You're building two video campaigns. One emphasizes craftsmanship, exclusivity, and aspiration. The other focuses on smart value, practicality, and savings. You suspect different income groups will respond differently, but you don't know where each message lands or fails.
Testing these frames with opt-in panelists and facial coding gives you second-by-second emotion data before you buy media. You'll see exactly when high-income viewers disengage from value messaging, or when budget-conscious audiences reject luxury cues. This guide shows you how to run that test.
Why income segments respond differently to message frames
High-income viewers often respond positively to scarcity, heritage, and aspiration cues. They're less motivated by price anchors or discounts, which can trigger skepticism or disinterest. Middle-income audiences balance both: they want quality signals but also need reassurance they're making a smart financial choice. Budget-conscious viewers prioritize transparency, savings proof, and functional benefits over abstract lifestyle imagery.
Facial coding captures these differences in real time. When a luxury message shows a hand-stitched leather detail, you'll see sustained positive affect in high-income panelists and earlier disengagement in budget-focused viewers. When a value message emphasizes "30% off," the pattern reverses. These reactions happen within 2-3 seconds and often contradict what people say in post-viewing surveys.
Set up your test with income-segmented panels
Start by recruiting opt-in panelists across three income bands: high-income (top 20% household income for your market), middle-income (40th-70th percentile), and budget-conscious (below 40th percentile). You need 25-40 panelists per segment for stable emotion timelines. EmotionTrac panels are pre-qualified and consent to facial recording before viewing.
Prepare two video variants: one with luxury framing (heritage, exclusivity, aspiration), one with value framing (savings, practicality, smart choice). Keep runtime identical (30-60 seconds for ads, 90-180 seconds for brand films). Panelists watch both videos in randomized order to control for fatigue effects.
Each panelist's webcam records their face while they watch. Facial coding algorithms track micro-expressions frame-by-frame, measuring valence (positive/negative affect), engagement, and specific emotions like surprise or confusion. You get three emotion timelines: one per income segment.
Read the emotion timelines for divergence points
After data collection, overlay the three timelines for each video. Look for moments where one segment's emotion curve diverges sharply from the others. These are your friction points.
In a luxury message test, you might see high-income viewers sustain positive affect through a slow pan of a workshop, while budget-conscious viewers drop into neutral or negative valence at the 8-second mark. That's the moment the message stops working for them. In a value message test, budget-conscious viewers might spike positive when "save $200" appears on screen, while high-income viewers show flat or negative affect at the same timestamp.
Engagement drops are equally important. If middle-income viewers disengage 15 seconds into a 30-second luxury spot, you're losing your swing audience before the call-to-action. If high-income viewers tune out during a price comparison montage, you've confirmed that frame doesn't work for them.
Identify which cues drive the divergence
Once you've flagged divergence timestamps, review the video content at those exact moments. What's on screen? What's being said? Common luxury cues that alienate budget-conscious viewers include slow-motion lifestyle footage without product context, abstract language like "crafted for those who understand," and price reveals that feel inaccessible. Common value cues that lose high-income viewers include aggressive discount graphics, comparison charts, and language emphasizing "affordable" or "budget-friendly."
Middle-income viewers often respond well to hybrid cues: quality signals paired with transparent pricing, or aspiration paired with practical benefits. If your test shows middle-income engagement staying high through both message types, you've found a flexible audience. If they disengage from both, you need a third frame tailored specifically to them.
Surprise spikes can reveal unintended reactions. A luxury message might trigger surprise (often negative) when the price appears if it's higher than expected. A value message might trigger surprise (positive) if the savings proof feels unexpectedly generous. These micro-expressions help you calibrate expectations before launch.
Rebuild messaging based on segment-specific emotion data
Use the divergence points to edit or rebuild each video. For high-income audiences, remove or shorten value-focused segments that caused disengagement. Extend scenes that sustained positive affect, like product craftsmanship or brand heritage. For budget-conscious audiences, cut abstract lifestyle footage and front-load functional benefits and savings proof.
If you're running a single video across all segments, prioritize the middle-income timeline. Find the overlap where all three segments showed positive or neutral affect, and build your edit around those moments. Cut everything that caused sharp negative divergence in any segment. You'll sacrifice some peak resonance with niche audiences, but you'll avoid alienating large groups.
Test your revised edits with fresh panelists from the same income segments. Compare the new timelines to the original test. You should see reduced divergence and higher sustained engagement across all three groups. If one segment still disengages early, consider separate creative tracks for that audience.
Try it: Schedule an EmotionTrac demo and see second-by-second emotion tracking in action. Or visit Audience for more information.
Run follow-up tests to validate media allocation
Once you've optimized each message, test media allocation scenarios. Show high-income panelists the luxury edit and budget-conscious panelists the value edit. Measure whether sustained positive affect increases compared to the original mixed-message test. If it does, you've validated separate creative tracks.
If you're working with a single video, test different media placements. Show the hybrid edit to middle-income panelists in different contexts (social feed, pre-roll, connected TV) and measure whether engagement holds across formats. Context can shift how income segments interpret the same message.
Track emotion data over multiple exposures if you're planning frequency-heavy campaigns. Some luxury cues that work on first viewing can feel repetitive or irritating by the third exposure. Value messages that feel urgent initially can lose credibility with repeated viewing. Facial coding across 2-3 exposures reveals wear-out patterns before you overspend on frequency.
Build a reusable emotion library for future campaigns
Save your segmented emotion timelines as benchmarks. When you test new creative, compare the new timelines to your library. If a new luxury message underperforms your best historical timeline with high-income viewers, you know it needs revision before launch. If a new value message matches or exceeds your budget-conscious benchmark, you can move forward with confidence.
Tag specific creative elements (voiceover tone, music style, visual pacing, price presentation) that consistently drove positive or negative reactions in each income segment. Over time, you'll build a pattern library that speeds up creative development. You'll know which cues to avoid and which to emphasize before you shoot.
Share these benchmarks across your team. Media buyers can use emotion data to justify budget allocation across segments. Creative teams can reference proven reaction patterns when briefing new work. Product teams can see which benefit claims actually land with target income groups versus which ones get ignored.
Sources
Höfling, T. T. A., & Alpers, G. W. (2023). Automated facial action coding system validation using transdermal optical imaging. Frontiers in Neuroscience, 17. DOI: 10.3389/fnins.2023.1125983
EmotionTrac Audience: https://audience.emotiontrac.com/