People Counting Validation Worksheet

First define what the number means

A camera or sensor reporting people counts can support staffing, space planning, queue analysis, and building operations. But an entry count, an exit count, and an occupancy estimate measure different things. Before buying a dashboard or trusting a monthly report, write a plain-language question: how many visitors crossed the main entrance during staffed hours, or how many people were present in a defined zone at a given time? State what the measurement does not include. A dashboard is only useful when its units, limits, and counting boundary are understood by the people making decisions.

Draw the counting line and list alternative routes

On a floor plan, mark the exact doorway or virtual line and the direction assigned to entry and exit. Note secondary doors, stairwells, shared lobbies, staff entrances, delivery paths, and emergency exits. A single sensor at the front door cannot reliably represent whole-building occupancy when people use other routes. Decide whether employees, delivery workers, repeat visitors, carts, and children should count under the intended business definition. Record these decisions before testing so a later discrepancy is not explained away by changing the rules after results are visible.

Build a short manual comparison sample

Select a manageable observation window that covers ordinary activity and, if possible, a distinct busy period. Assign an authorized observer, document the starting and ending times and the counting rule, and maintain the organization’s privacy procedures. Use a simple tally sheet for entry and exit events rather than collecting unnecessary identity information. Ensure the dashboard and observer use the same clock and time zone. When the comparison period ends, retain the raw count, displayed count, and a note about unusual conditions. Avoid extrapolating a five-minute result into a guaranteed accuracy rate for every day.

Look for specific sources of disagreement

People walking shoulder to shoulder, groups reversing direction, a person stopping directly under a sensor, doors propped open, reflective surfaces, moving carts, and poor lighting can complicate counting. Different sensor designs respond differently. Log conditions observed when a discrepancy occurred instead of assuming one universal cause. Compare entry and exit reports separately when both are available. If a queue forms directly across the boundary, identify whether that physical layout should be addressed before changing analytics thresholds. Have a responsible technical reviewer assess settings and placement instead of changing several variables at once.

Distinguish flow measurements from occupancy estimates

Counting entrances and exits can support an estimated in-building total, but errors accumulate if a route is unmeasured or a counter misses an event. Establish the initial occupancy value and a policy for reconciliation; for example, compare with a known empty-building state when appropriate and authorized. Never use an unvalidated estimate as the sole basis for an emergency evacuation or life-safety decision. If multiple doors feed one occupancy display, confirm which data streams are included and whether events from the same person can be counted twice. Reports should make assumptions and uncertainty visible.

Protect privacy and limit dashboard access

Ask whether the chosen system stores identifiable video or only anonymous count events, and do not infer anonymity from a vendor’s use of the word analytics. Document which data are recorded, where they are retained, who can access them, and how long they remain available. Check relevant workplace, tenant, and visitor notification requirements with the appropriate adviser. Aggregate reporting may accomplish the business goal with less sensitive data than tracking individual paths. The operating team should understand the distinction before enabling new features or sending reports outside the organization.

Create an ongoing quality-control routine

Repeat a manual spot check after camera relocation, firmware changes, lighting upgrades, entrance remodeling, or changes in traffic flow. Keep a version log for analytics rules, device placement, and report definitions. Track missing data or offline periods rather than treating them as days with zero visitors. When using counts for staffing or lease planning, include the sample size and known uncertainty in any presentation. Use trend comparisons only when the counting method is stable across the periods being compared. A documented correction is more useful than an impressive but unexplained number.

Create a sign-off record that others can reproduce

Include boundary photo or drawing, direction definitions, included populations, observation windows, manual tallies, dashboard output, known exceptions, approving manager, and the next validation date. Give each follow-up action a due date. For related discussion of business activity analytics within wider security operations, see NERSA’s business analytics resource. This page focuses on validating a count and preventing misleading conclusions, not on selling or installing a particular counting system.

Simple validation worksheet

Site and doorway; purpose of count; approved definition of entry, exit, or occupancy; alternative routes; data-collection method; sample dates and times; observer and reviewer; manual total; reported total; measured difference; observed circumstances; privacy owner; corrective action; and next comparison. A manager should be able to explain the result without relying on a vendor demonstration.

Check the scene and review method before relying on counts. Continue with Camera Coverage Walkthrough: A Property Worksheet and Transaction Video Review Preparation Checklist.