Coming out of 2016, the biggest thing that stood out to my bosses was how much of a gap there was between digital Democrats and Republicans. Gap had closed online but GOP was ahead of Dems. Party kept doors open for 2018 cycle. We didn’t just want to close the gap, but to diagnose it and why Dems were underinvesting online.
3 biggest lessons: creative, media buying and budgeting.
Creative testing is a requirement, not an option.
Impacts the effectiveness of the program. Will continue to do. What works in poll may not work in real world. No silver bullet in creative. We’ve gotten good at constructing polls to say what we want them to say but also doesn’t reflect how these messages will be communicated in real world. Hurting digital more than other mediums. Polling needs to better reflect digital needs. Add a step, produce concepts that will work in real world. Leaned on panels not in-field.
Testing has an incredible amount of ROI. Creative testing is supposed to help us eliminate ads that will do harm. Def happening. 12% of ads tested had no effect, did harm. Added benefit, we’re picking ads better than average ad.
We had a robust, creative team not just from the political world. Encouraged them to put out creative content that wasn’t traditional political content. Most important rule is there isn’t a silver bullet. Storytelling is a good way to communicate a message. Saw experimental content that worked and content that didn’t. Next cycle will gauge how humans can predict these results.
GOTV creative was led and produced by a black woman and it showed. Did better in testing, speaking to voters who looked like us. Staff all our departments.
Budgeting for Digital
Don’t base digital budget on percentage of overall paid media budget. Base it on reach, frequency, audience. Are there voters who are more online or not online that we should consider not budgeting for or over/under budgeting for? Used persuasive messages to measure recall. Found very simple takeaway: everyone is online. Every single voter is reachable online. In 2012/16, what about old people? They are online. Reachable.
Media allocations, we used TV market efficiency to guide us. Who are the cord cutters, what is their TV behavior, how to change budget. How does efficiency of a TV DMA change our allocation.
Staff and How to Scale
Scaling digital persuasion with the right staff. Digital marketing is complicated. Not straightforward. Looking at political ecosystem, difficult to navigate. TV is a mounted device on a wall, digital is way more than that. Not a curse but a blessing. It can solve goals that other mediums can’t. Allows us to reach people every point in the day. Targeting offers options but not scalable across every platform. View it as a luxury, value of search as a platform for changing minds. Requires a lot of talent and time.
In 2017, I wanted to build capacity inhouse. But given the turnaround, didn’t feel we could build a diverse and well-managed team. Instead of doing so, we built a training program instead. Now we have flexibility to build more capacity inhouse.
We ran a program that is my favorite. Could we run a simpler media plan and cause similar impact? Are FB ads enough? Muted video is not enough. Took two fellows, gave them $100K, less-competitive congressional districts and trained them. Four months embedded with team, comms, analytics, design. We gave them tools for creative, polling. They put together a media plan that was FB and Google. Held out half the district. Found on initial results, moved about a point. You can train new staff on how to think about digital and you can move people as you would move with a mail piece.
You can scale a lot more. We want to do more of this. Doing things is the best way to learn. In 2012, the best way I learned is we had a last-minute push in Colorado. Dove in, took risk. It worked.