Footfall Counter Software Visitor Counting, Occupancy Counting Software, and Customer Counting on the Cameras You Already Own
Footfall counter software counts the people crossing a doorway from a video stream instead of from a dedicated sensor mounted above it. It reads the RTSP or ONVIF feed your existing CCTV cameras already produce, decides that a given pattern of movement across a line is one person entering or leaving, and turns that into hourly visitor counts, live occupancy, dwell time and peak-hour curves. Nobody is identified. The honest trade is accuracy: a purpose-built 3D stereo sensor over the door will beat camera software, and whether that gap matters depends entirely on what decision the number feeds.
Almost Everyone Shops This on Accuracy. That Is the Wrong First Question.
Every footfall vendor leads with an accuracy percentage, so buyers line the percentages up and pick the highest one. The percentage only matters in proportion to the decision it feeds, and footfall numbers feed two very different kinds of decision.
If footfall is the denominator of a conversion rate, small counting errors get multiplied by everything you build on top. A 10 percent undercount at the door makes conversion look 11 percent better than it is, in every store, forever, and you will make merchandising and staffing calls on a number that is quietly wrong. That job wants a purpose-built sensor over the door and a documented accuracy figure you have validated by hand.
If footfall is an operational signal, the shape of the curve carries the decision and a few percent of drift changes nothing. Knowing that Saturday afternoon runs three times Tuesday morning tells you where to put labor whether the true peak is 300 people or 330. Queue length, dwell time and live occupancy behave the same way. For that job, software on the cameras you already own is enough, and it costs a fraction of a sensor rollout.
So the first question is not which system is most accurate. It is whether anyone downstream is going to divide by this number. Retailers building conversion reporting should buy the sensor. Operators staffing to traffic, managing queues, proving lease-negotiation traffic or watching occupancy should start with the cameras already installed and only add sensors where a doorway proves difficult.
What Each Counting Technology Costs and What It Actually Delivers
US market ranges for 2026. Accuracy figures are what the category typically achieves at a normal single-door entrance, not a best case on a calibrated test rig. Wide entrances, glass frontage, direct sun and groups walking abreast pull every one of these numbers down.
| Technology | Hardware Cost | Typical Accuracy | Best Fit |
|---|---|---|---|
| Break-beam / infrared | $200 to $500 per door | 85 to 95% | Narrow single doorways, low traffic |
| Thermal sensor | $500 to $1,500 per door | Around 85% | Privacy-sensitive sites, no imagery kept |
| 3D stereo sensor | $800 to $2,000 per door | 98%+ (category leader) | Conversion-rate reporting, retail KPIs |
| Dedicated AI camera counter | $1,000 to $3,000+ per door | 90 to 97% | Wide entrances needing extra analytics |
| Software on your existing cameras | $0 new hardware | 70 to 90% | Operational counts, occupancy, dwell, queues |
Add roughly $100 to $500 per device for installation on any hardware option, $20 to $100+ per sensor per month for the cloud analytics layer, and 8 to 12 percent of hardware value per year for support. Those recurring lines are why a sensor rollout across a chain costs far more than the sticker suggests, and why the camera-software route is usually the cheapest way to find out whether anyone will use the data at all.
Who Actually Publishes a People Counting Price
Most of this market quotes rather than publishes, which makes budgeting hard before you are already in a sales cycle. The figures below are the ones vendors put in public, checked directly rather than repeated from roundups.
V-Count
Publishes hardware and software separately. The Nano sensor runs about $299 to $799 per unit one-time, with the BoostBI analytics layer at roughly $9 to $49 per month.
Density
Waffle hardware about $229 per unit, software about $8 per unit per month for rooms on annual billing and $15 monthly. Worth knowing this is workplace occupancy, not a retail doorway counter.
RetailNext
Publishes no pricing at all. Neither do most enterprise footfall platforms. If a comparison article quotes you a firm RetailNext number, it was not taken from RetailNext.
Prices move. These were checked against vendor material and are worth re-checking before you build a budget on them.
Occupancy Counting Is Harder Than Footfall Counting, and the Reason Catches People Out
Footfall is a count of events. Live occupancy is a running balance: everyone who came in, minus everyone who went out. That difference has a nasty property. A counting error on a footfall report is a one-off, and tomorrow the report starts clean. A counting error in an occupancy balance never clears. It sits in the running total until somebody resets it.
The usual cause is not accuracy at all. It is an uncounted door. A fire exit propped open on a warm afternoon, a staff entrance nobody thought to include, a loading bay used as a shortcut: each one leaks in a single direction and the occupancy figure drifts steadily away from reality over the day. Every real deployment we have seen that produced nonsense occupancy numbers had a door missing, not a bad algorithm.
So if you need live occupancy, count every opening a person can use, not just the customer entrance, and schedule a nightly reset to zero when the building is empty. If you only need footfall, you can cover the main doors and accept that a side entrance is missing from the total. Buying occupancy and covering one door is the single most common way this project disappoints.
How Camera-Based Footfall Counting Is Set Up
Connect the camera stream
Point the platform at the existing RTSP or ONVIF feed. No new hardware at the door and no rewiring, which is what keeps a multi-site pilot cheap enough to actually run.
Draw the counting line
Set a virtual line across the doorway and mark which side is inside. Direction is what separates an entry from an exit, so getting the orientation right matters more than the exact placement.
Validate against a hand count
Count a busy hour manually and compare. Retail buyers commonly hold a counter to 95 percent or better against a hand count before trusting anything built on top of it. Do this before you report the numbers upward.
Report and alert
Hourly and daily counts per door and per site, live occupancy, dwell and queue length, with thresholds that notify a manager when a limit is crossed.
Where a Doorway Sensor Still Beats Camera Software
We sell the software route, so here is the honest version of when not to buy it.
| Situation | Better Choice | Why |
|---|---|---|
| Conversion rate reporting | 3D stereo sensor | Footfall becomes a denominator, so counting error propagates into every downstream KPI. |
| Camera aimed along the aisle, not at the door | Add a camera or a sensor | Counting needs a clean crossing line. A shallow oblique angle makes people occlude each other. |
| Wide mall entrance, groups walking abreast | Overhead sensor | Overhead views separate individuals that a side-on view merges into one blob. |
| Strong backlight or all-glass frontage | Thermal sensor | Silhouetting defeats visible-light counting for part of the day, every day. |
| Staffing, queues, dwell, peak hours, occupancy | Camera software | The curve shape drives the decision, and no new hardware means you can cover every site at once. |
| Proving traffic across many sites cheaply | Camera software | A chain can instrument every store in days for the cost of licenses rather than a hardware rollout. |
What the Same Cameras Give You Once They Are Connected
This is the practical argument for the software route. The counting license is rarely the only thing a site wants from its cameras, and the feed is already connected.
Dwell time by area
How long people linger in a zone, which is usually a better merchandising signal than the door count on its own.
Queue length and wait alerts
Notify a manager when the checkout line passes a threshold, which is the footfall-adjacent feature operators use most.
Live occupancy limits
A running in-minus-out balance per building, with the nightly reset that keeps it from drifting.
Natural-language search
Ask for a moment in plain English instead of scrubbing hours of recording to find it.
After-hours motion alerts
The same feed covers the security job outside trading hours, which is what most of these cameras were bought for.
Multi-site comparison
Rank sites on traffic and conversion-adjacent measures without a separate hardware budget per location.
Questions Buyers Ask About Footfall Counter Software
What is a footfall counter?
A footfall counter is a system that measures how many people enter, leave or move through a physical space. It can be a dedicated sensor mounted above a doorway or software that reads an existing security camera feed. Either way it produces counts and traffic patterns, not identities.
How does footfall counting work?
A sensor or camera watches a defined line across an entrance and classifies each crossing as an entry or an exit based on direction of travel. Camera-based systems detect the shape and movement of a person in the video frame. The system stores only the count and the timestamp.
How accurate is people counting software?
Software running on existing security cameras typically lands between 70 and 90 percent at a normal entrance, while a purpose-built 3D stereo sensor over the door reaches 98 percent or better. Retail buyers commonly require 95 percent or better against a manual hand count before trusting the data.
Can I use my existing CCTV cameras for footfall counting?
Usually yes, provided the camera sees the doorway with a reasonably clear crossing line and outputs RTSP or ONVIF. The common blockers are a camera aimed down an aisle rather than at the entrance, heavy backlighting from glass frontage, and a very wide entrance where groups overlap.
How much does footfall counting software cost?
Camera software avoids hardware entirely, so the cost is the platform license. Sensor routes add $200 to $2,000 per door for hardware, $100 to $500 per device to install, $20 to $100+ per sensor per month for analytics, and 8 to 12 percent of hardware value per year in support.
What is the difference between footfall counting and occupancy counting?
Footfall counts events over a period, so an error affects one report. Occupancy is a running balance of entries minus exits, so an error persists until reset. Occupancy also requires every usable door to be counted, including staff and fire exits, or the number drifts all day.
Does footfall counting identify individual people?
Counting does not require identification. The system decides that a pattern of movement crossing a line is one person, and records a number and a time. If you are evaluating any system that adds face-based features, treat that as a separate decision with its own state privacy questions.
Is footfall counter software worth it for a single store?
For one location the honest answer is that the value comes from acting on the curve. If nobody will change staffing, opening hours or layout based on the traffic pattern, the data will sit unread. Chains get more from it because comparison across sites creates the decisions.
Related Reading
For the wider market survey, our roundup of the best people counting software ranks the platforms and sensor vendors side by side, and how much people counting sensors cost breaks the hardware budget down device by device. If accuracy is the sticking point, people counting accuracy explains how the percentages are measured and why entrance conditions move them so much. Teams that want the counts inside a wider analytics view should look at people counting analytics, and anyone weighing the software layer on its own can compare people counting software.
Count Your Doors Without Buying Sensors
Connect an existing camera, draw a line across the entrance and see the counts the same day. Validate it against a hand count before you commit to anything.