Automatic Door People Flow Counting: A Procurement Guide for Commercial and Retail Environments

Every retail store, shopping center, airport terminal, and commercial building has a people flow problem. Decisions about staffing levels, store layout, opening hours, and lease negotiations are made on the basis of incomplete or inaccurate information about how many people are actually moving through the space.

People flow counting systems integrated into automatic door systems change this equation fundamentally. But 'door with a counter' covers an enormous range of technical capability. A basic infrared beam counter and a high-accuracy 3D vision system with API connectivity are both technically 'people flow counters' — but the gap between them matters enormously for the business value extractable from the data.

This guide helps operations managers and procurement teams understand the technology options, accuracy implications, integration requirements, and common pitfalls to avoid.

Part 1: Why 'Bring a Counter' Is Not the Same as 'People Flow Analytics'

A basic infrared beam counter counts pulses — interruptions of an infrared beam. It cannot distinguish between one person and two people walking side by side. It cannot handle a child who passes under the beam, a shopping cart, or two people walking close together. In typical retail environments, basic infrared counters achieve accuracy rates of 70-85% — meaning 15-30% of all events are miscategorized.

At 75% accuracy, a count of 1,000 events could represent anywhere from 750 to 1,250 actual people. For rough traffic estimation, this may be acceptable. For operational decisions with regulatory or contractual implications, it is not.

Part 2: The Three Technology Approaches

Approach 1: Dual-beam Infrared (Traditional)

Accuracy range: 75-88% under typical conditions. Performance degrades in crowded conditions, with children, with wheelchair users, and with slow-moving groups. Advantages: low cost, simple installation, no privacy concerns. Best suited for low-budget applications where rough traffic estimation is sufficient.

Approach 2: 3D Vision (Time-of-Flight or Stereo Camera)

Accuracy range: 95-99% under typical operating conditions. Handles groups, children, varying speeds, and bidirectional traffic substantially better than infrared. No privacy concerns with depth-only systems (no recognizable images captured). Best suited for retail stores, shopping centers, airport halls, commercial building main entrances — any application where accurate counts are operationally important.

Approach 3: Millimeter-Wave Radar

Accuracy range: 95-98% under typical conditions. Completely immune to lighting — works in total darkness, rain, fog, and temperature extremes. Inherently privacy-preserving (detects presence and movement without image capture). Best suited for outdoor entrances, transit hubs, environments where camera installation raises privacy concerns, or applications requiring reliable operation across extreme environmental conditions.

Part 3: The Analytics Chain — What You Can Actually Do With the Data

Layer 1: Traffic Volume and Pattern Analysis

Traffic volume reports enable staffing optimization, opening hours rationalization, and marketing effectiveness measurement. If analysis shows Saturdays 2-5 PM account for 35% of weekly traffic, staffing can be concentrated accordingly. Comparing traffic during promotional periods against baseline provides direct measurement of foot traffic uplift from marketing spend.

Layer 2: Occupancy and Density Management

Net occupancy (cumulative entry minus cumulative exit) maintained in real time enables capacity management for spaces with regulatory limits, customer experience management through queue activation and flow guidance, and zone-level density analysis across multiple entrances.

Layer 3: Conversion Analytics and Strategic Decision Support

For retail, combining people flow data with POS transaction data enables conversion rate analysis by time period, day of week, and staffing scenario — insights neither dataset could reveal alone. At portfolio level, multi-location traffic data informs lease negotiation (using verified footfall rather than estimated catchment), store format decisions, and new location evaluation.

Part 4: Integration Requirements — What to Specify

• API access: require a documented, open REST API with JSON output. Avoid proprietary protocols requiring specific software tools.

• Data refresh rate: sub-minute for real-time occupancy management; hourly aggregation may suffice for reporting.

• Historical data retention: minimum 12-24 months for trend analysis and strategic decision support.

• POS system integration: timestamp precision sufficient for conversion rate correlation.

• BMS integration: for larger commercial properties, HVAC and security coordination based on occupancy data.

• BI platform integration: ability to route data into existing Tableau, Power BI, or Looker environments.

Part 5: Five Common Procurement Mistakes — and How to Avoid Them

• Specifying 'people flow counter' without specifying accuracy. Fix: require minimum 95% accuracy in documented field conditions with third-party test results.

• Assuming the platform is more capable than it is. Fix: require a live demonstration using real data from a comparable reference installation.

• Underspecifying integration requirements. Fix: engage the teams who will consume the data before finalizing specifications.

• Ignoring ongoing platform subscription costs. Fix: request a complete schedule of all ongoing software costs; evaluate 10-15 year total cost of ownership.

• Treating all entrances as equivalent. Fix: develop a per-entrance analysis mapping traffic volume and analytics requirements to the appropriate technology tier.

People flow analytics capability is rapidly becoming a standard requirement for automatic door systems in commercial environments. For a medium-sized retail property with 500,000 annual visitors, even a 1% improvement in conversion rate from data-driven decisions can justify the people flow system investment within 12-18 months.