Grocery, convenience and FMCG · Physical Operations Intelligence

Where does configured store flow become queue pressure—and when can teams respond?

Queue pressure is often visible only after waiting has already spread. In a controlled pilot, supported views can keep entrance load, configured zone demand, service-counter coverage, and checkout waiting in one operating context.

Plan a Spatial Audit

First, we verify which camera views can reliably support each entrance, aisle, category, campaign zone, service point, and checkout area.

Evidence boundaryCameras can describe anonymous movement, dwell, queue, waiting, physical continuation, and role-level coverage. Basket, SKU, stock, sale, revenue, shelf availability, planogram compliance, and planned staffing require the relevant connected system.

Wide grocery store with aisles, campaign endcaps, checkout queues, and anonymous flow. — Illustrative operating view. No customer result or measured KPI is shown.

Camera-native grocery operations scope

Build the store-flow view from camera-verifiable operating signals.

The list below is a starting point, not a metric ceiling. A useful KPI begins with the store area, operating condition, and decision your team needs to understand.

Traffic and store load

Entry, exit, footfall, occupancy, and density
Anonymous groups, peaks, zone dwell, and mapped movement
Entrance distribution and first category reach
A supporting 2D heatmap and a primary 3D operating map

Aisles and campaigns

Aisle flow, category dwell, and physical journey drop-off
Campaign and endcap attention in supported views
Signage or promotion continuation to a named category
Layout, display, store, region, and period comparison

Checkout and service

Queue, estimated waiting, and configured thresholds
Staffed and self-checkout area pressure where covered
Role-level floor and checkout coverage
Configured pilot conditions with location, duration, priority, and reviewed action

Context and comparison

Physical continuation toward checkout
Chain comparison under the same operating question
Connected POS, inventory, planogram, cart, and schedule context
Location-specific layout and operating recommendations for review

Dashboard action signalsChain comparisonDashboardAPIsConnected retail context

Before the queue becomes obvious

Where does configured store flow begin to wait or build pressure?

Keep the store fixed. Change the question to see whether flow first weakens at the entrance, a campaign zone, an aisle, floor support, or the checkout approach.

Store question

Comparison view

Read the selected question inside one mapped store.

Where does configured checkout pressure become waiting?

Read arrivals to the checkout area, queue formation, estimated waiting, and nearby role-level coverage in the same physical view.

What the camera can read

Checkout arrivals
Queue and estimated waiting
Checkout-area load and role-level coverage

Checkout pressure

01Checkout approach
02Queue formation
03Waiting area
04Service point

Camera reads

Suitable views can read entry and exit, footfall, occupancy, density, dwell, anonymous groups, mapped movement, campaign or endcap attention, continuation to a named category, queue, estimated waiting, physical journey drop-off, and role-level team coverage.

Connected systems confirm

POS, cart, inventory, supported shelf vision, planogram, and schedule systems confirm basket and SKU context, transaction and sale, stock and shelf state, plan compliance, replenishment records, and planned staffing when the operating question requires them.

The page does not identify shoppers, infer a basket or purchase from movement, or score an individual employee. Camera suitability and the supported view determine which zones and events can be read reliably.

Configured pilot condition

Selected checkout areaQueue threshold and durationAttention

Queue and estimated waiting remain above the configured threshold while available checkout coverage is limited.

Manager-reviewed action

Review opening capacity, balancing staffed and self-checkout support, or routing available team coverage.

Location-specific learning

Keep each store’s configured operating context distinct.

In a standard deployment, camera images are processed on the on-site Edge PC. Continuous raw video is not sent over the WAN. Required anonymous operational metadata may synchronize to the cloud so model weights can learn recurring entrance, aisle, category, coverage, and checkout patterns for this store.

Learns the store

Keeps recurring flow, waiting, coverage, and layout patterns specific to this location.

Presents a recommendation

Layout and operating recommendations go to manager review. Nothing is applied automatically.

Compares the chain

Stores are compared under the same operating question while every location-specific model remains separate.

Start with camera suitability

Map the store-flow question you need to answer.

A Spatial Audit identifies usable views, blind areas, store zones, queue boundaries, live conditions, connected-system requirements, and the first operating question worth measuring.

Plan a Spatial Audit

All six grocery questions remain available without interaction

How does observable entrance load distribute across configured zones?Read covered entry and exit, first mapped direction, and continuation toward the named store areas without identifying the shopper. Entry and exit by covered entrance · First mapped direction · Continuation toward named categories Entrance load remains above the configured range while continuation to the intended first zones becomes uneven. Review the active entrance, wayfinding, and floor coverage while the condition remains live.
Does physical movement continue from campaign material to the configured category?Connect observed attention at the supported campaign or endcap zone to anonymous continuation toward the linked category. Campaign or endcap-zone attention · Named category approach · Paths that continue or peel away The campaign zone remains active while anonymous continuation to the named category stays below the configured range. Review placement, sightline, and the physical route before changing the campaign or linked category plan.
Where does category-area dwell meet blocked aisle flow?Read density, movement, and dwell together so a busy category can be separated from a physical bottleneck in the aisle. Aisle density and directional flow · Category dwell · Movement around the named zone Density remains above the configured threshold while directional flow slows around the selected category. Review display position, aisle clearance, and available floor support while the pressure is still active.
Where is configured zone demand active without visible role-level coverage?Read where anonymous demand is forming and whether the required operational role is present within the supported nearby coverage area. Demand by mapped store area · Role-level team presence · Coverage gaps around active demand Demand remains active beyond the configured duration while the required role-level coverage is not present nearby. Review available floor coverage and route support to the active zone without scoring an individual employee.
Where does configured checkout pressure become waiting?Read arrivals to the checkout area, queue formation, estimated waiting, and nearby role-level coverage in the same physical view. Checkout arrivals · Queue and estimated waiting · Checkout-area load and role-level coverage Queue and estimated waiting remain above the configured threshold while available checkout coverage is limited. Review opening capacity, balancing staffed and self-checkout support, or routing available team coverage.
Where does the mapped journey pause before checkout?Read anonymous paths that reach the checkout approach and paths that peel away. This is physical continuation, not a purchase or sale result. Journey reach to the checkout approach · Paths that continue · Paths that leave the mapped checkout route Anonymous journeys repeatedly reach the checkout approach but do not continue through the mapped physical route. Review the route, queue condition, and coverage, then use POS only if the question needs a recorded transaction outcome.

Stores, regions, and equivalent periods can be compared under the same selected operating question.

How does the same store-flow question change across the chain? Select a comparable entrance, category, campaign zone, aisle, checkout area, or operating window. Review the same camera-observed evidence across stores, regions, formats, and equivalent periods.

Camera reads: Suitable views can read entry and exit, footfall, occupancy, density, dwell, anonymous groups, mapped movement, campaign or endcap attention, continuation to a named category, queue, estimated waiting, physical journey drop-off, and role-level team coverage.

Connected systems confirm: POS, cart, inventory, supported shelf vision, planogram, and schedule systems confirm basket and SKU context, transaction and sale, stock and shelf state, plan compliance, replenishment records, and planned staffing when the operating question requires them.

Build the store-flow view from camera-verifiable operating signals. Entry, exit, footfall, occupancy, and density · Anonymous groups, peaks, zone dwell, and mapped movement · Entrance distribution and first category reach · A supporting 2D heatmap and a primary 3D operating map · Aisle flow, category dwell, and physical journey drop-off · Campaign and endcap attention in supported views · Signage or promotion continuation to a named category · Layout, display, store, region, and period comparison · Queue, estimated waiting, and configured thresholds · Staffed and self-checkout area pressure where covered · Role-level floor and checkout coverage · Configured pilot conditions with location, duration, priority, and reviewed action · Physical continuation toward checkout · Chain comparison under the same operating question · Connected POS, inventory, planogram, cart, and schedule context · Location-specific layout and operating recommendations for review

You define the operating question. We shape the KPI with you. Define the important areas and the decision you need to make. We combine camera-verifiable signals into a tailored KPI and add approved POS, inventory, planogram, cart, or schedule context only when the question requires it.

Location-specific learning: In a standard deployment, camera images are processed on the on-site Edge PC. Continuous raw video is not sent over the WAN. Required anonymous operational metadata may synchronize to the cloud so model weights can learn recurring entrance, aisle, category, coverage, and checkout patterns for this store.