Measuring is Knowing, Guessing is Missing: The Future of Footfall

Savvy retailers know that maximizing profits means smart marketing; however, it can be challenging to market efficiently in an offline world, at the point of sale, where measuring and analyzing customer behavior is next to impossible.

To support that effort, footfall – people counting – technologies have been created and refined over the years. Suppliers of footfall offerings use sensors, WIFI or IMEI data. In an effort to further improve their solutions, some also work with expensive 3d cameras or offer a combination of these technologies.

What are the promises of footfall? Counting people in shops, malls or train stations can help determine key metrics which are crucial for survival in today´s retail climate. The results can be used for increasing sales and driving profits up. For example: retailers can measure peak hours and match their staffing patterns, they can measure basic sales conversions (how many people in store compared to how many purchases) and assess how many customers return to their stores. On a broader perspective, managers can find out what marketing initiatives delivered more traffic to stores and which stores perform better than others.

Sure, collecting and using these data is better than not being aware of basic trends and figures. However, compared to the possibilities of online tracking and customer analysis, footfall solutions score not very highly. It is a fact: Online shoppers can be traced and analysed more effectively. Every click indicates attention, every point of friction, every churn event can be used for understanding customer needs and refining the customer journey.

New technologies give better insight into POS customer behaviour

A.I. technologies are on the rise and combined with video analytics capabilities they can make a big difference in understanding POS customer behaviour. The video-based A.I. technology DeepEyes is perfectly suited for footfall requirements as it operates on inexpensive standard equipment, which makes it the perfect solution for company-wide rollouts. Also, it can work standalone. Internet access is not needed as the innovative algorithm creates its own neural network on the processor.

Hence, customer data do not have to leave the company and privacy concerns would not necessarily arise. Also, the technology works very accurately and can be customized to any information requirement that can be “seen” with a camera. The algorithm will only extract the required data, e.g. motions or gestures. This means that, except, of course, for authentication purposes, faces do not need to be the target of recognition. Instead, motions, which are a biometrical trait like the iris or the face, can be used as unique identifiers, which makes it possible to count how many times a specific customer has entered a location.

Knowing your customers based on measurable “soft” insights

Besides the fact that the DeepEyes technology is an effective people counter, which can also “learn” who is staff and who is customer, and can give accurate “net” customer figures, it is also a good choice for the more sophisticated tasks: Depending on customer requirements it can easily track and analyse customer attention by catching the line of sight as well emotions like surprise, lack of interest or attention. Also, it can capture approximate age as well as gender and more specific traits.  Hence, the adaptive algorithm is also capable of analysing the commitment of sales representatives during customer conversations or the effectiveness of billboards or a specific promotion. More sophisticated use cases are assessing locations for products that lead to higher revenue or analysing customer needs in malls or shops and optimise their shopping experience. Also, the DeepEyes A.I. based video analytics technology can be an effective answer to shop lifting. It can identify fraudulent intentions by specific motions and gestures. Of course, the analytics algorithm is also capable of identifying core customers and alerting a sales representative on his mobile phone. A sales person who greets a major customer personally can help to increase customer loyalty.

Privacy concerns belong to the past

The DeepEyes technology can easily accommodate privacy concerns. Depending on specific recognition criteria the technology can also work with mere parametric information of faces and thus recognize numbers of visitors simply by comparing these specific data. This innovative recognition method which only stores selected pixels of faces, cannot be reproduced, even if the footfall database is being hacked or stolen. Additionally, the simple fact that the DeepEyes technology can work stand alone, represents an extra level of confidentiality and privacy.

The DeepEyes´ unique method of storing has a nice side effect: It enables fast search. As an example, the technology can identify a person in a data base of 10 million faces within a second with a standard I5 computer.

Besides, the algorithm can be adapted to fit specific national privacy requirements. As pointed out above, the motions of a person also represent a unique identifying biometrical trait. They can be used for “counting” and “re-identifying” a human being without breaching privacy barriers.

The future of footfall

Amazon has set the pace: With its first Amazon Go store in Seattle it has revolutionized the way we buy. The company has opened a supermarket with no checkout operators or self-service tills. Instead, it uses hundreds of ceiling-mounted cameras and electronic sensors to identify each customer and track the items they select. When they enter the store, shoppers walk through gates similar to those in the London underground: They simply swipe their smartphones loaded with the Amazon Go app. Purchases are billed to customers’ credit cards when they leave the store.

While this step may be difficult to take for retailers in Europe´s strictly regulated and privacy aware markets, video-based artificial intelligence solutions like DeepEyes can make a big difference and provide valuable, reliable data for analysing and acting on customer information at the point of sale. Managers can easily understand customer trends, they can evaluate the impact of store promotions and measure the effectiveness of their sales agents and their marketing spend. All in all, based on these data, they will plan with confidence for the future.