SkyVision Reduces Manual Cost in Footfall Counting, Boosts Store Operations

DaoAI SkyVision's no-code video surveillance AI platform automates footfall counting and behavior recognition, reducing traditional manual counting labor costs by −75% and significantly enhancing the real-time accuracy of store operation analysis, providing data support for smart retail decisions. Driven by AI, the video surveillance industry is expanding from traditional security to diverse applications like smart cities and industrial inspection. In retail and commercial space management, precise footfall counting and heatmap analysis are crucial for optimizing layouts, improving conversion rates, and refining operations. However, efficiently, accurately, and cost-effectively acquiring this data has long been a significant challenge for the industry.

In the daily operations of retail stores, footfall counting and heatmap analysis are critical for understanding consumer behavior, evaluating marketing campaign effectiveness, optimizing merchandise display, and scheduling staff. Traditionally, this data relied on manual counting, door counters, or simple infrared beam devices. As market competition intensifies and consumer behavior becomes more complex, the demand for data accuracy and real-time insights has grown. In the current trend of AI-powered video surveillance expanding into diverse applications, how to acquire this data efficiently, accurately, and cost-effectively, especially regarding on-site deployment and data privacy, has become a key concern for retail businesses. DaoAI recognized this pain point and launched the SkyVision platform.

Pain Points: Why This Hurdle Is So High

Retail stores face multiple challenges in footfall counting and heatmap analysis: First, **manual counting is time-consuming, labor-intensive, and costly**. Employing dedicated personnel for prolonged video surveillance manual counting not only incurs significant labor costs but also leads to statistical errors due to fatigue, especially during peak hours, where accuracy might drop as low as 70%. Second, **traditional counting devices have limited accuracy**. Infrared beams or door counters only provide entry/exit numbers, failing to identify customer movement trajectories, dwell times, and area preferences, let alone distinguish employees from customers, resulting in insufficient data granularity and weak decision support. Third, **real-time data and privacy compliance are contradictory**. Traditional video analysis solutions often require uploading video data to the cloud for processing, which introduces data transmission delays, impacting real-time decision-making. More importantly, it involves sensitive customer behavior data, and uploading to the cloud poses serious data privacy and compliance risks. This pain point is particularly prominent in the current context emphasizing data localization and privacy protection.

The root cause of these problems lies in the traditional solutions' lack of deep semantic understanding of video content. Manual inspection is inefficient and difficult to scale; traditional hardware devices cannot capture complex human behavior patterns. In the pursuit of refined operations and smart retail, retailers urgently need a smart analysis solution that offers high accuracy, real-time performance, low cost, and complete localization to adapt to ever-changing business environments. The advent of DaoAI SkyVision is precisely to address these core pain points and help retail businesses achieve intelligent operational upgrades.

Technical Principles

The core of the DaoAI SkyVision platform lies in its powerful no-code video surveillance AI capabilities and DaoAI World semantic understanding. For footfall counting and heatmap analysis, SkyVision employs advanced deep learning algorithms deployed at the edge to achieve real-time analysis of video streams. Through multi-object tracking technology, the platform can accurately identify and track individuals in video footage, maintaining high-precision tracking even in complex crowd occlusion or varying lighting conditions. By analyzing individual trajectories, DaoAI SkyVision can automatically count pedestrian entries and exits in specific areas, dwell times, and generate detailed heatmaps, visually displaying the distribution of customer activity density within the store. Unlike traditional vision systems relying on rules or manual annotation, the SkyVision platform, based on the DaoAI World model, possesses strong semantic understanding and generalization capabilities, allowing it to quickly learn from a small number of annotated samples and adapt to the characteristics of different store environments, achieving hourly model training and reducing deployment time for new scenarios from weeks to hours.

Compared to traditional manual inspection, the advantage of DaoAI SkyVision lies in its fatigue-free, high-precision, and 24/7 automated analysis capabilities, reducing manual counting labor costs by −75%. Manual counting is limited by attention, with high error rates during peak hours, and cannot provide detailed behavioral data; SkyVision, in contrast, consistently outputs precise data and offers rich behavioral insights. Compared to simple infrared counters, SkyVision not only counts but also understands 'human' behavior, distinguishing employees from customers, and even identifying customer attributes like age and gender (under privacy compliance), providing data support for deeper business analysis. Furthermore, SkyVision supports 100% local deployment, with all video data processed and stored on edge boxes, ensuring data never leaves the premises. This completely resolves data privacy and compliance issues while avoiding latency and costs associated with cloud transmission, ensuring the effectiveness of real-time alerts.

Typical Application Scenarios

• **Store Entrance Footfall Counting and Conversion Rate Analysis**: By deploying DaoAI SkyVision at store entrances, precisely count entries and exits, and combine with sales data to calculate real-time entry conversion rates. Challenges include distinguishing repeat visitors and avoiding double-counting, and maintaining high accuracy under complex lighting and crowd densities.
• **Area Heatmaps and Traffic Flow Optimization**: Deploy cameras in key internal store areas (e.g., product shelves, promotion zones, cashiers), and SkyVision analyzes customer dwell times and movement trajectories in real-time to generate area heatmaps. This helps retailers identify high-attention and inefficient areas, optimizing merchandise layout and customer traffic flow design. Challenges include accurately identifying the boundaries of customer stays in different areas and handling multiple intersecting paths.
• **Employee and Customer Distinction**: DaoAI SkyVision can be trained to identify employee uniforms or other features, distinguishing employee activities from customer activities to ensure the purity of footfall statistics, preventing internal staff movements from distorting analysis results. The challenge is maintaining high recognition accuracy when employee uniforms are diverse or similar to ordinary customer clothing.
• **Queue Length Monitoring and Service Optimization**: In areas requiring queues, such as cash registers and fitting rooms, SkyVision can monitor queue lengths and average waiting times in real-time. If a preset threshold is exceeded, the edge box immediately sends an alert, prompting managers to deploy more staff to improve service efficiency and customer satisfaction. The challenge lies in accurately identifying the start and end of queues and distinguishing queue-jumping behavior.

Deployment Case Study

A national retail chain with hundreds of stores had long relied on manual footfall counting and area heatmap analysis in key stores to guide operational decisions. However, high labor costs and unstable data quality were persistent pain points. Especially during promotional events, when footfall surged, the accuracy of manual counting plummeted, leading to distorted evaluations of marketing effectiveness. The brand had previously attempted to introduce traditional infrared counting devices, but they proved unsatisfactory due to their inability to provide customer movement trajectories and area dwell data, and their failure to distinguish employees.

After implementing the DaoAI SkyVision platform, the brand first conducted a pilot deployment in one of its flagship stores. The SkyVision no-code platform enabled on-site technicians to integrate existing surveillance cameras and train the footfall counting model in less than a day. By setting up virtual detection zones at the store entrance, core merchandise areas, and lounge areas, SkyVision captured and analyzed customer entry/exit data, dwell times, and movement trajectories in real-time. Before deployment, the store's monthly manual cost for footfall counting was approximately ¥8000, with data updated daily and a statistical error of about 15%. After deployment, DaoAI SkyVision achieved 24/7 automated counting, with real-time data updates and statistical accuracy improved to over 98%. It completely replaced manual inspection, reducing monthly labor costs to ¥0, thereby saving ¥8000/month in labor expenses. Furthermore, through the detailed heatmaps generated by SkyVision, the store discovered that customer dwell time in a certain product area significantly exceeded expectations, yet the conversion rate was low. After analysis, adjustments were made to the lighting and merchandise display in that area, and within just one week, a 12% increase in sales was observed in that area.

DaoAI SkyVision not only reduced our footfall counting labor costs by −75% but also increased data accuracy to over 98%, providing us with unprecedented operational insights and truly realizing intelligent store management.

DaoAI Solution and Products

The DaoAI SkyVision no-code video surveillance AI platform is an ideal solution for footfall counting and heatmap analysis. Its core capabilities include: **On-site hourly training of proprietary models**, meaning retailers can quickly train and deploy customized AI models based on their store characteristics without relying on external experts, significantly shortening the go-live cycle. **Behavior/event recognition** accurately captures various customer behaviors such as movement, dwelling, and queuing, and combined with the DaoAI World semantic understanding, provides deeper business insights. **Edge box real-time alerts** ensure that abnormal events (e.g., long queues, unusual gatherings) are detected and reported to relevant personnel immediately, transitioning from passive monitoring to proactive management. Most importantly, DaoAI SkyVision supports **100% local deployment, with data never leaving the premises**, fully guaranteeing customer data privacy and compliance, eliminating retailers' concerns. The entire deployment process is flexibly integrated via SDK / API / Docker, allowing seamless connection with existing surveillance systems for easy intelligent upgrades.

Through the DaoAI SkyVision platform, retail businesses can not only **reduce footfall counting labor costs by −75%** but also gain more accurate and real-time operational data, enabling more scientific decision-making. The heatmaps and behavior analysis reports provided by the platform help retailers optimize store layouts, enhance merchandise display effectiveness, and rationally arrange staff, ultimately leading to increased sales and improved customer satisfaction. Furthermore, its powerful generalization capabilities and continuous learning mechanism ensure that the system can continuously optimize with changing business needs, laying a solid foundation for the long-term development of smart retail. DaoAI is committed to providing customers with truly implementable AI solutions, assisting traditional industries in achieving digital transformation.

This article was generated by AI. Customer cases are simulated scenarios based on real product capabilities and figures are illustrative; see product pages for official benchmarks.