You've shot a 60-second CPO spot. It's polished, on-brand, and your agency loves it. But you're about to spend six figures on digital and broadcast placement. If the emotional arc is wrong, you'll know in week three when conversions flatline.
Second-by-second facial coding gives you a different option. Show the cut to opt-in panelists who match your buyer profile. Track their facial expressions frame by frame. Find the exact moments where engagement drops or confusion spikes. Then re-edit before you buy media.
Why CPO creative needs emotion testing
Certified pre-owned sits in a weird middle ground. It's not new-car aspiration. It's not used-car bargain hunting. Your messaging has to thread trust, value, and a little bit of pride all at once.
If your opening 8 seconds lean too hard on "like new," budget shoppers tune out. If you emphasize savings first, aspirational buyers feel cheapened. And if your warranty explanation runs long, everyone's face goes neutral and you've lost them.
Traditional focus groups catch big problems, but they don't show you second 23 versus second 34. Facial coding does. You see the micro-expressions that predict actual viewing behavior: when eyes wander, when brows furrow, when a smile breaks through.
Set up your pre-test panel
Start with 40 to 80 panelists who mirror your target demos. If you're selling CPO luxury crossovers to dual-income households aged 35-50, recruit that profile. EmotionTrac's automotive panels include conquest shoppers, brand loyalists, and cross-shoppers who are 90 days out from a purchase decision.
Panelists watch your video on their own devices. Webcams capture facial expressions. The system codes 7 core emotions in real time: joy, surprise, sadness, anger, fear, disgust, contempt. You get a timeline that maps every frame of your ad to aggregated emotional response.
Run the test before your edit is locked. Ideally, you have two or three cut variations. Test them all. The data will tell you which opening works, which benefit sequence holds attention, and which closing call-to-action actually lands.
Read the emotion timeline
Your first look should be the joy and engagement curves. In a strong CPO ad, joy spikes when you show the vehicle in a desirable context. It dips slightly during warranty talk (expected), then recovers when you restate value or show a happy owner.
If joy never recovers after the warranty section, your creative is back-loaded wrong. You've buried the emotional payoff. Move it up or cut the explanation shorter.
Watch for surprise peaks. A small surprise spike when you reveal price or financing terms is good; it signals "better than expected." A surprise spike during your brand promise means confusion. Panelists didn't see that claim coming, and not in a good way.
Contempt and disgust are your kill signals. If either emotion spikes above baseline, you've triggered skepticism or distaste. Common causes in CPO ads: over-promising ("like driving a brand-new car"), clumsy disclaimers, or visuals that undersell the product (a CPO luxury sedan shot in a bland parking lot).
Flag the drop-off points
Facial coding shows you where panelists disengage. Engagement is a composite score: are they looking at the screen, are their expressions active, are they leaning in or pulling back?
In CPO video, you'll often see a drop around second 35 to 45. That's when many ads transition from emotional appeal to feature list. If engagement falls and never recovers, your feature section is too long or too boring. Cut it or weave it into the story earlier.
Another common drop: right after the call to action. If your CTA is "Visit our website," and engagement craters immediately, panelists are done. That's fine if your ad is 30 seconds. It's a problem if you have 20 more seconds of brand montage. Either move the CTA to the end or make the closing sequence compelling enough to hold attention.
Compare cuts side by side
If you're testing two edits (say, one that opens with the vehicle reveal and one that opens with a customer testimonial), the emotion timelines will look different. The vehicle-first cut might spike joy earlier but lose steam in the middle. The testimonial-first cut might build slower but hold engagement longer.
Look at the area under the joy curve. Whichever cut has more cumulative joy will probably perform better in-market, assuming the drop-off points aren't catastrophic.
Also check the negative emotion floors. If Cut A triggers a contempt spike and Cut B doesn't, that's your answer even if Cut A has a higher joy peak. You can't afford skepticism in a CPO ad. Trust is the whole game.
Test your disclaimer and fine print
Legal disclaimers kill engagement. You know this. But you're required to include them. Facial coding helps you find the least-bad placement.
Test three positions: disclaimer at the start (get it over with), disclaimer mid-roll (split the pain), disclaimer at the end (standard). The data will show you which approach preserves the most joy and engagement in the surrounding frames.
In most cases, end placement wins. Panelists tolerate fine print after they've already felt good about the offer. But if your offer is complex (multiple trim levels, varying warranty terms), mid-roll might work better because it contextualizes the details when they're still paying attention.
Validate your call to action
Your CTA might be "Shop CPO inventory online," "Visit your local dealer," or "Get pre-qualified in 60 seconds." Each one sets a different expectation and friction level.
Facial coding won't tell you which CTA converts best (you need actual campaign data for that), but it will tell you which CTA feels most congruent with the emotional journey you just created.
If your ad built trust and comfort, and then you ask people to "Act now, limited inventory," you'll see a surprise or confusion spike. The urgency doesn't match the tone. If your ad was energetic and aspirational, and you end with "Speak to a certified specialist," engagement drops because that feels like friction.
Run a quick follow-up test with CTA variations if the first pass shows a problem. Swap the last 5 seconds and re-test. It's cheaper than guessing in-market.
Try it: Schedule an EmotionTrac demo and see second-by-second emotion tracking in action. Or visit Automotive for more information.
When to re-edit versus accept the data
If your test shows one or two fixable problems (a slow middle, a confusing transition), re-edit and re-test. You'll spend another week and a small panel fee, but you'll save the media budget.
If your test shows systemic issues (low joy throughout, multiple contempt spikes, engagement below 50% by the halfway mark), the problem isn't the edit. It's the concept or the script. Go back to creative development. Don't try to salvage a fundamentally weak idea with tighter cuts.
And if your test shows strong performance (high joy, sustained engagement, clean CTA response), lock the edit and move to media planning. You've de-risked the spend.
Build a CPO emotion benchmark
After you've tested 3 or 4 campaigns, you'll have a benchmark library. You'll know what "good" looks like for your brand's CPO messaging. You'll know the joy floor you need to hit in the first 10 seconds. You'll know how much engagement drop you can tolerate during warranty explanation.
Use that benchmark to evaluate new creative faster. If a new cut underperforms your historical average, you know it needs work before it sees daylight. If it outperforms, you know you've got a winner and you can weight your media buy accordingly.
Sources
Höfling, T. T. A., & Alpers, G. W. (2023). Facial expressions in automated emotion recognition: A review of research. Frontiers in Neuroscience, 17. DOI: 10.3389/fnins.2023.1125983
EmotionTrac Automotive: https://auto.emotiontrac.com/