How to Estimate Billboard Impressions From One Week of Footage

Once you have footage, the question becomes: how do you turn a handful of sampled hours into a weekly impressions number you can put on a rate card? This guide explains how to estimate billboard impressions honestly — from measured counts, to the multipliers that make them impressions, to the confidence range that has to travel with every projection.

It is the math companion to capturing footage with a phone and sits under the OOH audience measurement guide. The goal is a number that is generous enough to sell and honest enough to survive an agency's questions.

Key takeaways

• Impressions come in two steps: measure the counts in the visibility zone, then convert them with standard occupancy and visibility multipliers.
• Sampling works because a face's traffic follows a repeatable daypart-and-weekday pattern — you capture the shape, not every second.
• Every projected number ships with a confidence interval; the interval reflects count noise, not whether your sample was representative.
• AdWitness labels estimates low, medium, or high confidence and refuses to project a month from seconds of footage.

Step 1: From footage to counts

The engine detects and tracks each pedestrian and vehicle in your footage and counts the ones that cross your counting line inside the visibility zone. That gives you a measured count for each sampled window — say, 320 pedestrians and 210 vehicles between 5pm and 6pm on a Thursday. This part is observation, not modelling: it is what actually passed.

Step 2: From counts to impressions

A count is not yet an impression. An impression is an opportunity to see — and a car is not one person. To convert, the industry applies two standard multipliers:

• Occupancy (load) factor — the average number of people per vehicle, so 210 cars becomes roughly 210 × the load factor in people.
• Visibility adjustment — a factor for how likely a passer-by in the zone is to actually notice the face, based on things like the ad's size, angle, and illumination.

Impressions = counts × occupancy factor × visibility adjustment.

Be clear about what is measured and what is borrowed. AdWitness measures the counts from your footage. The occupancy and visibility multipliers are industry-standard averages, not something observed in your specific clip. A defensible certificate prints both the measured counts and the multipliers it applied, so a buyer can see exactly where the number came from — the honesty rule behind never selling on an unverified number.

Step 3: From a sample to a full week

Sampling works because a face's audience is not random — it follows a stable rhythm. Mornings look like mornings, Saturdays look like Saturdays. So the projection:

1. Averages the counts within each daypart (morning, midday, evening, late) and by weekday versus weekend.
2. Scales those averaged rates up to fill the hours you didn't film.
3. Sums the filled week into a weekly total, then a daily average.

The wider and more representative your samples, the truer the shape — which is exactly why good capture spreads across dayparts rather than sitting in one long block.

Step 4: Attach a confidence interval

No projection is a single exact number, and pretending otherwise is how OOH numbers lose credibility. Each estimate carries a confidence interval — a range within which the true figure most likely sits. You will see it as something like:

Weekly impressions: ~42,000 (most likely 34,000–50,000).

That band widens when counts are sparse and narrows when you have more data. Crucially, the interval reflects count noise only — the natural variation in how many people pass. It does not, by itself, tell you whether the hours you sampled were representative of a normal week. That second kind of uncertainty is handled separately, in the next step. The statistics behind the band are unpacked in what "accurate" really means for an impressions number.

Step 5: Label the confidence honestly

Because a tight interval on a tiny sample can look more certain than it is, AdWitness adds a plain-language confidence tier to every estimate:

• Low / indicative — a short spot reading. Useful as a first look, but too little footage to project a full day. The report says so, and it will not publish a monthly figure from it.
• Medium — enough footage to give a reasonable daily projection with a visible margin.
• High — a couple of hours or two or more separate observed days, enough to project a week with confidence.

This is the difference between a number that impresses a buyer and a number that embarrasses you when they check it. A "busiest hour" is only shown when there is enough footage to mean it; a monthly projection only appears when the sample can actually support one.

A worked example: how to estimate billboard impressions

Suppose you sample 6 spread-out hours across a week for a roadside face and measure an average of 480 people-equivalents per hour passing through the visibility zone after applying the occupancy factor. Scaled across the face's facing hours and adjusted for visibility, that projects to roughly 42,000 weekly impressions, with a most-likely range of about 34,000–50,000 and a high confidence tier because the sample spanned several days and dayparts.

That is a number you can put in a media kit: a headline figure, a range, a stated method, and a confidence label — everything an agency needs to accept it.

Turn your footage into a certified estimate

The Audience Certificate runs all five steps for you and prints the workings: measured counts, the multipliers, the weekly and monthly projections, and the confidence band on each. Certify one face for $99 to see your own estimate, or read how to read the certificate next. For the full landscape, return to the OOH audience measurement guide.