No screen - no problem: data stack for in-store customer flow and behaviour analytics

In previous articles we have examined the measurement of audiences, as well as the campaign and location performance and the data stack for digital screens - in Digital-Out-Of-Home Advertising and In-Store Media. Ongoing progress in Artificial Intelligence, Video Analytics and sensor technologies enabled a rich set of metrics for all aspects of DOOH and Digital Signage. However, when it comes to Retail stores, they have an even larger demand and capacity for data and analytics, beyond screens and screen audiences, and many of those requirements can be satisfied with a similar set of technologies.

Stores themselves can be thought of as a media for their customers, and there is a need to measure their audiences (passers-by, visitors, customers), their dwell time and engagement with particular shelves and products, as well as broader product categories, demographics, and sentiment. There is also room for audience targeting via store layout optimisation and merchandising.

This means another broad spectrum of applications for the same and similar audience measurement, people counting, video analytics and sensor systems in a slightly different area and for similar purposes.

Let us have a closer look at the needs of in-store customer flow and behaviour analytics, what data can be collected to address them and how it can help in creating smarter physical Retail spaces and make Retail experience more comfortable and rewarding.

People counting

One of the first application of sensors in Retail was historically around footfall measurement, as store owners wanted to analyse their most precious resource - customers - and understand visitation patterns to be able to optimise their operations. Once achieved with a simple infrared sensor, to a certain degree of accuracy, now it is often done with more sophisticated sensors or camera-based solutions - for a more robust and trusted data, often coming with some extra dimensions like unique visitor count, conversion measurement, traffic flow mapping etc.

Recently, people counting solutions started to also offer occupancy monitoring capabilities. While undoubtedly important in the COVID-19 age for safety and compliance purposes, real-time occupancy data brings lasting benefits allowing for better workload management, traffic spreading and, ultimately, improved customer experience.

Some people counting solutions come with a unique and returning visitors counting capability (for example, based on Wi-Fi mobile phone sensors). This data allows to better understand visitation patterns and analyse the effect of promotions on new and regular customers. Moreover, this provides an opportunity to analyse average customer dwell time as an aspect of the in-store customer behaviour.

Traffic flow mapping

Another aspect of customer behaviour analysis can be enabled with overhead sensors and Video Analytics. Top-down view on the whole store's floor or some of the most critical areas allows to see where customers go once they enter and how the traffic splits at every turn or intersection. This also allows to highlight areas with higher and lower traffic and build a traffic heatmap. Another useful dimension of such visualisation is the average dwell time at every point of the store floor. Dwell heatmap allows to visualise areas where customers spend more or less time and use these insights for merchandising and better layout management purposes.

Knowing how traffic flows and dwell time change over time and across the floor, retailers can in a more informed way plan promotions and expose certain products or offers to more traffic, or inversely, use promotions or product sections spatial positioning to draw more customers to certain areas of the store.

Queue monitoring

Being able to track people's location inside the store provides an opportunity to improve customer service and make shopping a more frictionless experience. With specialised camera-based analytics it is possible to monitor occurrences of congestion or the length of the queue and detect walk-offs, i.e. when customers lose patience and leave the queue before reaching the counter. Knowing the patterns around queue formation, as well as having a capability of early queue detection can help retain customers, drive loyalty and sales performance.

Analysing customer behaviour at points of interest

Oftentimes brands use their presence in-store to showcase new products or run marketing campaigns. This usually comes in a shape of an endcap or pop-up display, a special section layout or simply a static banner or a digital screen. Some brands have a permanent dedicated section inside the store, which they fully control and manage. Putting a lot of effort in such initiatives, brands value every bit of the first-party data that is able to show customer behaviour around and sentiment towards their offers or space. This data can also be used to gauge the impact of advertising through any media channel on the in-store customer activity and engagement with advertised products.

Much of the audience measurement technology, like Anonymous Video Analytics, can be used to capture data about customers' conversion, dwell, attention time, mood and demographic segmentation in front of such product displays or promotional materials.

Product pick-up monitoring

Knowing how often people pick up items from the shelf, when they put them back, and if they don't, provides another important perspective on customers' behaviour and their response to promotions or advertising campaigns, and this can also be detected and analysed with an AI-powered camera, product pickup sensor or an integrated smart shelf solution. In many cases such solutions also enable shelf stock and availability monitoring and replenishment, a vital part of the next-generation smart retail spaces.

Ordering kiosks and recommendation systems

Many retailers have kiosks or ordering systems inside their stores. This is especially prevalent at the Quick Service Restaurant business. Looking back at the customers making orders through such kiosks, with a camera sensor, can provide rich insights into their audience profile and preferences, and may help to improve the ordering process by providing relevant recommendations on the spot and optimising the whole ordering process over time.

Vehicle analytics

Drivethrough, Curbside Pickup and Petro-Convenience Retail environments benefit from real-time and historical analysis of the customer vehicle traffic. Detecting the presence of the vehicle, its colour, type, make and model may help to decrease wait times and make the ordering process faster and more efficient. In some environments, where License Plate Recognition is possible (and consented by the shoppers), for example, as a part of loyalty programs, it may help to quickly link every arriving customer to their order and make the whole experience as frictionless as practicable.

Sentiment measurement

Even simple mood and emotions detection via camera systems at the entrance and exit, or voice tone analysis solutions at the counter may help to track changes in customer sentiment during their visit or overtime and correlate that data with other metrics, which could help improve the experience and champion the customers' loyalty.

Real-time data such as described above can be very powerful if used properly. Even some isolated insights can inform business decisions that would in turn generate extra savings or revenue. Imagine if you could automatically change music background in a retail store dynamically, based on the aggregated audience profiles of the current clientele, their preferences and weather. Or if you could run interactive promotions that would take into account every shopper in the aisle. Or if you could put a new product on the shelf that is sure to generate maximum interest and sales. Every bit of data about what is happening inside the store can be precious for CPG brands trying to reach customers in the most efficient way. Imagine now if you could combine all essential metrics and sensors in a single platform. This is quickly becoming a reality, with computing resources and AI capabilities evolving so quickly.

Data about the in-store customer flows and behaviour, coupled with real-time audience measurement is essential in the process of creating the smarter retail spaces that are able to transform physical stores in the age of the accelerating ascent of online Retail. It has become an important part of Retail Analytics and will only grow in capabilities and relevance to the industry.

It is the right time to start building your smart retail space today.