Real-Time Occupancy and Queue Management: A Practical Guide

Occupancy and queue management is most useful when it changes what your team does in the next ten minutes. A live occupancy number tells you the floor is filling up. A wait-time number tells you the checkout is backing up. On their own, they are just figures on a screen. The value comes from the staffing or safety action each one triggers.

This guide explains the three measurements retailers and venue operators most often confuse: footfall, occupancy, and queue. For each one, you will see how it is measured from cameras you already own, what a sensible target looks like, and the action it should set off.

Every metric here is counted anonymously from your existing IP cameras over RTSP or ONVIF. There is no facial recognition, and the video is processed on-site rather than sent to a cloud. We will come back to that, but it is worth saying up front, because the privacy model shapes what these systems can and cannot do.

Key takeaways

• Footfall, occupancy, and queue answer three different questions. Pick the metric that matches the decision you are making.
• Live occupancy is a running count of entries minus exits, measured anonymously from your existing cameras with the video kept on-site.
• A capacity limit only works if you set an early-warning band below it, so staff act before a space feels crowded.
• Queue analytics measure length, wait time, and service time separately, which is what tells you when to open another lane.
• Every metric should map to one clear staffing or safety action with a named owner.

Occupancy vs. queue vs. footfall: three different questions

These three terms get used interchangeably, and that is where bad decisions start. They measure different things over different time windows.

• Footfall is how many people came in over a period, such as a morning, a day, or a week. It is a cumulative count. You use it to plan ahead: rotas, opening hours, marketing spend.
• Occupancy is how many people are inside right now. It is a live snapshot that rises and falls through the day. You use it to manage the space in the moment.
• Queue is how many people are waiting at a service point and how long they wait. You use it to decide when to open another lane or move staff to the front.

A worked example makes the difference clear. A store might record 1,200 visits across the day (footfall), peak at 180 people inside at once around lunch (occupancy), and see checkout waits climb past four minutes during that same peak (queue). Each number triggers a different response. You would not staff the whole day to handle a 20-minute lunch spike, and you would not judge a quiet afternoon by the morning's total.

The mistake to avoid is using one number to make every decision. A high daily footfall does not mean the floor was crowded at any point, and a long queue can form even when occupancy is moderate if too few lanes are open.

All three are measured the same way at the source: anonymously, from your existing cameras, with body-only detection and no facial recognition. The video is processed on the premises. The through-line for the rest of this guide is simple. A metric is only worth collecting if it triggers a staffing or safety action.

How live occupancy is counted from entries and exits

Live occupancy uses a running count. At each entrance, the system counts people walking in and people walking out. Subtract exits from entries and you have the number of people inside right now. As customers arrive and leave, the figure updates continuously.

This is why door placement matters. If a store has three entrances, all three need a counting view. Miss one and your total drifts, because people will be leaving through a door the system cannot see. For most layouts, a single overhead camera per door is enough, and many stores already have a camera near each entrance.

It is worth being honest about drift. Over a long, busy day, small counting errors at the door accumulate, the same way a manual clicker would. Crowding, people walking close together, and odd traffic patterns all add a little noise. A daily reset at open or a close-out at the end of trading keeps the figure dependable, so each day starts from zero rather than carrying yesterday's small errors forward.

You will usually want the count to reflect customers, not your own team walking in and out. Staff exclusion filters employees out of the running count, so the occupancy figure tracks the people you are actually trying to serve. You can read how that works on the staff exclusion page.

For the live capacity meter, the history behind it, and how the running count is presented to a manager, see the occupancy feature page.

Setting a capacity limit you can actually act on

A capacity limit is a decision you make, not a number CountPort certifies for you. The system measures how many people are inside. Choosing the safe ceiling is yours, and it should come from three inputs.

• Floor area. Larger usable space supports more people comfortably. Subtract fixtures, back-of-house, and anything customers cannot stand in.
• Comfort. A clothing store browsing experience and a packed sale event tolerate very different densities. Pick a number that keeps the space pleasant, not just legal.
• Local rules. Where capacity rules apply to your premises, factor them in. Treat them as one input to your own decision rather than a setting the software hands you.

Whatever number you land on, do not alert at the hard limit. By the time you hit it, the space already feels crowded and staff are reacting late. Set an early-warning band below the ceiling instead. For example, if your limit is 200, you might warn at 170. That gap gives the team time to manage the door before the space is uncomfortable.

When the warning triggers, the alert can route to a manager's phone or a dashboard on the shop floor. The point is that someone owns the response. Decide in advance who that is: a duty manager who starts metering entries, a greeter who slows the door, or whoever is closest. An alert with no named owner gets ignored.

For the full method on choosing a threshold and wiring up alerts, see How to Set a Safe Occupancy Limit and Get Alerted Before You Hit It.

What queue analytics measure: length, wait time, and service time

A queue is not one number. It is three, and separating them is what makes the data useful. Break any line into three stages.

• Join. A person enters the queue area.
• Wait. The time spent in line before reaching a server.
• Service. The time spent being served once at the counter or till.

Wait time and service time are timed separately, usually from a single overhead camera looking at the checkout zone. The system measures how long the queue area holds people and how quickly they move through it. That distinction matters: a long total wait caused by slow service is a different problem from one caused by too few open lanes, and the fix is different too.

Queue abandonment is the people who join the line and leave before being served. Measured qualitatively, a rising abandonment pattern is a strong signal that waits have crossed what your customers will tolerate. You may not catch every walk-away precisely, but a clear upward trend during peaks tells you the line is costing you sales.

As with occupancy, this is anonymous. The system times the queue area, not named individuals. It is not following a specific person from the door to the till. It is measuring how the zone behaves: how many are in it and how long they take to clear.

For the measurement detail and how the timings are captured, see the queue feature page and the deeper walkthrough in How to Measure Queue Wait Time and Abandonment at the Checkout.

Turning each metric into a staffing or safety action

Here is the part that earns its keep. Each metric should map to one clear action with a named owner. A simple trigger table keeps everyone aligned.

Signal	Threshold (you set it)	Action	Owner
Occupancy rising toward the limit	Early-warning band, e.g. 85% of ceiling	Meter the door, manage flow	Duty manager
Wait time climbing	e.g. crosses 4 minutes	Open another lane or pull a colleague to the front	Front-end lead
Queue forming with lanes closed	e.g. 3+ waiting per open lane	Open a till	Shift supervisor
Quiet hour	Footfall below planned cover	Reduce cover, redeploy to stocking	Store manager

The dashboard is not the point. The action is. A screen full of green and amber tiles changes nothing if no one is responsible for responding when a tile turns amber. Assign ownership for each trigger and review whether the actions actually happen.

To make this concrete, picture a store that opens a second till whenever average checkout wait passes a threshold it has set. When the alert fires, the front-end lead opens the lane and the queue drains. This is illustrative, and your thresholds and results will depend on your layout, staff, and traffic. The pattern, though, holds: measure, set a threshold, name an owner, act.

For turning hourly counts into actual rotas, see Staffing to Footfall: How to Match Rotas to When Customers Actually Arrive.

Where this applies beyond a single store

Occupancy and queue are not only retail problems. They matter anywhere people gather and wait, and often for comfort and safety as much as for sales.

High-flow public venues are the obvious case. A shopping mall watches concourse occupancy to manage crowding and plan cleaning and security cover. A grocery supermarket lives and dies by checkout wait time at the evening peak. A museum or gallery caps gallery occupancy to protect the experience and the exhibits, not to drive transactions. Libraries, gyms, airports, and travel retail all sit in the same family: capacity and flow are the headline concerns.

Then there are venues where the queue is the whole story. In banking, the branch wait is the experience, and a long line is a customer-satisfaction problem before it is anything else. In quick-service restaurants, order-to-collect time and the lunch-rush line decide whether a customer comes back. In those settings, queue analytics are not a nice-to-have. They are the metric that defines a good day.

If you want a fuller treatment of what cameras can and cannot tell you about a line, Queue Management Without Tickets or Apps: What Cameras Can Measure goes into the detail.

Privacy and where the video stays

None of this requires identifying anyone, and CountPort does not. Detection is body-only. There is no facial recognition and no biometric data is stored. The system sees shapes moving through a frame, not faces or identities.

Processing happens on-site. The video stays on the premises and is analysed there, on a back-office machine or supported camera. What leaves the device is the count and the timing data, not the footage. People are counted and timed anonymously, and there is no central video archive being shipped off to a cloud to be re-watched.

That is privacy by design. It is a sensible default for any operator that takes customer and staff privacy seriously, and it sits comfortably with a GDPR-minded approach. We are describing how the system is built, not claiming any regulatory certification on your behalf; how you deploy and document it on your premises is still your responsibility.

For the detection method and deployment detail, see the technology page.

What to set up first, and how it deploys

You do not need to instrument everything on day one. Start where the decisions are clearest.

A sensible first step is occupancy on the entrance, so you have a live count and a capacity warning, and then queue analytics on your busiest checkout. Those two cover the most common in-the-moment decisions: is the floor too full, and is the line too long. You can add zones, dwell, and conversion later once the basics are paying off.

Deployment is software, not new hardware. It runs on a back-office PC, a mini-PC, or a Mac Mini, or directly on supported cameras, reading from the IP cameras you already have. There are no new sensors to mount, no beam counters, and no tickets or apps for customers.

Pricing is flat per camera, which makes a pilot easy to scope. You can put a number on counting one entrance and one checkout before deciding to extend it. See the pricing page for the per-camera rates.

Frequently asked questions

What is the difference between occupancy and footfall?

Footfall is how many people visited over a period, such as a day. Occupancy is how many are inside right now, calculated from entries minus exits. You use footfall to plan ahead and occupancy to manage the space in the moment.

Can cameras measure occupancy and queues at the same time?

Yes. You can run occupancy on your entrance cameras and queue analytics on a checkout camera at the same time, because both are software running on-site against your existing IP cameras.

Does occupancy and queue tracking identify individual people?

No. Detection is body-only with no facial recognition. People are counted and timed anonymously, and all video is processed on-site, so no biometric data is stored.

What action should a long queue trigger?

When wait time crosses a threshold you set, the usual action is to open another lane or move a colleague to the front. The value comes from acting on the alert, not from watching the dashboard.

See it on your own cameras

The fastest way to judge whether occupancy and queue analytics fit your floor is to run them on the cameras you already have. If you want a fuller view of the metrics behind day-to-day decisions, Footfall Analytics Explained is a good companion read, and How to Manage Occupancy During Seasonal Peaks and Sale Events covers the busy periods that stress these numbers most.

When you are ready to see it live, Request a demo and we will set it up against your existing setup.