Down-ballot races don't get won with broadcast TV. They get won with a small budget spent in the right places at the right time. Here's an anonymized look at what that actually looks like, the plan, the spend, the choices, and what the result depended on.
The Race
A municipal assembly seat in a mid-size city. Open seat, three candidates in the primary, top two advance. Our candidate was a first-time runner with strong local credibility but no district-wide name ID. Total digital budget: about $20K across roughly 10 weeks.
The Audience
We started from the voter file, not from demographics. Three universes:
- Base: high-propensity voters in the candidate's party, used for GOTV reinforcement
- Persuadable: mid-propensity, mixed-party history, the universe the race would actually be decided in
- Low-propensity supporters: voters who lean our direction but don't always show up, prioritized in the final two weeks
Each universe got matched to digital audiences across Meta and programmatic. Search ran on a separate logic, anyone in the district searching the candidate's name or the opponent's.
The Channel Mix
- Meta, 45%. Largest reach, best for issue messaging and bio content. Two creative concepts in rotation, refreshed every 2-3 weeks.
- Google, 25%. Brand defense (the candidate's name), opponent name searches, and YouTube pre-roll on the persuasion universe.
- Programmatic, 30%. Display, video, and CTV through a single DSP, hitting the persuadable universe with consistent messaging across formats.
The Phasing
- Foundation (weeks 1-2): 18% of budget. Bio, intro video, "who is this candidate" content. Heavy on Meta, supporting on YouTube.
- Persuasion (weeks 3-7): 37% of budget. Issues, contrast, endorsements. All three platforms running in parallel.
- GOTV (weeks 8-10): 45% of budget. Vote-by-mail reminders, polling location info, urgency creative. Heavier weight on programmatic CTV and SMS-paired Meta.
The phasing was the most important strategic decision. Front-loading the budget would have meant running out of money in the final stretch when persuadable voters actually pay attention.
The Creative
Three concepts ran across the cycle:
- "Why I'm Running": vertical video, candidate to camera, 30 seconds. Foundation and persuasion phases.
- Contrast on a single issue: the issue voters in the district consistently named in polling. Static carousels and short video.
- "Vote Tuesday" GOTV: date, polling location, candidate name. Repeated across formats in the final 10 days.
Every cut had captions burned in, a clean disclaimer, and was sized to its placement. No auto-cropped reuse.
What Actually Moved the Result
- Brand defense on Google was the highest-ROI line item in the entire plan. About $300 in spend captured every voter searching the candidate's name through Election Day, at a fraction of the cost of any other channel.
- CTV on the persuadable universe hit completion rates above 92%. Voters who saw the CTV spot also converted on Meta retargeting at meaningfully higher rates.
- The GOTV pivot: moving budget aggressively into the final two weeks based on early-vote data was probably worth several hundred net votes in a race that was decided by under 1,000.
- Cutting losing creative early. One concept underperformed by week 3 and we killed it. The temptation to "let it run" because the candidate liked it would have wasted real money.
What We'd Do Differently
- Start search earlier. Brand defense should be live the day the campaign launches, not week 3.
- More creative iteration. Three concepts was enough, but four to five would have given the persuasion phase more lift.
- Better landing page. The campaign site was fine, not great. A purpose-built donate-and-volunteer page would have improved Meta CPL by 20%+.
Bottom Line
Down-ballot wins are made of unglamorous decisions: where to put the next dollar, when to cut a creative, when to shift the phasing. There's no single tactic that wins. There's a mix that fits the race and a willingness to change it as the data comes in.