Heat map analysis

Your floor space is producing data every second. Customer foot traffic across each aisle, dwell time at each display, bounce-back rate at the bar, queue length at the register. Most retail and hospitality operators have never seen any of that as a number — just as a vague sense of "the back corner is dead." A heat map turns the vague sense into a specific dollar value.

What a heat map actually shows

Modern heat-mapping pulls three signals from your existing cameras and assembles them into a per-zone picture:

• Foot traffic count — how many bodies pass through each zone per hour
• Dwell time — how long they stand or sit, on average, in each zone
• Conversion rate — what % of dwellers in a zone become a transaction (when fused with POS data per category)

The output is a colored overlay on your floor plan: green zones convert, yellow zones bottleneck, red zones leak. You stop guessing which 30% of your floor produces 70% of revenue and start knowing.

Retail use case: the dwell-without-conversion gap

The most valuable signal in retail is dwell-without-conversion — customers who stand in front of a display, consider, and walk away empty-handed. This is the gap between interest and purchase. It's where pricing, signage, or product placement is killing the sale.

A specialty retailer we analyzed had a $4,200 weekly average sales gap on one specific end-cap. The heat map revealed that aisle had the highest dwell time in the store — people were stopping — but the lowest conversion. After 30 minutes on-site:

• Price tags were 12-point font; older customers couldn't read them
• The cross-sell sign was below sightline (hidden by the upper shelf)
• Inventory was facing the wrong direction — UPC out, not branding

Three changes, zero new spend, two weeks later: that end-cap moved from 4% conversion to 11%. Annualized incremental revenue: ~$30K.

Restaurant use case: table turn + bar bottleneck

Restaurant heat maps look different. The signals that matter:

• Table dwell time — how long after the bill arrives do guests stay (predicts table turn)
• Bar queue length — how many people standing waiting for a drink (predicts the order-fewer-rounds phenomenon)
• Server route efficiency — which sections take longest to walk between

The bar-queue signal is usually the most actionable. If your data shows 6+ people consistently waiting at the bar between 7:00 and 7:30 PM, you're losing tickets to impatience. The fix is usually one extra bartender for that 30-minute window. The cost of the labor is dwarfed by the recovered second-round orders.

Multifamily use case: amenity ROI

A multifamily property runs amenities on the assumption they get used. They probably don't. Camera + access-control fusion produces per-day visitor counts by amenity. Common findings:

• Fitness room: 40-80 visits/day (works)
• Coworking lounge: 5-15 visits/day (questionable)
• Rooftop deck: 20-40 visits/day (works for marketing)
• Game room: 1-5 visits/day (kill it)

The game room is costing $400-600/month in HVAC, cleaning, and insurance. Closing it — or converting it to a more-used amenity — is a $5K-$7K annual win that property managers find inside their first month of heat-map data.

Privacy: nothing identifiable leaves your premises

The heat map is built from counts and dwell times, not identities. Raw video stays on the Sentinel device inside your network. What ships to the dashboard is structured numbers: "5 bodies in zone A, average dwell 22 seconds." No faces, no plates, no PII.

This is a meaningful difference from cloud-based analytics services (which upload your raw video to their servers). Edge-AI processing means your customer privacy posture stays clean — important for residential properties and any business with CCPA exposure.

What it costs

A heat-map deployment runs as part of a Sentinel Standard tier install: $15K hardware + $500/mo Sentinel Cloud. The map renders inside the Insights / BIE dashboard alongside the rest of your operating data. Vendor-agnostic on cameras — works with whatever you already have.

Typical payback: 60-90 days for a single-location SMB based on the dwell-without-conversion lift alone.