The 'Digital New Infrastructure 2020 (3rd) China FMCG Conference', hosted by New Distribution, was held from August 24-26, 2020 at Shanghai Fuyue Hotel. The event attracted over 3,000 representatives from distributors, manufacturers, and internet companies, with a full house and unprecedented scale. The following is the speech delivered by Huang Yu, Director of Strategic Cooperation at DuoDian, at the conference forum, now published for our readers. In a networked and information-based society, the digital reconstruction of traditional industries is an urgent task. At the same time, we have noticed that traditional retail industries face many pain points in digital transformation, such as how to digitize and standardize distributed stores, and how to reduce dependence on people. Below, I will share our five years of practice at DuoDian. To excel in online retail, you must first have a solid offline foundation. Take Hema as an example: their online traffic has already reached 50% of the total. Currently, increasing the online share remains a key performance indicator for Hema employees. When other leading enterprises undergo digital transformation, offline management is often difficult to standardize, making it hard to bring offline scenarios online and reduce reliance on people. In fact, the success of a store largely depends on its store manager. In Japan, a 7-Eleven store manager must first have a bachelor's degree; second, have worked in the convenience store industry for more than three years; and third, produce dozens or even hundreds of complex reports each month. In contrast, Chinese retail enterprises are growing at nearly 20% annually, with most retail growth coming from convenience stores, community stores, and fresh food stores. However, the annual turnover rate for store managers is as high as 100%—employees leave before they can even be trained as store managers within a year. In this context, how can digitalization be achieved? Let me introduce how DuoDian empowers these enterprises. DuoDian Dmall was founded in 2015 as a one-stop digital retail service provider, helping retail enterprises achieve comprehensive digital transformation, including online-offline integration, full-scenario coverage, full-chain connectivity, and omni-channel operations. DuoDian's retail operating system, Dmall OS, comprises 15 major systems and 800 subsystems. Through continuous algorithm optimization and iterative intelligence upgrades, it provides retail enterprises with a complete digital solution, achieving full online integration across membership, products, employees, services, marketing, and management. Online retail is essentially digitalization. If you replicate offline retail online, the underlying data must be accurate and connected in real-time. Therefore, we developed the Dmall and Mini systems to enable all supermarket operations to run in real-time on this system, reconstructing people, goods, and scenes. 1. People: This refers not only to customers but also to the hundreds of thousands of employees on the DuoDian platform and their daily tasks such as receiving goods, replenishing, organizing, cashiering, and inventory counting—all of which we have fully digitalized. 2. Goods: This includes not only product information and inventory data but also intelligent product selection, intelligent ordering, and convenience store ordering, which directly determine a convenience store's competitiveness. 3. Scenes: This includes offline hypermarket or community convenience store scenarios, as well as two new scenarios: buy-online-pick-up-in-store and home delivery. The buy-online-pick-up-in-store scenario involves placing orders online and picking up in store, concentrating customers' online purchasing habits and frequency on the online platform for precise targeting and marketing. At the same time, we push the entire supply chain—from product selection to supply chain efficiency, including source traceability—to ultimately achieve omni-channel support, full-scenario coverage, and full-chain connectivity, transforming stores into a store-warehouse integrated model to meet the needs of users within a three-kilometer radius. After five years of development, DuoDian's solution has gained recognition from many merchants. Currently, DuoDian serves 112 leading retail chain enterprises, with 13,000 stores, most of which are supermarkets and hypermarkets. The platform now has 100 million registered users, and platform information is shared among regional leading enterprises. These customers become Zhongbai customers in Wuhan, Xiaomate convenience store customers in Shenyang, Chongqing Department Store customers in Chongqing, and Wumart Metro customers in Beijing. We are doing in the traditional retail industry what Alibaba does, ranking first in the fresh FMCG sector, with GMV of 26 billion yuan last year. Digital empowerment in the retail industry, from a business perspective, should be divided into two lines: first, empowering B, and second, serving C. The concept of empowering B is essentially about improving enterprise efficiency. Previously, employees worked in a commercial scenario with unsaturation in time, making it difficult to precisely manage employee behavior. Through this approach, we can better serve the C-end through employee management and store reconstruction. For example, a convenience store can have one person handle multiple roles: replenishing goods in the morning, cashiering from 9 AM to 12 PM, picking and delivering online orders in the afternoon, and taking inventory in the evening. To effectively utilize these processes, we rely on task flows. A set of apps is centralized into the system, recording all employee actions. For instance, previously, stores operated like a communal pot, where employees focused on building relationships with store managers. Now, with the system, all store work is digitalized, and employee management is fair, just, and efficient. Serving C: We know that improving user experience is the only way to drive enterprise growth. We serve many hypermarkets and enterprises that face growth barriers online. For example, a hypermarket we served had only 100 orders a day, with no increase in online orders. There are two situations: First, high product out-of-stock rates. Second, low fulfillment rates. For instance, during a promotion, orders might surge from 100 to 1,000, but without enough delivery and picking staff, orders cannot be fulfilled, leading to poor customer experience and hindering future growth. This diagram shows how we break down store operations and digitize them. Through a distributed approach and task flows, employees can work autonomously. Previously, headquarters issued commands that were executed through layers of management. Now, employees complete tasks themselves, and if anomalies occur, their supervisors or higher-level leaders monitor them. For example, a store clerk might have 20 tasks a day. If a task is not resolved within five minutes, it escalates to the store manager; if the store manager doesn't see it, it escalates to the supervisor, and eventually to headquarters operations and even the chairman. Thus, the overall task flow is a closed loop. Another example: Previously, a convenience store chain wanted to offer a 10% discount across all stores in Shanghai. The traditional approach was to hold a meeting, communicate through layers—operations, supervisors, stores—with each department preparing in advance: stores preparing price tags, warehouses preparing goods. Even with a week of communication and preparation, the activity might not be executed well. Now, how do we do it? Directly publish a task via a mobile app, and all employees at various positions execute the required tasks on time. This allows the activity to reach the terminal directly. To excel online, stores need to be transformed into a store-warehouse integrated model. The core is to significantly increase online sales. We see the biggest pain point is that all production factors, including infrastructure, are built around the store, and store sales are still relatively high. In first-tier cities, we see that 80% to 90% of supermarket sales still come from the store. At the same time, we want to use store infrastructure to meet home delivery needs, so we need to renovate stores. With low-cost renovations, we can meet users' rigid needs. There are two core points: First, how can convenience stores cover a 1.5-kilometer business district? Second, how can we use a logistics system and store management system to systematically manage stores? Second, products have both online and offline channels. Through a product inventory middle platform, we ensure that out-of-stock rates are well managed, and we also need perfect order fulfillment to better meet user needs and provide a better experience. We know that supermarkets operate as a channel and cannot continuously drive user demand through promotions like restaurants. Therefore, it's more about providing standardized fresh food home delivery services. We understand customer behavior data and consumption habits. How do we reach these customers? The most important thing is to push what customers want: if they want water, push water; if they want food, push food. For example, how to promote a wine? There are two customers: one buys wine every three days, another every ten days. If I push the wine exactly on the third day for the first and the tenth day for the second, the customer feels like they wanted wine today. Also, at certain cycles, like sanitary pads: if a woman is about to buy them, I push the message one or two days before. Traditional enterprises would send a message to millions of members for a new wine, which is inefficient: first, they are not the target audience; second, each SMS costs 0.1 to 0.2 yuan, adding up to significant costs and annoying customers. Most importantly, this service is predictable, and only a good experience can lead to customer retention. For us, the next-month retention rate is leading in the industry. At the same time, in the digital era, user needs change rapidly, so our previous product structure is hard to meet user needs. How do we capture these needs? Now, 70% to 80% of user transactions are digitalized, giving us a means to reach users bidirectionally. During this process, user searches and feedback can be collected in real-time, which is very convenient. Through these methods, we can better iterate products, and also update products for each business district, achieving 'thousand stores, thousand faces' to meet surrounding user needs. For example, during the pandemic, we saw a surge in demand for grain and oil products, and we could quickly organize products to meet user needs. After digitalizing user needs and providing a data platform, there are more opportunities to reach users through new technologies. In the past six months, live streaming has been the hottest trend, and it has clear advantages in converting users. So we also use apps and mini-programs to better showcase new products. In product selection, how do we help stores find products? First, consider a four-quadrant model: the horizontal axis represents sales, and the vertical axis represents gross margin. Products with high sales and high gross margin are called 'hero products'; products with low sales and low gross margin are called 'replacement products'. Continuously updating 10% of replacement products is the core of product selection. If daily sales exceed 100 yuan per product, including cigarettes, there are 10 SKUs; for sales above 10 yuan, there are 150-200 basic products. Additionally, covering various small categories, a convenience store typically has 1,000-1,400 SKUs, and daily revenue exceeds 6,000 yuan. In second- and third-tier cities, daily sales of 3,000 yuan is the break-even point. What about the 10% replacement products? DuoDian has an intelligent product selection module that classifies and ranks nearly 10 million FMCG products, reducing millions of products to dozens, allowing purchasers to find suitable products for stores. This is highly efficient. DuoDian's intelligent ordering and replenishment have high accuracy, greatly saving store managers' time and allowing them to better serve customers. We believe that within 50 SKUs, human decision-making is better than the system; beyond 1,000 SKUs, the system is more accurate. The system knows when products should be cleared (down to the hour): for example, bread A sells out at 6 PM, bread B at 9 PM. For bread B, we might replenish three fewer; for bread A, three more. The system also considers weather, historical data, promotions, etc., to determine daily replenishment. Looking at our achievements in the past six months: First, Metro began deep cooperation with DuoDian in the first half of the year. Metro's fresh food standards are very high, so we emphasize fresh food and standardization online, such as -18°C cold chain delivery and product traceability. Through DuoDian's capabilities, we quickly replicated its cold chain standards and product standards online, achieving good results. Digital membership accounted for 66% of store customers, and home delivery orders increased rapidly. Its penetration in the business district improved significantly, with omni-channel accounting for 50%. Next, we will assist Metro in developing its paid membership and further enhance its product capabilities. Of course, all this is based on digitalization. Some modules are already using DuoDian's system. Now, let's look at the changes at Wumart and DuoDian. We have seen some good data in the first half of the year. During the first COVID-19 outbreak, online orders surged, but due to restrictions, many delivery personnel could not return to Beijing. We proposed the concept of 'community futures stations': treating each store and community as a fence, delivering goods in bulk to pickup stations for contactless delivery. This had two benefits: first, it truly achieved contactless delivery, addressing user concerns about contact with strangers; second, bulk delivery greatly improved efficiency. The underlying support for these points is DuoDian's digital system, which quickly transformed the store-centric model into smaller fences, using the community concept to refine store operations. During the second outbreak, we better coordinated source products to ensure vegetable supply. Not only vegetables, fruits, and meat saw significant growth, but we also ensured Beijing's mask supply was uninterrupted. Now that we understand DuoDian, let's look at why convenience stores should do online business. 1. To increase sales, especially after the pandemic. The convenience store industry is at 80% of previous business, but Wumart convenience stores achieved counter-trend growth of 20%, with online growth of 300%. After the pandemic, convenience store sales declined overall. For closed scenarios like schools and hospitals, stores closed directly. For areas with high foot traffic, like subway stations and bus stops, business was minimal due to reduced mobility. For community stores, some thrived while others struggled. Previously, stores had four doors open; now only one. Stores that opened sold well; those that didn't basically failed. The pandemic affected people's income, reducing spending at convenience stores. The pandemic boosted online development, with online growth of 300%, directly increasing sales. 2. Second is customer growth. Looking at offline, pure offline customers account for 59%, omni-channel customers (both online and offline) account for 26%, and pure online customers account for 15%. If we don't do online, we lose that 15% entirely. 3. Third is purchase frequency. Pure offline customers visit convenience stores 5-6 times a month; pure online customers order 3-4 times a month; omni-channel customers buy every 2-3 days, with high frequency. Online orders include both pickup and delivery, and pickup customers can also make additional purchases. How to excel online? 1. First, public domain traffic pools, such as Meituan and Ele.me, which charge a 20% service fee. Convenience store retail entrances charge 5% to 15%, depending on local coverage and number of stores. There are also private domain traffic sources, such as mini-programs, apps, community group buying, next-day delivery, and 15-minute convenience delivery. Why do we do private domain? Because riders on Meituan or Ele.me may increase prices. So we find ways to convert public domain traffic to private domain. First, rely on differentiated products. During the pandemic, only Wumart had masks in Beijing; other convenience stores didn't. Customers could order via our app and mini-program. Second, we can print our app and mini-program QR codes on packaging bags. Third, leverage price advantages on our mini-program to attract public domain traffic. 2. Convenience store products should be 'more, fast, good, and economical'. Hypermarkets focus on 'more' and 'economical', while convenience stores focus on 'good' and 'fast'. Ensure sufficient inventory of high-selling products. 1) Good products: The top category in convenience stores is beverages, followed by FF products like oden, fried items, and meals—unique to convenience stores. Third is daily distribution, like ice products. Additionally, we should pay attention to what young people like, such as internet-famous products like Zhong Xue Gao ice cream and 'Deadly Big Wusu' beer, and avoid stockouts. 2) Fast delivery: Meituan and Ele.me deliver within 30 minutes. Hypermarkets take 10-15 minutes to pick orders, while convenience stores take 1-2 minutes, achieving delivery within 15 minutes, better serving customers. The rest is membership improvement: divide members into four tiers—new, active, silent, and lost—and run different marketing activities for each to maintain stickiness. Also, convert omni-channel customers. Let me show you a case: the overall DuoDian convenience store situation. For store online traffic, we need to do the following:

  1. Analyze exposure and conversion rates based on different business districts and communities.
  2. Monitor output across product categories, identify high-interest categories, and list them on the online platform.
  3. For high-value, high-output products, ensure inventory levels, control out-of-stock rates, and maintain variety.
  4. Run different activities at different times based on store activities to keep products at the top and improve store ranking. In fact, community online retail is a deep science. What I've covered is only a quarter of the entire system. If you want to open a convenience store, this is what retailers, especially national leading retailers, need. If your tip is adopted, you will receive 400-2000 yuan.