news

Precise analysis of vegetable sales can also rely on algorithms

2024-08-07

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

How far has a seven-year-old fresh food e-commerce platform evolved? "The outside world has always misunderstood Dingdong Maicai as a forward warehouse company. In fact, Dingdong Maicai is essentially a food supply chain company focusing on fresh food." In the view of Dingdong Maicai CTO Jiang Xu, the platform can drive company decisions with full-link digital capabilities, strictly control losses and strengthen turnover through strong data accumulation, and forge the resilience of the fresh food supply chain. Although building a digital algorithm model is a hard job, it has also strengthened Dingdong Maicai's muscles and bones, allowing it to gain a firm foothold in the fierce competition.

Full-link digitalization

Consumers can get fruits, vegetables, grains and oils delivered to their homes within 30 minutes by just moving their fingers on the fresh food e-commerce platform. Behind this consumption method that people have long been accustomed to is the result of the long-term accumulation of digital technology capabilities in forward warehouses. According to Jiang Xu, Dingdong Maicai can now achieve digital drive for the entire chain of people, goods, warehouses and transportation, and the end-to-end loss can be controlled at 1%-2%. Behind this, the platform's forward warehouse model has undergone countless iterations.

In recent years, in order to gain a foothold in user consumption decisions, major e-commerce giants have increased their investment in instant retail, and the competition in near-field e-commerce has become particularly fierce. The forward warehouse model has broken the limitations of traditional retail fulfillment with its outstanding product delivery speed and service quality, and has become a major player in the short-distance track together with models such as errand running.

Take Dingdong Maicai as an example. The platform has deployed more than 1,000 forward warehouses across the country. Each warehouse has an average of more than 3,000 SKUs, which means there are more than 3 million combinations. This requires companies to have high operational efficiency and fresh food loss control capabilities. To this end, Dingdong Maicai has refined the granularity of planning and execution on the basis of full-factor digitalization. Every month, the company will regularly inventory and adjust various resources including traffic, marketing, categories, inventory, warehouse manpower, etc.

Jiang Xu said that in a large number of business scenarios, the company's business decisions are mainly driven by data algorithms, and digitalization has penetrated into various links such as forecasting and allocation, search recommendations, marketing pricing, user growth, fulfillment scheduling, warehouse network planning, and intelligent scheduling. The platform has established a relatively complete data indicator system, which can update many key indicators in real time at the minute level. The full-link transparency brought by digitalization can also enable the platform to achieve full traceability.

A typical benefit brought about by this is that the platform can observe the inventory and sales status of goods in real time. If it is found that the goods are overstocked or understocked, the platform can make adjustments in advance, such as conducting targeted discount promotions to attract users to buy and reduce the risk of loss.

The “cost” of innovation

Looking back at 2017 when Dingdong Maicai was founded, it was the period when the new retail format was booming. Many fresh food e-commerce companies, including Dingdong Maicai, competed for the market with home delivery and self-pickup models. Seven years later, some people are unable to get up, while others are struggling to survive. Forward warehouse companies were once questioned by the industry due to huge losses.

"In the early days, Dingdong Maicai lacked experience in the fresh food retail industry and paid a lot of tuition fees." Jiang Xu confessed to the Beijing Business Daily reporter the difficulties in the company's start-up period. In the past few years, every time there was a typhoon, heavy rain, high temperature, or snow, Dingdong Maicai's forward warehouse resource allocation mainly relied on human judgment, which was prone to problems such as warehouse explosions due to deviations. In order to change this situation, in recent years, the company has standardized the situation of severe weather in various cities through a large amount of data precipitation and analysis, and transformed human judgment into part of the system and model.

Taking the weather impact of heavy rain as an example, Dingdong Maicai can now analyze data such as rainfall time, rainfall, wind speed, etc., and combine it with sales fluctuations of specific categories to decide whether to adopt conservative or aggressive strategies for product stocking and marketing. In order to achieve such a refined prediction, Jiang Xu believes that the platform needs at least two years to accumulate various data indicators to obtain a relatively stable model. If a decision is wrong, it used to take a week to adjust, but now it can be shortened to 1-2 days.

Dingdong Maicai has chosen a difficult and long road. "This is the 'price' of innovation." Jiang Xu further explained that Dingdong Maicai adopts a "top-down" strategy, first building a full-link system, using data algorithms to drive each link, running first and then iterating. Once the flywheel is running, the process or data problems are supplemented according to key issues, forming a cycle of "try-error correction-feedback", and the optimization speed can be faster and faster.

Compared with the "bottom-up" digitalization strategy of traditional retail companies, Dingdong Maicai can hardly avoid paying a high price in the early stage. Traditional companies need to find ways to connect multiple systems and data based on existing outsourcing or self-built independent systems. Although the initial investment is small, the difficulty of digital reconstruction in the later stage will be very high.

At the same time, Dingdong Maicai has accumulated profound digital capabilities, but the company is not "superstitious" about digital technology. In terms of human-machine collaboration, Dingdong Maicai chose a compromise solution, that is, the model and system make judgments, and the operators have the right to modify. The company will control the modification behavior and assess the employees' adoption rate of the system judgment. "At present, the overall adoption rate can exceed 80%, which means that both parties can achieve a win-win situation." Jiang Xu said.

"Deadly Fight" Fresh Food Supply Chain

"Many people think that Dingdong Maicai is a forward warehouse company. In fact, the goal of Dingdong Maicai at the beginning was to become a food supply chain company focusing on fresh food." Jiang Xu said that Dingdong Maicai focuses on the fresh food supply chain and realizes the end-to-end supply chain process transformation in fresh food categories such as vegetables, pork, and aquatic products. "Since last year, the platform's vegetable category has achieved profitability. Unlike other companies that use fresh food as a traffic-generating product, Dingdong Maicai's fresh food category has always made a significant contribution to sales." This also means that Dingdong Maicai is already ahead of the industry in building a fresh food supply chain.

The financial report shows that in the fourth quarter of 2023, the user penetration rate of Dingdong Maicai's own-brand products has reached 73.6%. As of now, Dingdong Maicai has more than 20 own-brand products.

Next, Dingdong Maicai will continue to transform its supply chain. According to Jiang Xu, Dingdong Maicai began to expand from the dining table scene to the coffee table scene last year, focusing on leisure food. For example, in the fruit category, because fruits are easily damaged and the time to market is uncertain, the industry usually picks them at 50% to 60% maturity. At the same time, the flavor and sugar content will also be discounted. To change this situation, Dingdong Maicai has improved its demand forecasting capabilities, shortened the platform's shelf life, and compressed the cycle from picking to the table, so that fruits can be delivered to consumers at the optimal maturity.

At the same time, Dingdong Maicai will also move some manual operation links of the supply chain online. For example, there are still many links in the loss of fresh aquatic products in the aquatic product supply chain that need to be filled in manually. After the online transformation, it can help the platform model predict more accurately, improve data quality and control losses.

"We hope to be able to achieve data algorithm-driven in more business scenarios." Jiang Xu emphasized that since last year, the platform, including warehouse site selection and planning, has all been driven by algorithms, and has achieved good results in terms of fulfillment efficiency.

Making digital technology more comprehensive, even down to the minute level, is the direction that Dingdong Maicai will focus on in the future. Since the production cycle, production scale, and market demand of fresh produce are different from those of standard products, it is difficult to achieve a low price and scale cycle through economies of scale. Jiang Xu believes that if fresh produce retail is to succeed, it essentially depends on changing the supply chain to improve efficiency from production to consumption.

"Improving efficiency through digitalization and automation, thereby reducing losses, increasing profit margins, achieving supply-side reform, and developing more good products, this is what Dingdong Maicai will continue to do in the future," said Jiang Xu.

Beijing Business Daily reporter He Qian

Report/Feedback