Produced by Radar Finance and | Li Yihui, ed. | Deep Sea
The technocracy convenience bee is caught in the storm of public opinion.
In the impression of everyone, convenience bees are a chain of convenience stores distributed in first- and second-tier cities, where many office workers buy brunches and snacks. But what consumers find difficult to realize is that the cameras densely packed in the convenience bee shop act as a pair of eyes, and the "system" behind it is collecting customers' every move.
According to an investigative report by Southern Weekend, in convenience bees, such cameras not only stare at customers and employees, but also extend from the store to the outside of the store. In order to select the site, the convenience bee adopted the method of "sticking and shooting", shooting in residential areas, office buildings and even at the door of state organs, accumulating hundreds of thousands of shooting data in the background.
In this regard, the Beijing News commented that the crazy collection of passers-by information has facilitated the bee to break through the commercial boundary.
In June this year, the convenience bee was rumored to be listed in the United States. The company denies this. According to Tianyan, the company has disclosed three rounds of financing so far, and the investors include Zebra Investment, Tencent Investment and Gaoling Investment, and the latest valuation exceeds 10 billion.
Some analysts believe that data security issues may affect the listing of convenience bees.
"Crazy collection" of public image information
The founder of Convenience Bee, Zhuang Chenchao, is a serial entrepreneur whose most recent venture was Zebra Investment, where qunar, which was previously founded, was sold to Ctrip.
In December 2016, Zhuang Chenchao invested $300 million in convenience bees, and the founding team also included the original 7-11 and neighbor employees. In this venture, Zhuang Chenchao studied for a long time before he took the shot, and he wanted to find "a thing that has not been done in China, with an experimental nature." ”
But at that time, convenience stores were not on the outlet, and opening convenience stores in Beijing was not optimistic about the outside world, and 7-11 only opened more than two hundred stores in Beijing for more than a decade.
For the reasons for choosing the convenience bee project, Zhuang Chenchao told the media in September 2017 that the convenience store industry can be completely driven by algorithms, based on a large number of data analysis to form a "thousand stores and thousands of faces", from the site selection, selection, price display can give guidance.
In February 2017, the first store of Convenience Bee opened in Zhongguancun, Beijing. By the end of 2017, the total number of stores was only 60; at the end of 2018, the number was 570; by the end of 2020, it had opened 2,000 stores.
Such rapid expansion has benefited from the evolution of the data collection capacity and means of facilitating bees. In the early days of entrepreneurship, the company used apps and shared bicycles to understand the user's gender, age and other data. But soon figured out a way for cameras to capture information, and the company set up a special information collection department to hire full-time information collectors to collect all valuable data.
According to Southern Weekend, the way to facilitate the collection of bee information is to shoot videos, which are called "stickers" internally.
Zhuang Chenchao has publicly stated that 70-80% of the company's number is collected through vision, and a very small amount of power-using equipment will be collected with IoT, such as temperature, sound, etc. He even cites data collection as the "core advantage" of convenience bees, "we actually have no human judgment when we sign a contract, and we don't need any manual judgment or experience at all."
According to the aforementioned report, taking the location of the new store as an example, the targets include the community, school, office building, training institution, competitors, hospitals and government agencies within one square kilometer of the target store.
According to the director of convenient bee collection, some Beijing government agencies, including the Ministry of Commerce, the Ministry of Communications, the Ministry of Environmental Protection, the National Tourism Administration, the State Administration of Cultural Heritage, and the General Administration of Customs, are also within the scope of the posting. The total number of data for all office buildings within the fifth ring road in Beijing is as high as 43,000, and about 20,000 in Shanghai.
For the media's verification, the convenience bee replied that there has been no "secret shooting" behavior since the establishment of the company, and at the same time denied the shooting of residential quarters, business circles and national ministries and commissions. The reply stated that the source of the company's information was the purchase of third-party data.
However, the above report mentioned that the login of the convenience bee's "gold store management background" system confirmed the existence of information collection. And after the report was made public, the convenience bee has not yet clarified the statement.
Algorithm-driven "systems"
With the data, and then by the machine learning algorithm processing to form an automated decision-making, replacing the judgment of the store manager and the clerk, this is Zhuang Chenchao's proud innovation, such a system has always been the focus of the convenience bee.
In December 2020, in a program produced by Qingteng University (now renamed Qingteng), Zhuang Chenchao introduced how to build a digital model suitable for convenience store retail.
The first part that facilitates the bee to cut into is the underlying ERP system. Then enter the order, optimize the selection, scheduling, manpower links, and finally display, logistics optimization are handed over to the system automation to deal with. In this process, as far as possible, the data collection comes from the interaction between the device and the system, so that the accuracy and timeliness of the data are guaranteed.
When the system has an enabling effect, the algorithm gives an optimal solution. In order to solve the problem of strong sense of balance in human decision-making but weak adaptability, the convenience bee gave up manual intervention in all links, and from factory, logistics, display and even cleaning to the system for rapid decision-making, the model was able to run through.
At the same time, convenience bees have invested a lot of IT staff, and a vice president of a Japanese convenience store chain said that convenience bees MAY be four to five times their OWN staff. Convenience Bee officially replied on Zhihu: "On the surface, Convenience Bee is a chain convenience store, but in fact we are a data technology company. ”
Full chain automation, letting data-driven systems dictate, doesn't seem like good news for convenience bee clerks. Because employees have to face the camera without dead angle coverage in all directions, and the brain behind them. The system automatically identifies the picture in the store through manual and AI, and automatically alarms once it does not meet the requirements.
In a convenience bee on the street of Beijing, Radar Finance noticed that counting the two probes above the front and back doors, the store has 12 cameras, whether it is customers or employees, there is nothing to hide from the moment they enter the door.
When working, the clerk sorts out the shelves, cleans, heats fast food and discards expired products when taking photos and uploading, and there will be special people in the background to retrieve monitoring and comparative inspection.
To cope with this job, convenience bee clerks need to memorize up to 300 to 400 standard operating procedures. With this powerful system and algorithm, you can train a store manager in 6 months and a clerk in 5 days.
Customers' every move after entering the store will also be collected and become the basis for systematic decision-making. In addition, in the Zhihu Q&A, the company's official account once described, "Different from general convenience stores, convenience bees can capture the consumption data of local consumers on relevant e-commerce platforms, the review data of various social platforms, and the past sales data of other stores... Put in the formula, calculate the income ranking, and then choose the product. ”
With such an algorithm model, convenience bees not only have the ability to quickly copy, but also achieve operational efficiency improvement. Xue Enyuan, executive director of Convenience Bee, has publicly announced that by the end of 2021, Convenience Bee will open 4,000 stores and is expected to open 10,000 stores by 2023.
Compared with Lawson, which also has the "Ten Thousand Stores Plan", which has a franchise model, all the stores of the convenience bee are directly operated. Tens of thousands of directly operated stores are almost an impossible task in the industry, but the confidence of convenience bees comes from data advantages.
"In the past three or four years, convenience bees have opened nearly 700 stores in Beijing, the 'desert of convenience stores,' and they've been profitable." Zhuang Chenchao said that the machine learning algorithm has become more and more accurate after continuous iteration, and a more obvious feature is that the more newly opened stores, the higher the sales volume, the lower the proportion of rent.
Of course, there are also people who don't think that data-driven is necessarily low-cost. The vice president of the above-mentioned Japanese chain stores said that changes in climate, temperature, water, and surrounding environment may require dozens of dimensions of parameters, and it is very difficult to be fully automatic.
How far can alienated convenience bees fly?
The convenience bee, which regards itself as a data company, has a valuation of more than 10 billion yuan and has twice reported the news of listing.
In May 2020, convenience bee insiders revealed that the company has received hundreds of millions of dollars in Series C investment, and the cumulative raised funds have reached 1.5 billion US dollars (about 10.704 billion yuan), including the world's top private equity funds, national sovereign wealth funds, university investment funds and super large Internet companies.
Prior to this, on October 15, 2018, Convenience Bee received a US$256 million (about RMB1.827 billion) strategic financing from Tencent Investment and Hillhouse Capital, which was valued at US$1.6 billion (approximately RMB11.418 billion).
Since the beginning of this year, the news of the listing of the convenience bee has been reported twice, but both have been denied by the company.
On June 23, it was reported that convenience bees recently launched an IPO in the United States, planning to raise 500 million to 1 billion US dollars, goldman Sachs, Morgan Stanley, CITIC Securities and others as the lead underwriters, as soon as the end of this year or early next year to the United States to list. The convenience bee responded that "this message is not true"
On July 8, Bloomberg reported that convenience bees had secretly filed an IPO application to the United States or raised $500 million. People familiar with the matter said that convenience bees are working with advisers to sell shares, which could be done as soon as 2021. Since then, it has been denied by the convenience bee.
There has been no financing for more than a year, and for the rapidly expanding convenience bee, alleviating the financial pressure or the main reason for going public. However, according to the current regulatory environment changes, going public in the United States is obviously not the best choice.
However, before going public, the data security problems of convenience bees still need to be clarified. Article 6 of the Measures for the Administration of the Beijing Municipal Public Security Image Information System clearly stipulates that image information in public areas outside the scope of the unit shall not be collected, and no other unit or individual may set up a public security image information system in that area.
Many analysts also pointed out that the large-scale video collection of convenience bees in public areas is suspected of violating public privacy. Because the appearance of natural persons in public is not equivalent to giving up their right to privacy, for example, they do not want to be seen on certain occasions, or they expose their whereabouts and take pictures of them without permission, which is an infringement. In addition, there is also a great risk in how enterprises collect portraits and action trajectories of a large number of citizens, how to manage and use them. If this data is used for purposes other than business, or leaked out, it will leave the public's private data in a state of naked running.
Note: This article is the original of Radar Finance (ID: leidacj). Unauthorized reproduction is prohibited.