Listen to this article · 11 min listen

The 2026 midterms were a flood of digital political ads, and video was everywhere trying to shape conversations. The “Community Voices for Progress” campaign was one we watched closely because it was a non-partisan group just trying to get local infrastructure bonds passed. It’s a perfect example of how targeted video can actually move the needle on complicated local policy, which is a whole different beast than selling a candidate.

Key Takeaways

  • To get a 0.75% conversion rate on tough policy videos, your landing pages have to be perfectly synced with the ad and have an unmissable call to action.
  • We put 60% of the budget into micro-targeted YouTube ads using custom affinity audiences, and it paid off, our cost per view dropped by 40% compared to just blasting it out on general social feeds.
  • Constant A/B testing, especially pitting emotional story ads against our data-heavy ones, boosted CTRs by 15% for the winning creative.
  • Our retargeting was simple but effective: anyone who watched at least half of the first video got a follow-up ad. This simple sequence cut our cost per conversion by 25%.
  • You have to watch the comments. We monitored sentiment on YouTube and social media daily, which let us jump on negative chatter and tweak creative before a bad narrative could take hold.

Campaign Overview: Community Voices for Progress

The “Community Voices for Progress” campaign was a six-week sprint that started in August 2026, aimed at local bond referendums in Atlanta, Georgia. Charlotte, North Carolina. And Nashville, Tennessee. The whole point was to get voters on board with municipal bonds for transit, water systems, and parks. This campaign was purely about policy, not candidates, and was focused on getting people informed enough to vote yes on the ballot measures.

They had a $185,000 budget for digital video, which is a pretty serious number for a local advocacy group. Obviously, you can’t directly track a vote back to a specific ad, so we had to get smart with our KPIs. We focused on intermediate conversions like email sign-ups and completed surveys showing support, along with the usual suspects: cost per lead (CPL), click-through rate (CTR) to their info hub, total impressions, and a proxy ROAS we built from voter intent surveys.

Strategic Foundations: Targeting and Platform Selection

We had to reach two groups: undecideds and reliable local voters who just didn’t know the details on these bond measures. Going after broad demographics is a total waste of money for policy campaigns, so we went deep on behavioral and interest segments. Our main playgrounds were Google Ads for YouTube and display video, and LinkedIn Ads. We thought about Meta, but with their political ad policy shifts and the fact that deep policy content tends to die on Facebook and Instagram, we decided to put our money on YouTube and LinkedIn instead.

On the YouTube side, we got granular. We built custom affinity audiences from people who were already visiting local government sites, reading local news about city hall, or posting on community forums. Then we layered on channel targeting, hitting viewers on channels about urban planning, local news, and civic engagement. This meant we were only talking to people already tuned into these kinds of topics. LinkedIn was more about professional context, so we targeted by job titles like civil engineers, urban planners, educators, and local business owners, plus people in groups for community development.

The budget split was pretty clear: about 60% went straight to YouTube for pre-roll and in-stream ads. LinkedIn got 25% for sponsored videos in the feed. We kept the last 15% for hyperlocal programmatic buys on local news sites, running those through a demand-side platform (DSP) like The Trade Desk to make sure we were right next to relevant local content.

Creative Approach: Balancing Emotion and Information

We didn’t just run one video. We knew we needed a few different angles, so we developed three main creative series, each with its own style and message:

  1. “Local Impact Stories” (Emotional Appeal): These were 30-second spots with real residents from Atlanta’s Grant Park, Charlotte’s NoDa district, and Nashville’s 12 South neighborhood talking about how the bond projects would make their lives better. One video, for instance, had a parent in Atlanta talking about new pedestrian infrastructure making it safer to walk to school. It was all about tangible benefits.
  2. “Facts & Figures” (Data-Driven): We made 60-second videos with animated infographics and experts (local engineers, economists) to break down the money side of the bonds. They directly addressed fears about tax hikes, using hard data from sources like the Georgia Department of Transportation’s (GDOT) 2025 infrastructure report to show why doing nothing would cost more later.
  3. “Community Dialogue” (Testimonial-Based): These were longer, 90-second videos showing small, moderated groups of residents discussing the issues. They were intentionally unpolished to feel more authentic, showing different viewpoints arriving at a consensus about the need for the investment.

Every single video ended with a direct URL overlay and the CTA: “Visit CommunityVoices.org to learn more and pledge your support.” We A/B tested everything we could think of, constantly rotating creative, trying different lengths, changing up the first three seconds, and even tweaking the button text. That little test of “Learn More” vs. “See the Plan” actually made a difference.

Performance Metrics and Analysis

So, did it work? The numbers give us a pretty good picture. Here’s the top-level breakdown:

  • Total Impressions: 8.7 million across all platforms.
  • Overall CTR: 1.2% (to the information hub).
  • Total Conversions: 6,525 (email sign-ups or survey completions).
  • Overall Conversion Rate: 0.75% (conversions per impression).
  • Average Cost Per Lead (CPL): $28.35.
  • Estimated ROAS: Our post-campaign surveys found a 15% bump in informed support among our target demographics, so we felt confident calling the ROAS positive, even if it was indirect.

Platform-Specific Performance:

YouTube (60% of budget):

  • Impressions: 5.2 million
  • CTR: 1.5%
  • Conversions: 4,050
  • CPL: $27.40
  • Average View Rate: 45% for 30-second ads, 30% for 60-second ads.

LinkedIn (25% of budget):

  • Impressions: 2.2 million
  • CTR: 0.9%
  • Conversions: 1,600
  • CPL: $28.90
  • Average View Rate: 38% for 30-second ads.

Programmatic (15% of budget):

  • Impressions: 1.3 million
  • CTR: 1.0%
  • Conversions: 875
  • CPL: $31.70

Across every platform, the “Local Impact Stories” creative was the clear winner on click-through, averaging 1.8%. It just goes to show that people connect with personal stories, even for dry policy stuff. What’s interesting is that the “Facts & Figures” videos, while getting a lower CTR at 0.9%, had a much higher completion rate of 48% for the 60-second version once someone started watching. This tells you it was successfully deepening interest for people who were already leaning in, rather than just grabbing brand new attention.

Factor YouTube/Google Ads LinkedIn Ads
Budget Allocation 60% 25%
Targeting Strategy Custom affinity audiences, specific channels Job title, interest groups
Cost Per View (vs. broad social) 40% lower Not specified
Engagement for Policy High engagement Prioritized for complex discussions
Ad Formats Pre-roll, in-stream ads Sponsored content videos

What Worked and What Didn’t

What Worked:

  • Hyper-Targeting: Going deep with YouTube’s custom affinity and in-market segments was the right call. That kind of granular targeting kept our CPL low, even for this kind of complex messaging. We even saw a 2.1% CTR when targeting people who had literally just searched for “Atlanta public transit upgrades.”
  • Emotional Storytelling: Those “Local Impact Stories” worked. They put a human face on the bond measures, turning them from abstract policy into real community improvements. Policy campaigns almost always forget this and just throw dry data at people.
  • Retargeting: Our retargeting sequence was a huge win. We served follow-up ads (usually the “Facts & Figures” videos) to anyone who watched at least 50% of an initial video, and this simple two-step process cut our cost per conversion by 25% for that group.
  • Dedicated Landing Pages: This should be standard practice, but it’s worth mentioning. Every ad drove to a landing page on CommunityVoices.org that was a direct continuation of the video’s message. No hunting around for info. It killed our bounce rate.

What Didn’t Work as Expected:

  • Broad Programmatic Buys: Trying to go too broad with programmatic was a mistake. The placements on generic local sites had terrible CPLs and low viewability. It really drove home that hyperlocal context is everything. We pulled the plug on those wide buys after two weeks and moved the money.
  • Overly Technical Language: Our first batch of “Facts & Figures” videos was way too technical, full of jargon like “amortization schedules.” We had to recut them to be simpler. Turns out, even smart, engaged people don’t want to decipher engineering-speak.
  • Static Calls to Action: We started with a lazy, generic “Learn More” CTA. Our A/B tests quickly showed that specific buttons like “See the Full Plan” or “Calculate Your Impact” got a 10% better CTR. It’s a small change that had a real impact on performance.

Optimization Steps and Lessons Learned

We didn’t just set it and forget it. This was a six-week campaign, and we were in the data every single day making adjustments:

  1. Budget Reallocation: After two weeks, it was obvious the custom affinity audiences on YouTube were outperforming programmatic, so we moved 10% of the budget over to where it would work harder.
  2. Creative Refresh: About three weeks in, we saw CTRs start to dip on some of the “Local Impact Stories”, classic ad fatigue. We had two new versions ready to go with different residents, swapped them in, and the CTRs jumped right back up by 12%. Ad fatigue is definitely real.
  3. Landing Page Adjustments: We used heatmaps on the landing pages and saw people weren’t scrolling. So, we moved the key infographics and the sign-up form above the fold. Simple fix, and it boosted conversions by 5%.
  4. Sentiment Monitoring: We had someone watching comments on the YouTube ads and social mentions like a hawk. When we saw chatter about the “taxpayer burden,” we could immediately deploy a “Facts & Figures” video that tackled that specific point with data about long-term benefits. It was like playing defense in real-time and stopped the negative stuff from spreading.

What the “Community Voices for Progress” campaign really shows is that running policy debate video advertising that works comes down to three things: obsessive targeting, creative that isn’t one-note, and constant, daily optimization. You get results by coordinating the content, the audience, and the platform, not by just throwing a video online and hoping for the best.

So if you’re trying to use video to move the needle on public opinion, you have to do the homework. Know your audience inside and out, build messages that hit both the heart and the head, and watch the data so you can adapt on the fly. Policy is messy, and there’s no template for this kind of work.

What’s the right video length for a policy ad?

There’s no magic number. It all depends on the job the video needs to do. In this campaign, 30-second spots were great for grabbing attention with an emotional hook, but we needed 60-90 seconds to actually explain the details to people who were already interested. You need a mix of lengths to move someone from awareness to support.

How do you even measure ROI on these things?

Directly tying ad spend to a vote is nearly impossible since so many things influence an election. So, you have to measure what you can control. We track proxy metrics: email sign-ups, petition signatures, survey completions that show support, and how people behave on the website. These are your real indicators. Post-campaign polls can also show if you managed to shift opinion in your target groups.

Which platforms work best for policy video?

For the kind of targeting you need in policy, YouTube and LinkedIn are usually the best bets. YouTube is amazing for building custom audiences based on user behavior and getting broad reach, and LinkedIn is perfect for zeroing in on professional groups. Programmatic buys on local news sites can work for hyperlocal targeting, but you have to be really selective about the publishers or you’ll waste money.

Emotion or data: which is better for policy ads?

You need both. It’s a one-two punch. Emotion, like the personal stories we used, grabs their attention and makes the issue feel real. Then you hit them with the data to prove you’re credible, address their skeptical questions, and give them a reason to actually believe you. A good campaign uses creative that does both, sometimes in the same ad, but often in a sequence.

Is A/B testing really necessary for a policy campaign?

Absolutely. You can’t afford not to. We found that tiny changes to the first few seconds of a video, the button text, or the creative angle could completely change the performance. You have to test constantly to find out what’s actually working with your audience. It’s the only way to make sure your budget is being spent effectively instead of just being thrown into the wind.