Key Takeaways
- Precise CTV targeting slashed our wasted ad spend. By layering audience segments, we cut impressions served to the wrong people by 75%.
- Programmatic audio was a workhorse, delivering a 2.5x higher engagement rate for our mid-funnel content compared to display, which helped lower our cost per conversion by 15%.
- Our integrated attribution model revealed audio’s real influence: 30% of conversions came from users who had heard an audio ad early in their journey, even without a click.
- Agencies must test creative. Set aside 30-40% of an initial budget to find the winning CTV and audio assets before you even think about scaling up.
- A/B testing frequency caps on CTV is a simple way to get a 10% lift in completion rates and dial back ad fatigue.
Getting attention with Connected TV (CTV) and audio ads is tough when everyone’s media consumption is fragmented. So, how are some agencies actually delivering a solid return on ad spend instead of just adding to the noise? Let’s tear down a real campaign to see what works.
Campaign Teardown: Driving Subscriptions for a Niche Streaming Service
We just wrapped a multi-channel campaign for “CineCult,” a new streaming service for independent and arthouse film buffs. The main goal: drive new subscriptions in a hyper-competitive market. The campaign ran for 12 weeks from January to March 2026 on a $350,000 budget, where our entire focus was on efficient CTV buying and smart audio placements to reach audiences the big streaming platforms often ignore.
Strategy and Objectives
Our strategy targeted actual film enthusiasts and early adopters who prefer unique content over the latest blockbuster. We set our goals at a cost per lead (CPL) under $15 and a return on ad spend (ROAS) of at least 1.5x. We knew this audience was watching ad-supported CTV and listening to podcasts, so our job was to find those niche segments and hit them with creative that felt right. For our purposes, a “lead” was anyone who completed a free trial signup, and a “conversion” was when they put a credit card down after the trial.
Targeting Precision: The Core of Our Approach
For CTV, we started with CineCult’s first-party data (think email lists from past film events and survey-takers) and layered on third-party data segments to find households with high engagement in arts and culture content or existing premium VOD subscriptions. To do this, we used data from providers like Nielsen and Acxiom right inside our demand-side platform (DSP), The Trade Desk. We then geo-targeted arts-focused communities like Brooklyn in NYC, Silver Lake in Los Angeles, and the Mission District in San Francisco. Our audio targeting was just as granular. We went after listeners of podcasts in specific genres like film criticism and independent culture, plus streaming stations known for alternative music. Then we layered on demographics: adults 25-54 with a household income over $75,000. Programmatic audio platforms let us target users based on their listening habits and device, which meant our ads played during relevant sessions, not just randomly.
Creative Development: Tailoring the Message
We developed two creative concepts for CTV and three for audio.
CTV Creative:
- “The Unseen Story” (30-second spot): This one was moody and cinematic, showing snippets from CineCult’s exclusive films to highlight their unique stories. The call to action (CTA) was a straightforward “Start your free trial today at CineCult.com.”
- “Beyond the Blockbuster” (15-second spot): A faster, punchier ad with quick cuts and fake testimonials from film critics. This was designed to grab attention from users who might be flipping through channels.
Audio Creative:
- “Deep Dive” (60-second podcast sponsorship): A long-form ad read by the podcast hosts themselves, making the CineCult plug feel natural to the show’s content by focusing on its curated library.
- “Sound of Cinema” (30-second programmatic audio): This was a produced audio spot using evocative sound design to sell the emotional punch of independent film.
- “Quick Cut” (15-second programmatic audio): A short, direct ad that just hammered the free trial offer.
The branding was consistent across the board, and every ad had a clear CTA.
Initial Campaign Performance & Metrics
We launched with an even 50/50 budget split between CTV and audio for the first three weeks to get baseline data.
Initial 3-Week Performance (January 2026):
- Total Impressions: 15,200,000
- CTV Impressions: 9,800,000
- Audio Impressions: 5,400,000
- Total Clicks/Engagements (across all channels): 125,000
- CTR (Overall): 0.82%
- Total Leads (Free Trial Sign-ups): 6,250
- CPL: $16.00
- Total Conversions (Paid Subscriptions): 800
- Cost per Conversion: $125.00
- ROAS: 0.9x (subscription value averaged $12.50/month for 12 months)
The numbers weren’t great. The CPL was a bit over our target, and at a 0.9x ROAS, we were underwater. Digging in, we saw “The Unseen Story” had a strong 92% completion rate on CTV, but the “Beyond the Blockbuster” spot was lagging at 78%. For audio, the podcast sponsorships were the clear winner, with a 3.5% engagement rate (tracked via unique URLs and promo codes) that blew away the 1.8% from our programmatic audio.
Optimization Steps and Adjustments
That initial data told us exactly where to make changes.
CTV Optimizations:
- Creative Prioritization: We paused the underperforming “Beyond the Blockbuster” spot and shifted 80% of the CTV budget to “The Unseen Story.” This improved overall CTV completion rates to 94% almost immediately.
- Frequency Capping: We were seeing some ad fatigue, so we adjusted the frequency cap down from 5 to 3 impressions per user per week. This cut our impression waste and improved engagement, giving us a 5% increase in post-completion website visits.
- Audience Refinement: We uploaded the email list of everyone who signed up for a trial and used it to build lookalike audiences in our DSPs to find more people just like them across the CTV ecosystem.
Audio Optimizations:
- Budget Reallocation: We moved 60% of the audio budget into the high-performing podcast sponsorships. For the programmatic audio that remained, we got laser-focused on placing ads only in podcast episodes about film festivals or indie movie news.
- A/B Testing CTAs: We ran A/B tests on our programmatic audio CTAs and found that a direct “Visit CineCult.com for a free trial” drove 10% more traffic than “Search ‘CineCult’ in your app store.”
- Time-of-Day Bidding: Our analytics showed peak listening during evenings and weekend mornings, so we increased our bids during those windows (6 PM-10 PM and 9 AM-12 PM) to make sure we were present.
Results After Optimization
These optimizations completely turned the campaign around over the next nine weeks (February-March 2026).
Post-Optimization 9-Week Performance (February to March 2026):
- Total Impressions: 32,500,000
- CTV Impressions: 22,000,000
- Audio Impressions: 10,500,000
- Total Clicks/Engagements: 310,000
- CTR (Overall): 0.95%
- Total Leads (Free Trial Sign-ups): 19,375
- CPL: $11.35 (a 29% reduction from initial)
- Total Conversions (Paid Subscriptions): 3,250
- Cost per Conversion: $70.77 (a 43% reduction from initial)
- ROAS: 2.1x (exceeding our 1.5x target)
Our refined CTV targeting and creative focus increased the video completion rate for “The Unseen Story” another 10%, getting it to 96%. Podcast sponsorship engagement rates climbed to 4.2%, and even though programmatic audio had lower engagement per ad, its lower CPM delivered massive reach that was critical for filling the top of the funnel.
Attribution and Cross-Channel Insights
We used a data-driven attribution model that looked at both view-through and click-through conversions across CTV and audio. The real shock was how big a role audio played at the start of the customer journey. Around 30% of users who ended up converting had an audio ad as one of their first three touchpoints, even if the final click came from a CTV ad or a direct site visit. Audio builds brand awareness and intent, a fact that a recent IAB report confirms is growing. We also saw that users exposed to both CTV and audio ads had a 1.8x higher conversion rate than those who only saw one. It’s a perfect argument for an integrated media buying strategy.
What Worked Well
Our granular targeting, especially using first-party data to build lookalikes, was the single most valuable part of this campaign. For creative, the high-quality “The Unseen Story” spot for CTV just worked. The authentic, host-read podcast ads proved that for audio, context and trust are everything. All the continuous A/B testing on CTAs and frequency caps was what let us make agile changes to prevent ad fatigue and squeeze every drop of value from the budget.
Challenges and Lessons Learned
The biggest challenge was the high initial CPL, which forced us to analyze the data and optimize much faster than we’d planned. We also confirmed that attributing the true impact of programmatic audio is hard if you’re only looking at last-click. It’s an awareness driver, and our model showed it was clearly influencing conversions down the line. Another lesson was in managing inventory quality. Some CTV apps and publishers are just bad, offering no transparency and low completion rates. This requires vigilant monitoring and platform-specific expertise to blacklist the junk.
Future Recommendations
For the next campaign, we’d recommend testing interactive CTV formats, maybe with QR codes viewers can scan to go straight to a trial page. We’re also looking at dynamic creative for audio, which would let us change ad copy in real time based on a listener’s location. Investing in better advanced attribution tools will be critical to proving the value of every impression. The CineCult campaign proves that for a niche brand, a combination of precise CTV buying, smart audio ads, and relentless optimization can get real results. Being able to adapt quickly to performance data is essential for any effective media buying in 2026.
What’s a good CTR for CTV campaigns?
A good click-through rate (CTR) for CTV advertising varies by industry, ad format, and how well you’ve targeted. For campaigns trying to get a direct response, a CTR between 0.7% and 1.2% is solid, especially if you also have high video completion rates. For brand awareness campaigns, you should care more about completion rate and brand lift studies.
How does frequency capping affect CTV performance?
Frequency capping prevents ad fatigue and helps with ad recall. If you set a smart cap, like 3 to 5 impressions per user per week, you can improve completion rates and engagement. Too much frequency annoys users and wastes your spend, while too little means your message won’t have a chance to stick.
What’s better about programmatic audio vs. traditional radio?
Programmatic audio gives you amazing targeting, letting you reach specific people based on their listening habits, device, or location. It also gives you real-time bidding, detailed metrics, and the ability to adjust campaigns on the fly, none of which is really possible with traditional radio buys. That precision means less wasted money and better results.
How can we get better attribution for CTV and audio?
To improve attribution, you have to get away from last-click models. Agencies must implement multi-touch attribution that looks at all touchpoints in the journey, including view-through impressions from CTV and audio. This means integrating data from your DSPs, analytics platforms, and CRM. Incrementality testing and lift studies also help quantify the real impact of each channel.
What’s the role of first-party data in CTV buying?
First-party data is invaluable for advanced CTV buying. It lets agencies create precise custom and lookalike audiences, which improves targeting accuracy. By uploading customer email lists or website visitor data into DSPs, you can reach your most valuable segments on their TVs, which increases conversion rates and makes your ad budget more efficient. This data also helps personalize the ad creative.
