With the advent of the digital era, the application of new technologies by FMCG companies has become increasingly important. For enterprises, the point of sale is the closest place to consumers. Collecting data from terminals and efficiently conveying it to management can help quickly analyze market changes and make decisions. Mike Wang, Senior Vice President of Trax China, delivered a speech titled "FMOT High-Frequency Data Drives Changes in Distribution Management" at the 2023 (8th) China FMCG Innovation Conference, elaborating on the impact of mastering terminal data on distribution management.
Two Sets of Numbers Illustrate the Importance of Mastering Terminals
The traditional consumer goods market is largely offline, with at least 80% of business controlled in various types of offline stores. Therefore, mastering data from each type of offline store determines success or failure at critical moments. Today, my topic is the First Moment of Truth (FMOT). What is the evolution of terminal retail distribution management? Before that, let me share two sets of numbers with you.
The first set concerns consumers. On average, consumers spend about 15 minutes in a store, and about 15 seconds in a product area. 40% of consumers who cannot see or find the product they want to buy will switch to another product. On average, a consumer stands in front of the shelf for about 6 seconds to make a decision. The placement of FMCG products in terminal displays is crucial for purchase conversion rates. Whether it's O2O, community group buying, or online e-commerce, whether in public or private domains, with saturated data as a foundation, targeted intelligent push and operational efficiency will be high. But if in the vast number of offline stores, products are not placed where consumers can easily find them—for example, a consumer wants to buy a bottle of Head & Shoulders shampoo at a store, but it's placed in a hard-to-find corner—99.99% of consumers will not go to a store three kilometers away to find that product; they will simply buy a similar product on the shelf. Therefore, for brands, poor product placement can cause significant sales losses.
The second set concerns retailers. We have surveyed retailers in different countries around the world. Major retailers spend an average of approximately $340 million annually on labor costs related to product shelf availability. They employ many people to continuously patrol stores to ensure shelves are not out of stock. Even so, on average, 8% of products are still out of stock, causing an average annual sales loss of 4.1% for each retailer. So why mention the First Moment of Truth? When we can master granular terminal information, we can take timely terminal management actions. When we don't master this information, products may not be displayed on designated shelves, resulting in poor sales data. When we master each terminal, we can ensure all products are displayed in the best positions, and sales will be maximized. There is a saying: "You can only get what you can control." This control requires relatively saturated and sufficient data, similar to e-commerce and O2O, to master overall terminal performance.
Changes in Terminal Distribution Management
Many large enterprises face a pain point: they have thousands of salespeople nationwide, different city managers, and even many promoters, and they separately hire many inspection teams to investigate terminal performance. Even so, data feedback efficiency is still not high enough, hindering business growth. Suppose you are a senior executive. Do you rely on real-time frontline information from terminals to make decisions? If the terminal is a salesperson or a city manager, how can they report competitors' market actions to leadership at the first moment to gain company support? In this process, opaque channel information leads to lower offline management efficiency and response efficiency compared to online. In this era, more technology companies provide advanced technologies to help improve information blockage between channels.
We can divide the channel context of retail distribution management into past, present, and future. In the past and present, many enterprises have sales representatives, distributors, or other hired personnel in stores, using paper checklists to record product distribution status. However, SFA (Sales Force Automation) and SaaS-based terminals can automate information acquisition and real-time recognition. In the future, all underlying data will be connected without manual acquisition. If data is manually acquired, the granularity of data management is limited to the frequency of human store visits. Therefore, the increasing application of robots and IoT technology devices will help break through this blockage. The future market will definitely be a retail digital twin market, which will transform physical shelves into digital twin digital shelves. All information on the shelves will become data, provided to all upstream and downstream channel partners. So, we are already transitioning from the current human+AI approach to managing terminal distribution, gradually shifting to a machine+AI approach, so that the First Moment of Truth can be more thoroughly realized.
In fact, when operating an enterprise, whether you are a marketing director or a sales director, the most important point you want to understand is: after products are distributed from the warehouse, what is happening at the shelf end before the consumer checks out? Why do customers buy this product? What driving factors cause customers to buy more products? In this process, it is necessary to use "eyes in the store" to view consumer actions in real-time, including the impact of location on consumers, which can be achieved through technology. Initially, all enterprises recorded manually. Now, it has basically advanced to a human+AI approach, where taking photos or videos and recognizing terminal data in seconds can identify product placement, price tag appearance, price management, and relationships with competitors, all fed back through technology. However, the data granularity of this method is limited to the time and frequency of human store visits. With the underlying data of the First Moment of Truth, at least the overall offline sales rate can be improved.
"Machine + AI" Enhances Sales Power
Decompose sales power into three parts: penetration, execution, and repair. What is penetration rate? Suppose a brand has 50 SKUs, and all 50 SKUs are distributed to stores for sale. During promotions, the thoroughness and efficiency of execution are optimal. Because the store itself is dynamic, after products are sold or replenished, there may be deviations from the initial planned display standards. By maximizing penetration, optimizing execution, and making repair most efficient, problems can be discovered and fixed in time, and terminal sales efficiency can definitely be exponentially improved. This is the greatest role of the First Moment of Truth and the driving force at the bottom.
Store displays include basic display distribution, shelf share, and display position. Now, no manual identification is needed; AI technology can digitize these. Many pieces of information in stores, such as floor displays, POSM materials, and even cut cases and open windows commonly used by beverage and beer manufacturers, have all been conquered by AI technology in this era. But is that enough? The development of technology is moving from human+AI to machine+AI. We need an application that can see changes in the store 24/7. This trend has already occurred.
Machine+AI brings a more technological dimension. First, data dimensions can control overall terminal performance by day or even hour. For example, if drinks sell out at lunchtime and no one replenishes them, I need to replenish them in time; if I find that a competitor's product is more attractive than mine a day earlier, I need to handle my product quickly. When you grasp this information at the first moment, you can obtain the underlying data basis for high-frequency decisions. At the same time, this data no longer relies on people, because many data reported by people are based on human credibility, which may have issues of omission or authenticity. The so-called data shelf stitching and restoration of the overall store information will be presented through a retail digital twin, allowing everyone to feel as if they are there. For example, if you are in Chengdu and want to see a store in Hailar or Xinjiang, there will be a VR-like environment where every product's data information can be presented instantly before your eyes.
How to achieve this? Through automatic inspection robots and other equipment, these are the most direct devices for obtaining high-frequency, high-granularity data in the future. With these devices, the data of the most scarce "goods" and "field" in the "people-goods-field" framework can be fully realized. When robots walk through all stores, the distribution, display, facing, out-of-stock, shelf share, and even prices of every shelf in the store will be fed back on an hourly basis. In this way, the decision-making dimension gains a very high-frequency data foundation, and everyone's decision-making efficiency will be different.
From Data to Insights, from Insights to Action
Many products on shelves are bought out by consumers and not replenished in time; many products are often not purchased, and whether they should be replaced with more popular products; products that often sell out need to improve front-end replenishment and stocking efficiency. All these require data or finer-grained data as a basis. When you have this data, management and direct sales layers will have a more transparent, data-linked decision-making architecture. Similarly, when we have transparent data, we can directly send people to stores to solve problems. For enterprises, this part of the labor can be reduced through fine-grained applications. This is the future trend.
Based on fine-grained, high-frequency real-time data, new management, assessment, and decision-making bases are formed. Of course, these changes in decision-making and management bases require changes in the overall enterprise architecture and organizational structure to be realized.
The following video shows AIOT devices in a supermarket. When the robot walks through the entire store, all shelves it passes are digitized. The display position of each product, product type, competitor performance, and price tags all become real-time understandable numbers, presented in a new system. These devices can identify more granular data, including daily, weekly, monthly, and quarterly data, as well as hourly data. The most important concept is that when you first master this data, you have entered a higher dimension of daily and hourly management, while others are still managing monthly. Your decision-making efficiency will be higher than others; this is a dimensionality reduction attack. Therefore, in the digitalization process, whoever first applies AIOT technology and grasps market trends is likely to occupy a more important first-mover advantage in a highly competitive market. In the future, data will be the most important asset.
Applying digital technology can turn physical shelves into usable data. Based on this data, business insights are realized, driving action changes at every level from management to terminals. In this way, the overall organizational structure will transform and upgrade towards the most efficient decision-making. Finally, this data will simultaneously act on brand owners, retailers, and consumers. When all data is used to adjust display combinations in the most efficient way, consumers will also find the products they need as quickly as possible in central stores, and brand sales efficiency will be guided to new changes based on fine-grained data. The above is a case study of FMOT as a refined distribution transformation technology application driven by fine-grained data.