Emerging from Lockdown: Security Solutions for the Retail Sector

With the majority of the world’s population compelled to minimise the time they spend outside their homes, and in the presence of those outside of their households (for the time being), many aspects of our lives have been turned upside down. Shopping is no exception. 

Non-essential stores - for example, those selling clothing or homeware - have been forced to shut their doors, leaving eCommerce as consumers’ only option to purchase items in these categories. As for those supplying indispensable products - such as groceries - supermarkets have remained open, but tasked with restricting the number of shoppers allowed in at the same time and encouraging them to keep a safe distance. This has caused long queues near the doors and cash registers.

On the whole, the retail sector is experiencing one of its most testing periods to date, with record falls in sales and unprecedented obstacles to overcome.

As we now move towards a new phase of the lockdown, with movement restrictions eased and non-essential shops soon able to once again welcome customers, retailers will finally start seeing a more normal procurement landscape. Whatever this phase might look like for vendors, we know that, while it will help solve some of the challenges they are currently facing - ie blocked footfall - it is set to bring about new ones as well. For instance, as grocery store traffic remains steady and other outlets re-open their doors, enforcing social distancing within these premises or outside store perimeters will be vital. Not only to continue flattening the contagion curve, but also to protect the well-being of store staff, shoppers and the wider population.  

Retail security challenge

Social distancing can be challenging when it comes to compliance, and particularly so in those stores such as supermarkets where, typically, there is a high footfall. The first step to enable it is controlling the number of people permitted in the premises by counting those who enter and those who leave.

Monitoring in-store occupancy is something many retailers appreciate the importance of and are already taking care of. We’ve all experienced incredibly long waits due to the one in, one out policy observed by our neighbourhood supermarkets, and we’ve noticed the signs on the floor and the walls reminding us to keep two metres distant from other customers. 

However, despite all the design and infrastructure that retailers can implement, whether customers will respect the rules is hard to tell for sure. As a consequence, complying with current guidance is no easy feat. Retailers must take matters into their own hands to keep on top of how many individuals are behaving in their stores, what they are up to and whether they are keeping safe.

At the moment, social distancing is largely enforced by staff keeping a watchful eye on the shop floor. However, this is hardly sustainable and is unlikely to be the most convenient and effective method. Some other physical controls that retailers have put in place are expensive to operate and can create conflict between staff and customers. Automation-enabled technology located at the entrances of the store could, for example, be a solution.

Cameras for counting

There are certainly a number of systems available that use counting cameras to establish the occupancy levels by using analytics to count people in and out. In turn, these can be connected to a simple two-light traffic unit, signalling green to allow ingress and red to request people to remain in the queue. This solution could also feature a visual display unit that generates a Stop and Go sign, as well as marketing and advertising messages, thereby enhancing the optimisation of such end user investment.

While these units work very well, they can bring about other issues, such as making General Data Protection Regulation (GDPR) compliance more difficult and increasing cost. Additionally, these products would require more cameras being installed and would, therefore, have the potential disrupt regular in-store activities.

To tackle some of these challenges, using standard video analytics with correct camera positioning is essential. Generally, cameras need to be located above the entrance door. Even when this is ensured, inaccurate counts are commonplace. Needless to say such errors can cause customers to queue for unnecessarily long intervals, potentially putting them off from shopping altogether or, more alarmingly, lead to an exceeding of the pre-set occupancy level, thereby putting store staff and customers at higher risk of contracting the virus.

Either way, a more effective and tailored approach is required.

New type of analytics

A potential solution to some of these issues is to make use of what retailers already have: surveillance cameras. However, instead of using individual cameras loaded with capabilities such as facial recognition, it’s about using analytics which can provide data-fuelled insights which respect shoppers’ privacy. No additional cameras, but more information which adds value and can help compliance with current regulations such as the requirement for social distancing. 

One such method is deploying video analytics capabilities that can track gait analysis. Gait is unique to everyone. Its analysis is the study of human movement and its possible application in retail - and, specifically, occupancy monitoring - is fascinating. 

Pose estimation, or skeletal analytics, uses points of recognition within the human skeleton to identify and count each person entering, exiting and moving within the store. It’s therefore not reliant upon a traditional count line and, as a consequence, doesn’t require an overhead camera. Quite often, retail establishments have a camera installed which faces the entrance. This device is optimally positioned to enable pose estimation analytics to create an accurate count of every person entering and leaving the premises. This technology can be employed on pre-installed cameras covering every entrance and exit. The data collected from exits and entrances is then aggregated to provide a total occupancy figure, in turn enabling the control of ingress and egress of customers. 

Pose estimation recognises the individual nature of each gait and skeleton and, therefore, assigns a unique data-driven ID to each human object. This ensures the count is extremely accurate, minimising queue time and empowering retailers to adhere to store occupancy guidelines

Finally, pose estimation can gather and provide entirely anonymous data which isn’t stored in the system. Therefore, this kind of surveillance protects customers’ privacy and prevents any GDPR infringements. 

Quick and cost-efficient

One of the key advantages of pose estimation analytics is that it can be installed on existing surveillance cameras and the set-up procedure involved is minimal. For these reasons, the cost of installing this technology is less than half that of a conventional people-counting system. Furthermore, there’s no disruption to the in-store experience, and there’s no need for workers to be present outside of store opening hours in order to install new cameras at the entrances.

The implementation entails the installation of a simple Black Box, which is then connected to the cameras already used at the premises and set up with the ancillary devices chosen by the retailer. For instance, a traffic light system or VDUs. The entire operation should take less than three hours.

Pose estimation (or gait analysis) is a fast-growing medium designed to enhance and accompany video analytics, and is set to enable exciting advancements over the coming months and years. An interesting retail application for this technology could be, for instance, counting individuals in queues at checkouts, not only to help comply with social distancing regulations, but also to optimise the customer journey.

It can also help to create heat maps and monitor slip, trip and fall incidents, thereby supporting security staff. Retailers who are considering the purchase of a traditional footfall-counting solution should evaluate whether it would be the most beneficial option, and whether or not it could unlock long-term advantages for the business. The risk is that such a system would simply represent an additional expense: a quick fix aimed at dealing with what we all hope is a short-term problem, but might in fact impact the way in which retail stores are laid out for the longer term.

In order to future-proof their businesses, and in light of the global pandemic as well as other existing challenges within their sector, retail distributors have the opportunity to deploy an incredibly useful and appropriate tool. What players in this industry need now is an economically and operationally viable solution to address these issues: Pose estimation analytics is one such solution.

What lies ahead

Retailers have been the backbone of people’s well-being during the first few phases of the lockdown, and it must be said that navigating this challenging time successfully would not have been possible without their hard work. 

The future has never been so shrouded in confusion, not just for businesses in the retail sector, but for companies and individuals alike. We’re not sure what lies ahead, but it’s clear that what we are witnessing isn’t a temporary change. We are looking at a long-term evolution that will allow physical retail to exist in this new world. Gait analysis has the potential to be an invaluable tool in the retailers’ arsenal to ensure customers and staff are safe and, ultimately, weather the storm.

Kevin Waterhouse is Managing Director of VCA Technology