China Food

Haidilao’s digital intelligence makes the ultimate service online

digitization is both a development problem and a survival problem.
In 1994, Zhang Yong opened the first Haidilao hot pot restaurant in Sichuan. Many years later, Haidilao has become a national direct chain, and has opened its stores overseas. By the end of 2019, Haidilao has 768 global stores with 54.73 million members and 100000 employees.
The increase of the number of stores and members makes the original system slow down. Behind the rapid growth, there must be corresponding digital new infrastructure as the bottom support.
According to Shao Zhidong, chief information officer of Haidilao, digitization is both a development problem and a survival problem. The development of business and technology will naturally drive the digital process of enterprises.
In 2016, Haidilao began to put its core business systems into the cloud;
In 2018, it cooperated with Alibaba cloud to build the infrastructure of data, business and mobile platforms. On this basis, it upgraded the Haidilao super app and restructured its membership system. In the same year, it opened the first smart restaurant in Beijing, using automatic cooking machine, intelligent vegetable delivery machine and intelligent kitchen management system;
In 2020, it is planned to complete all the core business systems in the cloud, and push the automatic cooking machine and intelligent vegetable delivery robot to hundreds of stores.
In the catering industry, Haidilao has been in the forefront of digital intelligence.
From “fishing at sea” to “fishing on cloud”
After more than 20 years of development, Haidilao has established a variety of systems, including 136 systems, such as ordering cash, member management, supplier management, inventory management, etc. However, all these systems were not deployed in the cloud before 2016, but in the traditional IDC room. It is difficult to expand, with high failure rate, and vulnerable to DOS attacks, resulting in network congestion.
In 2016, Haidilao decided to go to the cloud. The first one to go to the cloud is the ordering cash register system.
Going to the cloud is not a simple migration. The original system needs to be reconstructed to solve the problem of compatibility. The interfaces between systems need to be re commissioned or deployed, and the rules of application firewall should also be adjusted.
Haidilao shop by store modification and debugging, ants move, took two years to complete the reconstruction of the cloud. After Alibaba cloud was launched, when the number of stores expanded to 2000 and 5000, the system failure was greatly reduced.
The second cloud is the membership system.
In May 2018, Haidilao cooperated with Alibaba cloud to implement the super Haidilao app project, and the new app was officially launched on October 16. This app project actually does two key things:
The first is the construction of the middle office, that is to establish the infrastructure of mobile, business and data platforms on Alibaba cloud, which is equivalent to digging a good foundation to provide the underlying support for the “building” demand of business departments in the future;
The second is to reconstruct the cloud on the whole membership system to support the demand of high concurrent traffic. The new system can now easily support the number of 100 million members and the activities of tens of millions of participants.
In 2019, Haidilao’s ordering and queuing system has also been moved to the cloud. In 2020, it is mainly to put the PeopleSoft personnel system and SAP ERP logistics systems into the cloud, so that all core business systems from the front end to the back end will be put into the cloud, and Haidilao will fully realize the “cloud fishing”.
After going to the cloud, the network service capacity has been enhanced several times, the infrastructure failure has been basically eliminated, and the efficiency has been greatly improved.
Offline ultimate service extends to online
Haidilao has always been famous for its excellent service. However, for traditional restaurants, the connection with customers will only occur when customers arrive at the store for consumption. When customers leave the restaurant, the connection will be terminated. Now, in addition to offline services, catering brands can also reach them through online ways, and have more interaction with customers.
For example, Haidilao super app is a mobile online portal. Haidilao’s stores are often full during peak dining hours and need to wait in line for a long time. Therefore, Haidilao has spent a lot of efforts on Queuing service, such as providing snacks, manicure, origami, etc.
Now, the queuing problem can also be solved online. On the super app, you can book in advance or arrange the number of the day, and you can choose the seating options such as “by the window”, “by the amusement park” and “relatively quiet location”. The birthday party is also a scene option for students’ party, and special needs such as “preparing baby chair and children’s tableware” can be noted.
Whether customers log in to order through pad, super app or other ports, every consumption information will be deposited in the data platform system, and some personalized membership labels will be generated, such as love spicy food, like vegetarianism, etc.
In this way, when a customer arrives at any store in Haidilao, the server can immediately know his preferences and provide meticulous services, such as delivering ice water he loves to drink, preparing fish soup or cushion for pregnant women, recommending some dishes he likes, or offering a snack he once praised for free, so that the customer can still feel at home and be cared about in a foreign country.
Large middle stage, small front desk
Although Haidilao has established digital systems for a long time, these systems are isolated chimney systems in the early stage, and there are many overlaps and redundancies.
For example, Haidilao app has its own front-end and back-end, and the official website also has its own front-end and back-end. In the past, it used to be a chimney like construction system, “repeatedly building wheels”. In fact, a large part of their capabilities can be shared.
For example, the payment function, before each system has its own set of payment system, causing a lot of trouble, can not quickly unified settlement.
In addition, each system will also involve stores. Every time a store is added, all information must be logged into one system and filled in one by one. The data can not be updated in time, the business can not be real-time collaboration, and the top management of the headquarters can not accurately grasp the dynamics of all stores.
In 2018, Haidilao did a very important thing when cooperating with Alibaba cloud in the super app project, which was to build a business and data platform architecture.
Among them, the business platform is to precipitate the common capabilities in various businesses. For example, many business end will use the payment function. So, it is very convenient to build a sharing platform of the payment center in the business platform for the unified call of each small front desk. In the future, the front desk develops new business, which can be easily expanded.
At present, some consumer related service functions have been set up in Haidilao’s business platform, such as marketing center, order center, payment center, etc. in the future, midrange capabilities related to internal operation and supply chain will also be established to form a “large middle platform and a small front desk” to make the business more agile and efficient.
In order to enable the technology department to better meet the needs of the business department, Haidilao has created a new position called itbp (IT business partner), which is dedicated to meet the needs of the business department, and decompose and translate it into it requirements that developers can understand.
The new Haidilao app emerges as the times require on this dual middle platform and mobile middle platform which is dedicated to mobile application service. With such an infrastructure, Haidilao can quickly iterate various functions on the app according to its own business needs.
At first, the app only had basic functions such as queuing, booking, ordering, and takeout. Now, it has added various functions such as shopping mall, popular activities, games, communities, short video sharing, etc., and intelligent customer service is online 24 hours a day. Basically, the functions and layout will be iterated every month according to the changes of consumers’ demands. The establishment of the middle platform makes such rapid iteration easier.
Waiting area interactive games
Data driven refined operation
In addition to building various applications on the business platform, Haidilao has a very important work in recent years, which is to extract and integrate the scattered data from various systems and inject them into the data platform.
Cheng Qiwu, Minister of Haidilao’s information department, said that the role of data center is first to unify the storage, standardize the data generated by various business systems and integrate them into various master data (Masters) Data (shared basic data), such as the master data of dishes and stores, and then provide the master data to all business parties with the ability of terminal API (application program interface), so as to realize cross department and cross system data sharing.
Secondly, the data center also has a great advantage in timeliness. In the past, Haidilao used to extract data from various business systems every day, which may be a T + 1 (day) model, and in the traditional data warehouse, the calculation is very slow. Now, with alicloud’s dataphin intelligent data construction and management platform, plus a layer of maxcompute big data computing service, the calculation is very fast.
Third, because the data center can timely feed back the unified data to the business system, it can make some attempts and improvements in the business based on the data, such as the recommendation of dishes – “guess what you like” and so on.
In the data center, all kinds of membership tags are generated by collecting, settling, analyzing and processing membership data from various ports.
Haidilao’s member operation Department can check these tags in the marketing center of the business center to delineate the customers they want to reach, and automatically carry out accurate marketing for them through the system, such as pushing the contents or activities of related topics in the app.
This means that each member may see different contents when opening Haidilao super app, which is similar to Taobao app’s “thousands of people and thousands of faces”.
For example, some customers often go to Haidilao for consumption during the night snack. When the crayfish products suitable for supper are launched, Haidilao app will push the crayfish tasting activities to them; some customers have consumed Haidilao beer, and when Haidilao beer comes out with a new model, they will also make corresponding contact with these consumers.
Intelligent algorithm improves decision efficiency
As one of the important paths to achieve the strategic goal of “changing the cost structure of catering enterprises with new technologies”, Haidilao launched its first smart restaurant in 2018, exploring the use of digital and intelligent means to improve customer experience, reduce costs and increase efficiency, and ensure food safety.
How to improve the intelligent degree of the whole chain from the production and supply side to the consumption end, the smart restaurant has made a beneficial attempt.
The core of hot pot is the bottom and seasoning. Everyone has different acceptance of spicy and spicy flavors. Haidilao has launched the “private customization” of the bottom of the pot. According to the customer’s requirements for spicy, salty, sweet, and oil-water ratio, the waiter can transfer out a flavor and store it in the CRM system. Through the precise configuration of raw materials, auxiliary materials and fresh materials, the automatic pot matching machine can produce a customized pot bottom.
Customized bottom
In order to ensure food safety, the intelligent food warehouse of smart restaurant maintains a constant temperature of 0 ~ 4 degrees, and the mechanical arm automatically collects and delivers food. Each dish has an RFID tag, which will alarm automatically after more than 48 hours, and will be eliminated from the shelf. After the dishes are made, several intelligent food delivery robots share the work of delivering the dishes to customers’ tables.
Intelligent vegetable delivery robot
In order to reduce cost and increase efficiency, smart restaurant cooperates with Ali in inventory management, forecasts sales volume through “intelligent demand” algorithm, tracks inventory in real time, looks at how many ingredients and dishes are consumed every day, judges how much goods need to be replenished, and keeps the inventory at a reasonable level, so as to reduce losses and ensure supply.
This algorithm needs a deep learning process. As more stores apply intelligent purchasing algorithm, accumulate more high-quality data over a long period of time, and consider more decision-making factors that may affect the results, the prediction effect will be improved.
For catering enterprises, site selection is another crucial link. Haidilao cooperates with Alibaba cloud and Gaode to carry out big data intelligent site selection.
In the past, they used to run around with their own experience. According to the average speed of opening a new store every 1.2 days in Haidilao in 2019, if all the developers run on site, they will break their legs. Now they sit in the office and point out the data analysis screen. The intelligent algorithm will comprehensively evaluate the factors such as population density, population consumption level, store distribution, surrounding catering and shopping facilities, and transportation convenience, so as to improve the probability of successful site selection 。
The digital intellectualization of catering enterprises is a hard road. As Shao Zhidong said, intellectualization is based on digitalization. Only on the basis of big data can there be so-called artificial intelligence. For catering enterprises, one size for one thing is a big mountain. The cost is too high. It is not only the catering enterprises themselves, but also from the food materials provided by the suppliers, they should do one thing at a time.
Yang Li, head of Haidilao supply chain, also believes that it is difficult to standardize the raw material supply data of catering enterprises. For example, if the beef and mutton are cut in half and put back, the spinach yield rate is 85% in cool weather (85% in a catty of spinach) and 50% in hot weather, which makes it very difficult to digitize.
However, despite the huge challenges, Haidilao is still firmly on the road of digital intelligence.
When it has only a few stores, it can do well through the service of meticulous people; when it expands to hundreds or even thousands of stores, only digital intelligence can form scale effect, healthy development, and eliminate the “wealth disease” of large enterprises, such as bloated system, waste of resources and low process efficiency.
Now, Haidilao has a good underlying infrastructure – data platform and business platform in the cloud, which can support the rapid development of business in the future and accurately grasp the consumption trend and competitive environment.
With the development of technology and the promotion of digital intelligence in the whole industry, Haidilao’s challenges in the supply chain will eventually be solved, and the service will be leapfrogged from meticulous to intelligent.
Author: Wu Beihong; source: new retail business review (ID: xinlingshou1001), reprinted with authorization. Join the community: Cherry (wechat: 15240428449); business cooperation: Amy (wechat: 13701559246). the “careful machine” behind Nestle’s accelerated business restructuring | Mintel Asia Pacific food and beverage industry trend | will 2021 be the first year of cheese snacks? |Japanese yoghurt: a century of bumpy road | how to stand out from the competitive category track | U.S. yogurt has been popular for 40 years |
daily new product collection: Isee trends | editor’s | reader’s

Read the original

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *