Options and Best Practices for the Use of Video Analytics in the Retail Sector

Video analytics refers to the collection and analysis of images to extrapolate clear and actionable outcomes. Indeed, with video analytics, operators can maximise their video management deployment to achieve more than surveillance. Through automatic detection and data interpretation, security teams can speed up maintenance and react faster to critical events.

However, with the rise of unmanned stores, the ECR Retail Loss Group estimates self-checkout processes account for 23% of total unknown store losses. This indicates self-checkouts as the focus for future video analytics deployments, with most retailers surveyed planning trials in product identification and non-payment alerts. However, as the need for all areas of retail to collaborate across organisational boundaries becomes more important to increase profits and combat the progress made by competition from online retailers, in-store security systems in retail applications must expand beyond security to include other business functions and departments within the retail industry.

Unlocking Consumer Insights, Improving Productivity, and Boosting sales
To fully realise the potential of video analytics, cross-functional use will be critical to success. Security systems are constantly capturing valuable information that can be used to transform businesses. Advanced retail analytics solutions can interpret such data to unlock consumer insights, improve staff performance, enhance customer experience, and boost sales conversion rates. Such functions can be added as part of a unified security solution and can be scaled to a single store or multiple locations globally.

By tapping into video gathered via a physical security system, retail analytic applications use modules like visitor counting, conversion rates, queue management, heat maps, and directional analysis to better understand customer behaviour and make real-time informed decisions that increase both consumer engagement and in-store profitability.

Using conversions as an example, studies have shown that 90 percent of consumers are more likely to convert from just browsing to a sale when they are helped by a knowledgeable associate. An advanced retail analytics solution can be configured to notify management which departments are crowded so that they can deploy enough associates to help with any inquiries. This can both accelerate conversions and reduce the cost of labour through better employee management. Losing sales due to long lines at the checkout counter? Retail analytics solutions can help reduce customer abandonment caused by long queues by notifying management to increase tills when long lines are detected. These solutions help managers act at a moment’s notice and turn undesirable situations into desirable and more profitable outcomes.

Display and promotion effectiveness – and other longer-term issues – can be improved with a feature like heatmapping. By visualising the areas in a store that represent hot areas – where shoppers dwell – compared to cold areas – which receive very little foot traffic – managers can measure the effectiveness of certain displays and promotions and make changes to address insufficiencies.

Best Practices for the Successful Deployment of Video Analytics
Despite video analytics’ potential as a powerful tool in loss prevention and generating insight, its deployment can often come with its challenges. Here are a few crucial elements organisations can consider when evaluating and deploying video analytics to increase success in a retail environment.

Organisations must first start by identifying the problem they’re looking to solve. Then, set the right performance expectation, and define metrics for success. Video analytics offer the best insights when deployed as a solution to a problem rather than a solution in search of a problem – the latter is often challenging to evaluate.

Having defined the problem, finding the appropriate video analytics for the job will help allocate resources appropriately and limit overestimating results. Companies need to consider each analytics’ intended environment of operation and judge how well it matches their scenario. Using video analytics outside of their intended parameters makes performance unpredictable, often to the detriment of business goals.

Another consideration is the server-based or edge-based nature of the video analytic processing and its effect on network bandwidth. Both video analytic processing models have their limitations and advantages that relate to a system’s architecture.

• Edge-based analytics reduce bandwidth usage by transmitting only the results of video analytics processed by the camera. However, this requires specialised cameras equipped with the hardware needed to run video analytics.
• Server-based analytics process video streamed to a server which increases bandwidth usage. However, this allows operators to take advantage of the existing video surveillance infrastructure.

The right security system will include event-to-action, alarm management, and map-based monitoring to leverage video analytic data. It becomes even more powerful when paired with a decision management system that can trigger workflows when suspicious activity is detected, or correlate multiple events into a single incident trigger.

After selecting and deploying a video analytic solution, it is crucial to define metrics to measure performance continuously. Video analytics is not a “set-and-forget” type of technology. High accuracy has traditionally been hard to obtain, especially in open areas with many moving parts and people. Often, factors in performance issues are video quality as well as lighting and positioning. Monitoring performance and working with a trusted vendor will facilitate the fine-tuning adjustments that make the difference between a fair purchase and a valuable return on investment.

A clearly defined set of success conditions before deployment also makes it easier to establish the return-on-investment A good example is the people counting video analytics used to prove compliance to occupancy regulations. This use case’s ROI compares the cost of the video analytic solution against hiring staff to count customers and the cost of any violation of occupancy regulations.

Another way to increase ROI is to look for more ways to use video analytic data. Deploying security in a unified platform allows for the combination of POS data with video recordings and video analytics to pinpoint suspicious transactions. For example, motion detection video analytics can reduce investigation time for fraudulent returns where no customer was present by narrowing down the search to only transactions where a return was completed, but no motion was detected in the area customers would typically occupy.

Thinking beyond the immediate results of video analytics and combining them with other data sources on a unified platform empowers operators with a complete view of events in each retail location.

In Conclusion
Retail analytics software applications can make the most of existing security resources and use the invaluable data they already gather to benefit a variety of departments in a retail establishment – retail marketing, retail operations, and merchandising. One of the most important benefits offered by retail analytics solutions that are part of a unified security solution are the additional information they can provide physical stores to compete with online retailers.

Online retailers have long had the distinct advantage to easily collect vast amounts of data and analytics from every customer visit. Today, traditional bricks and mortar retailers are wising up and are beginning to bridge the data gap, by fully leveraging the data that they are already collecting to improve sales, operations and customer experience.