The new DéporTienda, the same as Deportivo de la Coruña, is not a simple store merchandising with a facelift. It is a case study of how spatial analytics, computer vision and predictive models can impact the retail sports. And a club from, for now, the Second Division has done it. RC Deportivo celebrates its 120th anniversary immersed in a modernization operation that goes beyond what happens on the pitch: museum, hospitalitysports city… But there is a project that says more about the direction the club wants to take than any signing or renovation of the stands. It’s your store. La DéporTienda has just reopened in its historic location next to the Abanca-Riazor stadium after two months of works. And more than just the furniture or decoration has changed. The nervous system of space has changed: it is now monitored, analyzed and, according to those responsible, it is capable of anticipating decisions before they are made. The cameras watch The project started with an observation phase of at least four months in the old store. The company responsible for the technological layer is Noumena Group, a Barcelona company founded in 2011 and specialized in spatial analytics with computer vision and Machine Learning. Its CEO, Aldo Sollazzoexplains that the system they have implemented in Riazor has the same DNA as the one they developed for Barcelona City Council in the analysis of Superapples and the green axles: cameras that process images to understand how people move through a space. Material for sale, and in the background, the wave made of PETG. Image provided. “It works through cameras connected to a brain based on edge computing“explains Sollazzo. “Each camera converts the input visual in a string text that guarantees anonymity and complies with the GDPR and the European AI law. We do not store images: we store spatial data,” he explains to Xataka by phone call. The system maps movement flows within the store, segments visitors by estimated age ranges and gender, generates heat maps and cross-references all this with product distribution and sales data. So far, descriptive analytics. What Noumena proposes as differential is the predictive layer. Predict before moving a shelf “The difference with other systems is that our data are not numbers disconnected from space,” says Sollazzo. “They are linked to the built environment. And that allows us to train models to predict what would happen if we redistribute the product, move the furniture or relocate the payment points.” In practice, this translates into decisions such as estimating whether an alternative display arrangement maximizes product exposure. Or whether relocating cash registers based on traffic reduces queues on busy days. Or calculate how many staff the store will need for a specific event. All this without the need for trial and error in the field. Heat maps on the store floor (top-down view). They represent the distribution of visitor flows in three different scenarios or moments, probably different furniture configurations or different temporal moments (normal day, match day, special event…). The most intense areas indicate a greater concentration of people, the dark areas are cold areas where few people pass through. It is clearly seen how the traffic distribution changes depending on the scenario: in some configurations there are obvious bottlenecks at the entrance and next to certain exhibitors, while in others the flow is distributed more homogeneously. The numbered white elements are the furniture and display modules. Image provided. Three-dimensional visualization of the same flow data as in the previous image. Instead of a flat heat map, occupancy data is represented as volumetric columns that “grow” from the floor: the higher the column height, the higher the traffic density at that point. The isometric perspective allows you to see the store as a built space (you can sense the walls, the modules) with superimposed spatial data. It is the representation that connects analytics with architecture, which is the point that Sollazzo emphasizes as differential: the data is not abstract, it is anchored to the physical environment. Image provided. Panel with a comparison of key indicators: net sales, visitors, items sold, conversion rate, average ticket… All within the framework of an event day, with a specific anniversary that drives attendance and purchase. Image provided. This tab crosses spatial data with sales data by product. The graphs show the ranking of suppliers, where Kappa logically dominates as it is the technical brand of the club; the interactions by exhibitor or the comparison of interest versus sales by supplier, where it is seen that Kappa generates a lot of interest and many sales, while Soricastel and Texprint (merchandising, not technical clothing) generate interest but convert less. This type of data is what allows us to decide if a supplier needs a better location, better price or less space. Image provided. Thanks to this research, the club has increased its working square meters by 10% and enhanced product exposure by 15%. For a football club store next to a stadium, where on match days the footfall can multiply by twenty (a figure declared by the club), those margins matter a lot. The system, furthermore, cross internal data with external variables: weather forecast, events in the city or at the stadium, mobility data in the surrounding area… “In celebration of the 25th anniversary of the League We had 250% more occupancy compared to a normal day,” recalls Sollazzo. “Without this technology it would be impossible to anticipate how much stocks “Do you need, how much workforce, how to redistribute the points of sale so that the queue does not block the flow of the store.” What is measured and what must be demonstrated The KPIs that the club will monitor are those expected in any serious operation of retail: average ticket, conversion rate, average stay time, number of visits, distribution of flows by module and visitor profile. The return on investment, according to Sollazzo, is expected in one year. The idea, in theory, is to sell more, of course, but also to make better, … Read more