lack. The answer is simple: work with an experienced partner and pay close attention to the changing dynamics of the campaign to experiment and fine tune the messaging as needed.

Deepak Puri is a Silicon Valley veteran and co-founder of Democracy Labs, a San Francisco-based non-profit that connects technical and media experts with progressive campaigns.