How does camera-based audience measurement work for out-of-home?
Sneh Dhruv · Updated
Camera-based audience measurement works by processing live video feeds at the screen to detect faces and calculate exact dwell time and attention. You can read measured attention, dwell and exposure directly from Trillboards, pulling data from 5,456 screens whose camera measured the audience to verify your exact physical reach.
The numbers
| Measure | Figure | Sample | Data window |
|---|---|---|---|
| Per-face audience records in the latest 28 settled days | 5,712,174 | 5,712,174 records | Sep 13 – Oct 10, 2026 |
| Billboard passes measured from moving vehicles | 23,518,207 | 23,518,207 passes | Jul 13 – Oct 10, 2026 |
| Median legible seconds per legible billboard pass | 6.4 | 14,904,064 legible passes | Jul 13 – Oct 10, 2026 |
| Seconds a passenger faced the board, per camera-judged legible pass | 1.40 | 846,637 judged passes | Jun 28 – Sep 25, 2026 |
| Share of camera-judged legible passes with at least one second facing the board | 28.1% | 846,637 judged passes | Jun 28 – Sep 25, 2026 |
Moving from modeled traffic to physical measurement
Out-of-home advertising historically relies on modeled traffic counts and mobile-panel estimates to determine how many people pass a location. Organizations like Geopath combine cellular mobility data with vehicular traffic models to establish a baseline currency for the market. These models give you an Opportunity to See metric for your placements. Camera-based audience measurement builds on this foundation by processing live video feeds directly at the screen to detect physical faces. Edge-analytics vendors like Quividi and AdMobilize deploy computer vision to count the exact number of people present during an ad play. You capture hard evidence of your audience, reviewing 5,712,174 per-face records in 28 days to confirm who stood in front of the display.
A buyer for a quick-service restaurant can verify exact footfall at a transit shelter. You evaluate campaigns using physical gaze data. Modeled traffic counts and mobile-panel estimates provide a baseline for planning. Camera systems process the visual field locally to extract demographic estimates and count real impressions. You see the exact composition of the crowd standing near your placements. The industry relies on these edge-AI cameras to bridge the gap between ad playout and actual physical viewers.
Evaluating passing vehicles and pedestrians
Cameras mounted on digital screens evaluate the physical environment frame by frame. The system detects faces and measures the geometry of the space to determine how long a person remains in the viewing area. You can apply this technology to roadside inventory to measure vehicular traffic. The camera logs 23,518,207 billboard passes measured from vehicles to quantify the exact volume of cars moving past the display. The system calculates the exact duration the screen remains visible to the driver and passengers. You see 6.4 seconds legible, the median per legible billboard pass. This physical measurement gives you a precise understanding of the exposure window available for your creative. You know exactly how long a commuter has to read your message.
The camera tracks the trajectory of people and vehicles through the field of view. You measure the exact dwell time for every person who walks past an urban kiosk. The system records the moment a person enters the frame and the moment they leave. You use this data to determine the optimal length for your video assets. A buyer for a consumer banking brand can adjust their creative rotation based on the exact amount of time pedestrians spend waiting near a digital display.
Quantifying attention and likelihood to see
The Media Rating Council divides out-of-home exposure into Opportunity to See and Likelihood to See. Camera systems measure Likelihood to See by tracking head yaw and eye gaze to determine if a person actually looked at the screen. You move beyond basic presence to measure active attention. The system records 1.40 seconds facing the board per legible pass to show you the exact duration of engagement. You can evaluate the efficiency of different creative formats by looking at the percentage of people who turn their heads. The data shows 28.1% of camera-judged legible passes had a passenger facing the board for at least one second.
A media buyer for a regional hospital system can use these metrics to confirm their awareness campaign captured active attention from drivers. The camera evaluates the angle of the face relative to the display. You count an attention impression when a person looks directly at the screen for a full second. You compare the attention rates across different dayparts and locations to identify the most engaging environments. You allocate your budget to the screens that consistently deliver high gaze times.
Transacting on verified camera metrics
The digital out-of-home market operates on programmatic Cost-Per-Thousand models. Standard screens average five to fifteen dollars per thousand impressions. Premium locations command twenty to fifty dollars. You can buy this inventory through demand-side platforms and content management systems like Broadsign, Vistar Media and StackAdapt. The camera metrics feed directly into these platforms to verify your impressions. Screen operators pay flat annual software subscriptions or monthly per-screen licenses to run the edge hardware.
You use the resulting data to optimize your bidding strategies and select the highest-performing locations. You tie your media spend directly to verified physical viewers. The Interactive Advertising Bureau and the Digital Place-Based Advertising Association establish strict auditing guidelines for these systems. You rely on these standards to ensure the accuracy of the camera data. Programmatic out-of-home represents a massive portion of the market as buyers demand these verified metrics. You integrate the camera data into your attribution models to measure the true impact of your campaigns. You build a complete picture of your audience by combining physical camera signals with your existing measurement tools.
Sources
- arsa.technology
- otegomedia.com
- digitalsignagetoday.com
- b2bleadnavigator.com
- indooradvertising.org
- ariadne.inc
- moviamedia.com
- assets.contentstack.io
Related on Trillboards
- Who are the DOOH audience measurement providers?
- Where can you get dwell time data for retail screens?
- Where can you get eye tracking data for outdoor advertising?
- Is there an API for out-of-home attention data?
- How does foot traffic attribution work for billboards?
- Every measurement and data answer
- Waiting Audiences: price and index
- Published CPM floors for every audience
- Every audience we sell
- Screens and audiences by city
Plan this on Trillboards
Attention, dwell and exposure are measured and published with their sample sizes and data windows. Plan it on live inventory to see the screens, weekly views and effective CPM for the audience you pick.
Frequently asked questions
- Do digital billboards record video of the people walking by?
- Digital billboards process live feeds to detect faces and calculate dwell time. Some networks send selected camera frames to a cloud model to estimate demographics and keep those frames for hundreds of days. The systems do not store or transmit continuous video or audio recordings of the public.
- How do publishers price edge analytics hardware and software?
- Screen operators typically pay for the technology through flat annual software subscriptions or monthly per-screen licenses. These costs run alongside the initial purchase of the edge hardware. You buy the resulting ad inventory through programmatic channels using standard cost per thousand impression models based on the measured audience.
- Can advertisers link camera measurements to mobile devices?
- Advanced measurement systems link the faces detected by the camera to the device keys of the phones seen with them. The network connects these physical sightings to mobile advertising identifiers. You use these connections to understand the digital profiles of the people standing near your out-of-home placements.
- What demographic data do these vision systems estimate?
- The computer vision models evaluate the captured frames to estimate the age band and gender of the people present. The cloud models also estimate ethnicity, social composition, attire, carried items and mood. You review these demographic breakdowns to ensure your creative reaches the correct target audience.
- How do demand-side platforms receive the audience data?
- The screens send bid requests that carry the live audience measurements along with the IP address and location of the display. You receive these signals directly in your buying platform. You evaluate the live conditions to decide if you want to purchase the available ad play.
Plan it
Open the planner in the planner to see the screens, the weekly views and the price before you book, or browse the audience catalogue.