The Wonderful Journey of a Commodity: Data-Driven Smart Retail

The retail industry may be an ancient industry with a large scale and complex roles that has existed for a long time since human beings have traded commodities. Since ancient times, the development of the retail industry has always been in the process of continuous innovation. After entering the digital age, the retail industry has completely overturned the constraints of time and space and entered a new stage of development.

In this process, how has digital technology affected the business model of the retail industry? What is the current state of intelligence in each sub-sector? What are the key issues that need to be addressed in the digital transformation of the retail industry? How to better use data to drive business development?

On April 10th, with the theme of "Digitalization Facilitates the Fast Circulation of Commodities", TVP celebrities took the participants into the retail industry where they had never been together, and explored new scenarios, new opportunities, new future.

1. Technological achievements Mengniu’s digital transformation

Zhang Jue, CDO of Mengniu Group and Tencent Cloud TVP industry ambassador, gave a keynote speech titled "Technology Achieves Mengniu's Digital Transformation". Starting from the nature of business and business methods, he shared Mengniu's strategic thinking and practical experience of digital transformation.

(1) Change or not change: the nature of business and the way of business

Before starting to share, Mr. Zhang Jue threw a classic question: When we talk about digital transformation, does it mean that the essence of business has changed? From the perspective of Mengniu Group, the essence of business has not changed, and it still hopes to sell more and sell more profitably.

The five words digital transformation cannot be called digital transformation if it is only digital construction and is promoted by a single IT department. Digital transformation must be the use of digital means and technologies to support business transformation. In the process of promoting digital transformation of Mengniu Group, the senior management team and business backbones have participated in the top-level design, and found that the essence of business has not changed, but the way of business has changed. Because of technological breakthroughs, consumer habits and the way companies reach consumers have changed.

Zhang Jue explained that the digital age is actually a consumption age without time lag. In the time dimension, consumers can access commercial services all the time; in the spatial dimension, not only physical stores, e-commerce websites, but also various channel spaces can enjoy a consistent shopping experience; in the mental dimension, not only can they meet the needs of daily life It can also meet the customized needs of special groups of people.

Against such a background, how to meet the time-lag-free consumption demands of consumers is a major challenge for enterprises in the digital age. Mr. Zhang Jue borrowed the classic AIPL model to analyze from multiple dimensions such as knowing, understanding & transforming and purchasing, repurchasing, sharing, and retaining, how to build a data-driven, full-link empowerment system around people and goods yard with full-dimensional insight. Mode is the key to the digital transformation of enterprises.

(2) Mengniu’s strategic thinking on digital transformation

Based on such thinking, Mengniu Group has formulated a digital transformation strategy in 2021. By building a new digital and intelligent infrastructure for Mengniu, it will help transform the role of traditional fast-moving consumer goods to technological fast-moving consumer goods, and realize a new Mengniu strategy. Use digitalization to lead business development, and transform from a traditional fast-moving consumer goods company to a technology fast-moving consumer goods company.

Under such a new strategy, the existing business model needs to be completely innovated. In the top-level design, Mr. Zhang Jue summarized three focus directions:

The first is global consumer operation, which requires accurate user insights, differentiated automatic marketing and services, in-depth user operations, and agile demand response.
The second is the ability of omni-channel sales/contract fulfillment. It is necessary to achieve transparent channel management, integration of BC, and online and offline inventory.
The third is intelligent supply and production. To achieve intelligent factory, end-to-end supply chain visualization, and intelligent collaboration between upstream and downstream of the supply chain.

At the same time of business transformation, it is necessary to achieve data empowerment of the entire value chain, realize the precipitation of integrated reusable capabilities, and do a good job in key nodes such as collaborative organization and cooperation.

Specifically, at the technical level, Mr. Zhang Jue proposed that a key trend in Mengniu's digital transformation is to create a three-in-one micro-service architecture collaboration platform for Mengniu. Through the organic combination of dealer trading platform, channel operation platform, and boundless performance platform, the productivity of the business can be maximized and data-driven and empowered business development can be realized.

In addition, Mr. Zhang Jue made specific introductions on key directions such as supply chain business transformation, digital transformation, and manufacturing business transformation. She concluded that data decision-making is the core competitiveness of Mengniu as a fast-moving technology company in the future. To this end, Mengniu Group will escort data decision-making with four major measures: data innovation and application, data governance, data operation organization and data middle platform.

(3) Technological breakthroughs achieve business transformation of Mengniu

Before introducing the help brought by technological breakthroughs to Mengniu's business transformation, Mr. Zhang Jue raised another very classic question: what is the difference between informatization and digitalization? She said that informatization is actually just a support. Initially, some integrated displays were done with numbers, which gradually played a role in reducing costs and increasing efficiency. The current digitalization is a higher-dimensional existence, which itself can help businesses to make money.

After moving from informatization to digitization, the functions of enterprise IT have also changed and evolved into the current dual-mode IT form. Mode 1 does the work of predicting/improving and transforming known areas, and mode 2 is doing adaptive/experimental work to solve new problems, and this is the key to winning digital transformation.

For this reason, Mengniu Group changed the Information Technology Department into the Digital Science and Technology Innovation Department, in which the Digital Technology DT team is equivalent to a middle-office team. capability pack). In this process, the biggest role of the IT team is to sort out and decouple all business requirements according to the degree that they can be assembled and combined, and support the "intelligent collaboration and value co-creation" platform of the assembled enterprise, and finally realize the rapid realization and maximum A degree of reuse and LEGO-style plug-and-play flexibility.

Mr. Zhang Jue finally concluded that the core equation of Mengniu's digital transformation: exponential growth high-dimensional business model = (talent density × innovation breadth × technology intensity) ^ the dimension of the business model. After the sharing, Mr. Zhang Jue also gave wonderful answers to the key questions raised by the participants.

2. The current situation of intelligent retail of fresh food

Chen Hualiang, CTO of Daily Fresh and TVP industry ambassador of Tencent Cloud, gave a keynote speech titled "Intelligent Status of Fresh Food Retailing". Starting from the background of opportunities and challenges in the fresh food retail industry, he introduced intelligent recommendation, intelligent products , intelligent supply, intelligent distribution and other specific use cases of algorithms, and analyzed the intelligent future in his eyes.

(1) Opportunities and challenges

Teacher Chen Hualiang introduced the current situation of my country's fresh food retail market in the opening session. The data shows that by 2020, the transaction scale of the fresh food market has reached 2 trillion, and the weekly purchase frequency is very high. The reason is not difficult to understand. People take food as their priority, and the imagination of the entire fresh food market is very huge.

On the other hand, the large-scale market means that there is often a long development process behind it. In this huge trading network, there are diverse, complex, scattered, and traditional industry characteristics. In such a market environment, Daily Youxian has restructured the entire industry into five key elements: people, stores, front-end warehouses, warehouses and suppliers. On this basis, it has done several things in a targeted manner: business standardization ; Business systemization; system dataization; data intelligence.

To this end, teacher Chen Hualiang selected four aspects: intelligent recommendation, intelligent commodities, intelligent supply, and intelligent distribution.

(2) Intelligent recommendation

Teacher Chen Hualiang pointed out that intelligent recommendation is a very classic ability in the consumer Internet, and it relies on the core technology of recommendation algorithms. Daily Youxian has added algorithm recommendation functions in the homepage, category page, similar search, shopping cart and other scenarios. From manual setting to algorithm recommendation, the conversion efficiency has increased by more than 20%, and 90% of GMV comes from the products recommended by the algorithm. .

Behind this remarkable effect, it is the algorithm ability of intelligent recommendation, which is realized by the technical architecture shown in the figure below. It provides recall service through recall model, and provides sorting service through estimation model and mechanism model. Display services to realize the transaction of goods. Every action of the user behind this will generate a corresponding log, which is abstracted into real-time features and offline features, and iterates continuously through online and offline model training, and finally realizes a complete set of intelligent recommendation algorithm framework.

Teacher Chen Hualiang introduced the technologies behind intelligent recommendation, including features and capabilities such as representation learning, Exploit, Top, and CF, as shown in the figure below.

(3) Smart products

The traditional product selection process relies more on human experience and judgment, but Daily Youxian has found in practice that algorithms can also replace manual decision-making in this process, and the efficiency is generally improved. Specific to the technical framework of smart commodities, it includes expert knowledge and data knowledge. Through these two kinds of knowledge combined with commodity intelligence technology, intelligent decision-making of commodities is finally realized.

Teacher Chen Hualiang gave an example of brand name, commodity name and category identification. He said that the noise from people in this link is very loud, which will have a great impact on the decision-making of smart products. Through the ability of the daily excellent fresh core algorithm technology, effective information is obtained after removing artificial noise, and the accuracy rate is more than 95%.

In addition, in terms of price elasticity in smart products, the price elasticity model constructed by the algorithm can also effectively help to provide price increase or price reduction suggestions.

(4) Intelligent supply

Teacher Chen Hualiang pointed out that under the intelligent supply chain, the core is the forecast of sales. Under the industry characteristics of the fresh food market, the circulation, duration and loss of commodities in each link will affect the actual sales. Therefore, the requirements for sales forecasting are very high, and the calculation requirements behind this are also very high, which cannot be completed by manpower.

Therefore, relying on algorithms to predict sales is a feasible but technically difficult method. Daily Youxian abstracts this problem into a mathematical problem, builds an algorithm model to make corresponding predictions, and continuously optimizes the model to achieve Improve forecast accuracy.

Finally, Mr. Chen Hualiang introduced the application and practical effects of this complex model to the participants based on the actual business scenarios.

(5) Intelligent distribution

The scenario of intelligent distribution also involves a lot of decision-making processes, such as order placement, printing, picking, review, packaging and other operational processes, and then to the selection of delivery personnel, the delivery process within one hour, all require accurate distribution. Under this specific business scenario, Daily Fresh has built an algorithm framework for intelligent distribution. Through the core predictions of outbound time estimation, pressure estimation, ETA estimation, VRP path, etc., the algorithm is adjusted according to the corresponding constraints. The model finally realizes intelligent distribution through algorithms.

(6) The future of intelligence

At the end of the sharing, Mr. Chen Hualiang said that he is very optimistic about the prospect of intelligence replacing manual labor in multiple links, because people should not and do not need to do complicated and tedious work that they are not good at. Taking Daily Fresh as an example, machines can Participate in each element link and get involved in the actual physical world.

In this way, the physical world and the digital world have achieved a perfect linkage, achieved higher efficiency, and finally achieved an end-to-end intelligent high-level operation. He looks forward to the future, fresh food retail will be upgraded to an intelligent fresh food retail industry through the combination of big data, AI and IoT technologies.

At the end of the sharing, Mr. Chen Hualiang also answered the participants' questions about the advanced recommendation algorithm with wonderful answers and wonderful metaphors.

3. Smart Retail: How Technology Helps Retail Best Practices

Wu Chunze, Commercial Director of Tencent's Smart Retail, gave a keynote speech entitled "Smart Retail - The Best Practices of Technology Helping Retail". Starting from the trend of digital transformation in the retail industry, he shared Tencent's layout and segmentation in the retail industry. A classic case in the industry.

(1) Trend discussion: industry digitalization from shallow water areas to deep water areas

At the beginning of the sharing, Mr. Wu Chunze first gave a set of data: 85% of enterprises attach importance to digital strategy, of which "improving consumer experience" is the highest priority category. He said that in the current market environment, digitalization is already a must-have topic.

In such a market environment, there are three very obvious trends in the retail industry, which also indicate that digital transformation has moved from shallow water to deep water.

Trend 1: Omni-channel digitalization brings new business growth. From digital exploration of online channels to digital integration and empowerment of online and offline channels. The data shows that the total online retail sales in China and its proportion are increasing year by year. Among them, the GMV of WeChat mini-programs alone is close to 4 trillion, accounting for 27% of the total online sales. Mini Programs have become the core scenario for the digitization of multiple online and offline business scenarios.

Trend two, private domain has become a new way of life. Enterprises need to use system + operation to provide consumers with a more comprehensive experience. In my country, the private domain now has the characteristics of high penetration, strong stickiness, easy habituation, great influence, and frequent transactions. For example, he said that a clothing company used private domain marketing to achieve an impressive achievement of 300 million GMV in 2 days.

The third trend is that consumers' full life cycle operation brings higher business output. The contacts between consumers and brands have increased, and how to extend the long-term life value of consumers has become a compulsory course for enterprises. Teacher Wu Chunze pointed out that in the era of decentralized traffic, consumers often need 9-15 marketing contacts to realize the purchase action before making a decision. Therefore, it is the key to operate with consumers as the center.

(2) Tencent’s layout: from strategy to organization to fully support industrial interconnection

Under such a trend, Tencent's determination to fully support industrial interconnection from strategy to organization is highlighted. Today's Tencent has become a social retail transaction ecosystem that perfectly complements the social field, content field and trading field.

Teacher Wu Chunze said that Tencent's current resources in multiple business groups and ecosystems have been integrated to provide retail companies with smart retail technology and product capabilities.

At present, Tencent has joined hands with nearly 1,000 companies in various vertical industries of retail to support retail companies to achieve the goal of exceeding 10 billion GMV on the WeChat applet.

(3) The Stone of the Mountain: Interpretation of Cases in Various Subdivisions of Retail Industry

During the sharing, Mr. Wu Chunze also cited a number of cases to interpret the typical cases of Tencent Cloud's cooperation with various retail industry segments, focusing on the scenarios of successful practice to help the retail industry achieve their respective digital transformation best practices.

For example, a shoe and apparel head company hopes to improve the GMV of the brand's private domain, increase the level of users in the brand's private domain, and hope to have innovative consumer operation strategies, but there are problems with the original platform technology architecture, poor continuous iteration ability, and insufficient agility. Weak continuous operation and maintenance capability. To this end, the company joined hands with Tencent Cloud to reconstruct its online business, and achieved the effect of doubling private GMV through a set of GMV growth plans.

Another example is a leading supermarket company, facing the pressure of innovative industries such as fresh food delivery, community group buying, brand self-operated malls and other new channels to impact traditional offline stores, and the platform load capacity requirements are high. Tencent Cloud cooperated with the company to help upgrade the company's self-operated mall and improve its ability to undertake private domain transactions. The system has been online for nearly 3 years, and its stability and reputation are very good.

At the end of the sharing, Mr. Wu Chunze concluded that Tencent Smart Retail can provide industry partners with efficient drainage from the public domain, to refined operations in the private domain, to internal cost reduction and efficiency enhancement, data precipitation and feedback, and then to the full connection of the underlying cloud resources. Ecological digital closed loop to help customers succeed.

In the end, Mr. Wu Chunze also patiently answered the attendees' questions about the existing capabilities and future layout of Tencent's smart retail.

As the sharing draws to a close, the host of this session, Tencent Smart Retail Business Senior Manager Zhang Jun, said that the TVP industry ambassadors' insights into the smart retail industry and their judgment on trends will help Tencent Smart Retail better understand the industry and better help industry partners success.

Epilogue

For a long time, the retail industry, as a large-scale industry with complex roles and complicated links, has a sincere desire for the application of digital technology and data-driven business forms. The digital transformation of the retail industry has never been just a simple so-called intelligent ERP system or a fashionable microservice architecture, but a complete set of end-to-end digital solutions.

In the process of TVP's development, what TVP wants to provide to the industry is not just pure technology sharing or fashionable technical debate, but a thought inspiration that directly hits the source, a little spark that may change the world. From the very beginning of its establishment, TVP hopes to "influence the world with technology", so that technology can benefit everyone, and practice the original intention and original intention of science and technology for good. It is hoped that in the years to come, together with various sub-sectors of the retail industry, and even thousands of industries, we can go deep into the deep-water area of digital transformation.