Solution · Age & gender recognition
Gender & Age RecognitionAnonymous demographic analysis on the Nano AI sensor
V-Count’s AI gender and age recognition runs on the Nano AI sensor itself. As each visitor passes, the sensor estimates an age band and a gender, then discards the frame. Only anonymous metadata goes to BoostBI.
Your footfall count tells you how many people came in. Demographic analysis tells you who they were, so you can match products, staff and marketing to the audience that actually visits.
Three jobs, one Nano AI sensor
Age and gender recognition is not a separate camera system. It runs on the same Nano AI ceiling sensor that counts your visitors and keeps your staff out of the count. One device per entrance gives you all three, in one data stream and one BoostBI account.
Explore Nano AIPeople counting
Entries, exits and live occupancy with up to 99% accuracy at an indoor entrance line. Every demographic figure sits on top of this count.
People countingAge & gender recognition
An estimated age band and gender for each counted visitor, produced on the device. No photo, video or facial template is kept.
How it worksStaff exclusion
Staff wear a reflective shoulder tag or lanyard and Nano AI keeps them out of the visitor count, with 90%+ staff exclusion accuracy.
Staff exclusionOne sensor per entrance, one data stream, one BoostBI account: no second camera network, no images to transmit and nothing to retain.
How AI gender and age recognition works
AI gender recognition and AI age recognition run on the sensor itself. A Nano AI unit combines a people counting camera with an on-board AI processor, so the visual analysis happens on the device rather than in the cloud.
A visitor passes under the sensor
Nano AI is mounted on the ceiling above an entrance or zone. The same sensor that produces your footfall count also produces the demographic attributes attached to it.
The AI-on-chip estimates age band and gender
The on-board processor analyses the camera frame, estimates an age band and a gender classification, and then discards the frame.
Only anonymous metadata leaves the device
What leaves the sensor is a row of anonymised metadata: a timestamp, a direction of travel, an estimated age band and an estimated gender.
- No photograph
- No video clip
- No facial template
- Nothing that identifies or re-identifies a visitor
BoostBI joins it to the rest of your data
BoostBI places the demographic split next to footfall, dwell time, occupancy, queue metrics and sales, so you can read conversion rate per segment.
Estimates, not identities: estimates are probabilistic and reported in bands rather than as exact ages. That keeps them useful for planning while remaining non-identifying. The counting technology underneath is explained on our people counting page, and the hardware on the Nano AI people counting sensor page.
Demographic analysis for physical locations
Digital teams have had audience breakdowns for years. Demographic analysis brings the same view to stores, shopping malls, airports, museums, showrooms and venues, using the doorway and zone sensors you already need for footfall.
Customer demographics data, structured on four axes
The demographic attributes come from the same sensor and the same pass as your footfall count. That means you can read them at the same level of detail: by hour, by entrance and by zone.
- Age bands, never an exact age
- Estimated gender, aggregated rather than held per person
- Read by time and by zone, like your footfall
- Not facial recognition: no one is identified or re-identified
Age band
Visitors grouped into broad ranges such as child, young adult, adult and senior, never an exact age.
Gender
An estimated classification per counted visitor, aggregated rather than held per person.
Time
The same interval granularity as your footfall data: hour, day, week, season or campaign period.
Zone
Entrance, department, aisle, floor, tenant unit or queue area, depending on where sensors are placed.
What a footfall count answers, and what age & gender adds
| Question | Footfall count only | With age & gender recognition |
|---|---|---|
| How many people came in, and when? | Yes | Yes |
| Which age bands and genders visited, by hour and zone? | No | Yes |
| Which audience converts in a zone, and which walks through without buying? | No, one site-wide conversion rate | Yes, conversion rate per segment with sales data connected |
| Did a campaign aimed at an age band or gender change who walked in? | Total visits only | Yes, the mix as well as the total |
| What is the visitor profile per entrance, floor and time of day? | No | Yes |
From counting people to understanding them
On its own, a demographic split is interesting. It becomes decision-grade once it is joined to the rest of your data in BoostBI.
What BoostBI reports
- The demographic split of counted visitors by age band and estimated gender, by time interval and by zone or entrance
- Held alongside footfall, dwell time, occupancy, queue metrics and any sales data you connect
- Conversion rate per segment instead of one site-wide number
- The visitor profile of two stores or two floors, side by side
- How the mix shifts through the day
- Export of customer demographics data to your own BI or data warehouse through the API
Footfall, conversion and customer demographics data in one view is the practical difference between counting people and understanding them. It is the same reporting layer described on our retail store analytics page.
Explore BoostBIDemographic profiling of customers: use cases
Demographic profiling is not an end in itself. It is an input to decisions that were previously made on assumption.
Merchandising and assortment
Match ranges, planograms, category adjacency and product placement to the audience that actually visits each store, not to the chain-wide average shopper.
Staffing and scheduling
Roster by segment as well as by volume, so the hours when a particular audience dominates are covered by the staff and language skills best suited to them.
Marketing attribution
Check whether a campaign aimed at a specific age band or gender changed the mix of people walking in during the campaign window, not just the total count.
Media and DOOH audiences
Report the size and composition of the audience passing a digital screen or advertising site, so inventory can be priced and verified on measured exposure.
Tenant mix and leasing evidence
Malls use demographic data in two directions. Internally, it shows which audience arrives at which entrance and floor at which hours. That guides events programming, marketing spend, guest services staffing and the placement of promotional space.
Externally, it supports leasing. A prospective tenant wants evidence of the visitor profile near a unit, and measured data is far more persuasive than a survey. It also shows the gaps a new tenant category could fill.
Shopping mallsIndustries we serve
For a longer treatment of the business case, read our guide to everything you need to know about demographic analysis.
Privacy and compliance in AI demographic analysis
AI demographic analysis is only worth deploying if it is defensible. V-Count’s approach is built on three principles.
Anonymisation by design
The sensor performs inference on the frame and discards it. No face images, video or biometric templates are written to storage or sent to the cloud.
No identification
The system estimates attributes of an anonymous passer-by. It cannot recognise, match or re-identify an individual across visits or locations. It is not facial recognition and is not intended to be used as such.
Aggregation
Reporting in BoostBI is built from counts and shares by segment, not from records about named people.
What the law looks at, and what you should check
Under the GDPR and the UK GDPR, the analysis of data that does not identify a natural person and is not retained in identifiable form falls outside much of the regime that applies to personal data. The system does not process special-category biometric data for the purpose of uniquely identifying someone.
V-Count is ISO 27001 certified.
Demographic analysis is not facial recognition
| Demographic analysis (V-Count) | Facial recognition | |
|---|---|---|
| Question it answers | Roughly which age band and estimated gender an anonymous passer-by belongs to | Who a person is |
| How | Estimates attributes on the sensor, then forgets the visitor | Matches a facial template against a stored database |
| Facial templates | Not created, stored or matched | Created and matched |
| Same visitor on a later day or at another site | Cannot be re-identified | Matched to the same identity |
| Output | A count with attributes attached | A record about an individual |
| Data protection view | Anonymous metadata | Biometric processing of personal data |
Our diversity approach
Gender estimates are a statistical tool for planning. They say nothing about who a person is.
At V-Count, we value every aspect of human diversity, including all gender identities.
Our technology estimates gender based on visual indicators for statistical purposes only and does not define or assume biological sex or personal gender identity. We are committed to respectful and inclusive data practices.
Estimates are reported in aggregate, by segment, and are never held as a label on a person.
Adding age and gender recognition to your sites
Start at the doors and zones where demographics matter most, and keep one continuous footfall history in BoostBI.
Send a floor plan and a few photos
A floor plan and photos of each entrance are enough. V-Count recommends sensor positions within 24 hours.
If you already have counters, our team reviews your current device list and identifies which doors to convert first.
Mount Nano AI where demographics matter
Nano AI mounts on the ceiling at 2.2–7 m, runs on 5 V USB-C (optional PoE splitter) and connects over Wi-Fi.
A legacy beam-break, thermal or basic 3D counter cannot produce age and gender by software alone. Keep those counters elsewhere: BoostBI ingests both.
Validate placement during commissioning
Lighting, sensor angle, mounting height and occlusion all affect the estimates, so V-Count validates placement during commissioning.
Check the footfall count against a manual reference count, and exclude staff with a reflective shoulder tag or lanyard.
Choose the licence and read the results
Demographics is a separately selectable licence in your BoostBI package, alongside staff exclusion, queues, heatmaps and the API.
Related solutions and products
Gender and age recognition FAQ
GDPR, accuracy, privacy and sensors for AI age and gender recognition with V-Count Nano AI and BoostBI.
Ask an expertIs AI gender recognition GDPR compliant?
It can be, and V-Count’s implementation is designed for it. AI gender recognition on a Nano AI sensor produces an anonymous estimate and discards the raw frame, so no face image or biometric template is stored and no individual is identified.
Because the output is not linked to an identifiable person, much of the GDPR and UK GDPR regime that governs personal data does not apply in the same way, and no special-category biometric processing takes place.
How accurate is AI age recognition?
AI age recognition estimates a broad age band rather than an exact age, and accuracy is highest at the band and aggregate level, which is how it is meant to be used. Any single estimate can be wrong, particularly near the boundary between two bands, but across thousands of visitors the distribution is stable enough to compare stores, hours, zones and campaign periods.
Lighting, sensor angle, mounting height and occlusion all affect results, which is why V-Count validates placement during commissioning. We do not publish a single accuracy figure, because it is meaningful only for a stated band structure and installation condition.
Can demographic analysis work with my existing people counter?
It depends on the hardware. Demographic analysis requires an AI sensor with on-board processing, so a legacy beam-break, thermal or basic 3D counter cannot be upgraded to produce age and gender by software alone.
Most customers add Nano AI units at the entrances and zones where demographics matter and keep their existing counters elsewhere. BoostBI ingests both, so you keep one continuous footfall history rather than restarting your data. If you are planning a phased rollout, our team can review your current device list and identify which doors to convert first.
What is the difference between demographic analysis and facial recognition?
They answer different questions. Facial recognition tries to determine who a person is by matching a facial template against a stored database, which makes it biometric processing of personal data.
Demographic analysis asks only what an anonymous passer-by looks like in broad terms (roughly which age band, which estimated gender) and then forgets them. V-Count sensors do not create, store or match facial templates and cannot re-identify the same visitor on a later day or at another site. The output is a count with attributes attached, not a record about an individual.
What customer demographics data does BoostBI actually report?
BoostBI reports the demographic split of counted visitors by age band and estimated gender, broken down by time interval and by zone or entrance. It holds that alongside footfall, dwell time, occupancy, queue metrics and any sales data you connect.
In practice, you can read conversion rate per segment, compare the visitor profile of two stores or two floors, track how the mix shifts through the day, and export the customer demographics data to your own BI or data warehouse through the API. See BoostBI.
Does AI demographic analysis store face images or video?
No. Each frame is processed on the sensor and discarded; only anonymous metadata leaves the device. There is no image archive, no video retention and no facial template database anywhere in the V-Count stack. That is both a privacy safeguard and a practical benefit, since it removes the bandwidth and storage burden that a camera-to-cloud architecture would create.
Which V-Count sensor gives me a people counter with age and gender recognition?
Nano AI. It is a single ceiling-mounted device that delivers footfall counting with up to 99% accuracy at an indoor entrance line and on-device demographic estimation at the same time, so one sensor covers both requirements at a door or in a zone. See the Nano AI people counting sensor.
For outdoor entrances or large open floors, our team will advise where Nano Outdoor or Nano Prime fits better, and how to mix device types across a site without fragmenting your reporting.
Can one Nano AI sensor count people, estimate age and gender and exclude staff?
Yes. The same Nano AI sensor counts entries and exits, estimates an age band and gender for each counted visitor on the device, and excludes staff who wear a reflective shoulder tag or lanyard, with 90%+ staff exclusion accuracy. All three report to one BoostBI account. See people counting and staff exclusion.
How is demographic profiling of customers used in shopping malls?
Malls use it in two directions. Internally, it shows which audience arrives at which entrance and floor at which hours, which guides events programming, marketing spend, staffing of guest services and the placement of promotional space.
Externally, it supports leasing: a prospective tenant wants evidence of the visitor profile near a prospective unit, and measured demographic profiling of customers is far more persuasive than a survey. The same data also underpins pricing for advertising and DOOH inventory inside the centre.
See age and gender data from your own doorway
Request a demo and we will walk you through a live BoostBI dashboard built on Nano AI sensor data: footfall, age band and gender, with staff kept out of the count, from one sensor per entrance.
Request a Demo