At the FDIC2018 China FMCG Digital Innovation Conference, hosted by New Distribution and co-organized by the FMCG Branch of the China Electronic Commerce Association, Zhongshang Huimin received the '2018 China FMCG Supply Chain Recommended Service Provider' award. Su Xiaoxin, Executive President of Zhongshang Huimin, attended the award ceremony and delivered a keynote speech titled 'Channel Digital Operations, Industrial Data-Driven Future'. The following is a transcript of the keynote speech. At the end of last year, Zhongshang Huimin organized a small-scale summit forum for brand owners, sharing insights on data-driven marketing and refined management with some brand representatives. We also released a white paper and the Huimin Index. In our communications with brand owners last year, we focused on what I call the 'Three Transformations' in my PPT: Standardization of data production, Deepening of data insights, and Ecosystemization of data applications. We have set standards for our customers, tagging them with over 80 labels, and thus we have seen the accumulation of five years of customer data in our system. Today, I will focus on data mining, which is the second and third stages, that is, data application. In the future, we hope that data will become a marketing ecosystem. Let me first briefly introduce Huimin's data architecture. This diagram, viewed from bottom to top, shows that both front-end and back-end production systems serve as the source of data production. This data is then entered into a data warehouse, which is the middle part. We perform modeling and analysis on the data to achieve basic visualization, because people like to see charts. After analysis, the data moves to two systems for data mining and application: one called Huiyan (Huimin Eye) and the other called Huijing (Huimin View). JD.com uses 'Jing' (meaning 'wisdom'), but we use 'Hui' (meaning 'benefit'). Huiyan is like an MRI before surgery in a hospital; it can identify problem points and opportunity points in the data, and it provides these to the next system for implementation and application, called Huijing, which realizes the operational implementation and landing of the analysis results from Huiyan. How is Huiyan structured and applied? It comes from our standardization, that is, the definition of production data. Based on the tagged foundational data, we analyze and compare the static and dynamic transaction behavior data of users, thereby providing data support for Huijing's intelligent marketing, forming productivity, and enabling our business decisions to be based on visualized, evidence-based, and historically accumulated experience. This is the entire technical system of Huiyan. Huijing's data source comes from Huiyan. Huijing is like performing surgery, based on the conditions diagnosed by the MRI. It evaluates the population, system recommendations, efficiency, and effectiveness, and generates action plans and behavioral decisions within the Huijing system. It then implements operational decisions for production, products, logistics, retail, partners, and consumers. The results of the implementation are fed back to Huiyan for new data accumulation and mining before the next implementation. The most important aspect of data is how to use it to form productivity. Our BI project team spent 8 months completing the third phase of Huijing and Huiyan. Our goal is to provide certain business decisions for upstream brand owners and downstream retailers, and we hope it can truly overflow with productivity. For brand owners, the value we can provide is, first, incremental growth, and second, efficiency improvement, allowing brand owners to significantly reduce promotional costs and operational expenses. How do we do it? Still using these two systems. Huiyan is the data endpoint and the final data warehouse interface, locking onto target customers. Huijing handles everything from procurement to sales, including tiered sales management, all implemented within the Huijing system, and the results are fed back to Huiyan. Case 1: Shuanglu Battery No. 5 and No. 7 Distribution. We have conducted many deeply integrated activities. When both parties received such operational activities and discussed them, we leveraged Huiyan's capabilities. Based on the manufacturer's new products and the potential customer segments that might accept them, we first identified these customers as the initial batch of prospective sales customers for the new product. Additionally, we pushed their historical transaction data to the Huijing system, and based on the store's needs, we provided a quantitative new product brand coupon, some for one-time use, some for repeat purchases, depending on the average transaction value and category mix. Furthermore, we released the consumption cycle and behavioral results as the core basis for precise targeting in the second week. Based on the small store's restocking rhythm, we do not blindly conduct long-term promotions because small stores have turnover. We follow the store's cycle and provide orderly, regular, and intermittent support to ensure the new product's retention and effectiveness after its introduction. This is a simple two-week wave campaign result from a sales company. It didn't cost much because we had a 'scope' and an 'MRI' to screen and precisely push based on the historical battery sales data of these stores. The results exceeded my personal expectations. Based on the prediction of customer purchase intervals, we achieved even better results in the second wave than the first. When I used to work in FMCG, we distributed products by intuition. Now with big data, and with Huiyan and Huijing, we have achieved 'thousands of people, thousands of faces,' precisely targeting customers and achieving effective retention. Case 2: Coca-Cola Core Product Distribution. Coca-Cola is a giant brand, but in the Beijing market, there is still room for growth. This is about developing new sales points for a major brand. Both teams aimed for 1+1>1, so we completely excluded Coca-Cola's sales representatives' online sales points. We needed to screen Coca-Cola's sales points; if they were not within the scope of this distribution, we avoided internal friction. After screening, there were about 19,000 sales points in the city. Based on their online accumulation and offline actual surveys, we identified the shelf availability of various beverages in these stores. We then applied different operational methods based on the data accumulated in Huiyan for each store. For example, for Ice Dew, we targeted blank sales points. Based on the data intervals accumulated in Huiyan for each product, we provided effective operational measures in Huijing. These were all automatically pushed. When we wrote historical accumulation into a formula, we could basically achieve automated, scientific operations for that specific product. What were the results? Still, we spent very little money but achieved quite good results. The implementation cycle was very short, and our hit rate basically exceeded 80%. Both parties considered this a very good result, so we subsequently conducted many further refined activities. From these two cases, you can see that for our business companies, we have a research and development team of over 200 people. We do not develop for the sake of development. Our brand and B2B communication: B2B has data, but if B2B only has data without making that data helpful to everyone, I think that data actually adds to everyone's management costs and work pressure. We have always been striving to figure out how to monetize data and how to let data improve our operational efficiency and reduce brand owners' investment costs. So, I used two small cases to give you a direct view of the effects achieved in our Huiyan and Huijing systems. For downstream customers, we provide two-way empowerment, with B2B in the middle. For downstream customers, it's more about enhancing our service experience. We currently cannot achieve 'one person, a thousand faces,' but we can achieve 'thousands of people, thousands of faces' service. Why? Still because of deep data mining. When we defined FMCG, we called it a 'mom-and-pop store.' Today, we define it as a 'mom-and-pop store run by a female boss,' or 'a mom-and-pop store run by a 35-year-old female boss.' As we refine the user profile and add dynamic factors such as activity status and account status, we can provide this customer with priority delivery, exclusive discount products, priority purchase of products, etc., bringing better service to customers. For small B (small businesses), the biggest need when purchasing is definitely price, but each store has different expectations for price because small stores always want to buy at relatively lower prices. Each store has different price sensitivity, different expectations for brand completeness, and different expectations for after-sales quality assurance. We tag different customers based on their different requirements and expectations, and we can provide personalized services in the Huijing system. For example, for a small store operating 2,000 SKUs, the first need is that the search function at the top must be user-friendly. A 7-inch phone screen isn't that big, and the experience is very poor. The first thing they do upon entering is go straight to the search section and search for what they lack. Is it possible to have a comprehensive view of previously purchased related items and complete the purchase in the shortest time? Unbeknownst to the small store, we have achieved 'thousands of people, thousands of faces' service. The store owner doesn't know what we do in the background, but they feel the service is increasingly considerate. That's enough for us, because it's also a soft power. Our entire ID data project team spent 7-8 months to reach version 3.0 and achieve a data ecosystem. In the future, we hope that the entire Huiyan and Huijing, with big data BI as the foundational data platform, can achieve intelligence across Zhongshang Huimin's eight functional divisions, completely generating productivity. We hope that big data analysis can enable the separation of group procurement and sales, and production and ordering. In our Shanghai and Beijing branches, we have achieved automatic ordering, linked to the brand owner's minimum order quantity. I no longer need manual intervention for this order; the system can place the order directly. With Huiyan and Huijing below the city level, we can achieve the combination of product operations and customer operations. When we truly realize the Huimin ecosystem, we can achieve the concept of grand operations. For brand owners, everyone will enjoy more comfortable, more precise, and more cost-saving services. Embrace digital innovation and create a smart future for the industry. The core of B2B is not data; the core of B2B is to make data truly valuable for upstream brand owners and downstream retailers. Thank you. Click Read Original to see more highlights from the 2018 FDIC China FMCG Digital Innovation Conference... -END-