Need for Speed 2: Can the front-end warehouse model ever be profitable?
MissFresh versus Dingdong Maicai and the market of fresh produce delivery
Still from Dingdong promo clip.
Things that caught our attention
The front-end warehouse model on paper
Two IPOs and the fall of MissFresh
Others using the front-end warehouses
Can the front-end warehouse model be profitable?
Front-end warehouses versus alternatives
Things that caught our attention
James Hu recently interviewed Rui. They talk about AI, starting the Tech Buzz China podcast, distortion of truth, her current work as COO for a start-up and more. You can watch the interview on YouTube.
Earlier this year, Ed delivered a two-part session on how Chinese e-commerce changed in the past decade and how some of these innovations are coming our way. The session was filmed, and the first half on the evolution of e-commerce in China is now available on YouTube.
EV maker BYD sold 2.07 million vehicles in the first nine months of 2023, up 76% from a year earlier. We wrote about BYD earlier this month in Manpower over Machines: The Calculated Gamble of BYD.
Temu, Pinduoduo’s cross-border platform, is now available in 47 countries. Temu is rumoured to be planning to charge merchants a 0.5% commission for outsourced customer service. (source) Temu has denied the rumours.
Temu has also been shown to impact the bargain retail business. We wrote extensively about Temu in Temu: from $0 to $3 billion in 10 months.
Cotti Coffee has been recruiting franchisers outside China. We wrote about Cotti last months in Coffee Wars: How the arrival of Cotti has sparked a price war.
Introduction
A few months ago, we explored two different forms of instant retail: delivering goods within one hour from local businesses.
In Redefining Fresh: Alibaba's Freshippo and the New Frontier in Grocery Business, our article on Alibaba’s Hema supermarkets (Freshippo), we saw an example of the store + warehouse model. In this model, orders are picked in a supermarket and attached warehouse and are home-delivered within a 3-5 km radius. Some other supermarket chains like JD.com’s 7Fresh (七鲜 Qixian) and Yonghui have also followed this model, while Meituan pulled the plug on its Ella (小象 Xiaoxiang) Hema-clone a few years ago.
In Need for Speed: Instant Retail, we discussed a different form of instant retail in which marketplace platforms like Meituan’s Shangou, Dada Nexus’ JD Daojia and Alibaba’s Ele.me enable offline retailers to open online shops and provide them with couriers for home-delivery. Many Chinese retailers, from the smallest pop-and-mom shops to big chains like Walmart, have used these third-party services, and some have seen the share of online orders in their total sales grow to as much as 50% through this model.
In this new article (and an upcoming one on Meituan), we will explore the remaining business model of front-end warehouses (前置仓 qian zhi cang). This is the domain of businesses like Meituan Maicai, Dingdong Maicai and the now-defunct MissFresh. How do these platforms work, and are they profitable? Why did Missfresh perish while Dingdong survived? And who is Pupu, another important player in the market?
We hope you’ll enjoy this deep dive into the sector.
Freya Zhang, Ed Sander & Rui Ma
(Click on the images above for information on the Tech Buzz China team.)
The front-end warehouse model on paper
As with other forms of grocery delivery, the interest of China's internet companies in front-end warehouses lies in the frequency of traffic they can generate for their apps. Having the consumer return several times a week offers recurring opportunities to pull them further into the ecosystems and sell more goods and services than just groceries. Among the leading players, this is especially true for Meituan.
Unlike Hema and JD.com's 7Fresh, companies that operate through front-end warehouses typically have no stores but only small- to medium-sized warehouses close to residential areas or office buildings. These front-end warehouses, sometimes called DMWs (Distributed Mini Warehouse), normally deliver within an hour in a radius of 3 kilometres. Front-end warehouses have multiple temperature facilities such as refrigerators, freezers, and sometimes reheating for fresh product processing.
Still from Dingdong promo clip.
Rental costs for front-end warehouses are said to be lower than for a store (according to MissFresh, just one-third), and they also allow for storage of more products on the same surface. Shops often cover 800 to 1000 square metres, and a front-end warehouse is 200 to 400 square metres. As a result, the sales per square metre is about three to four times higher than at a supermarket, or so MissFresh claimed. That might be true, but the warehouses also have limited SKUs, putting a ceiling on the average order value.
Front-end warehouse companies usually have a high proportion of fresh produce, around 55%. The most prominent feature is the aquatic and seafood categories. This category doesn’t just attract many customers; it also provides high margins.
In traditional retail companies in China, spoilage rates for fresh produce are often high because retailers, for cost reasons, usually do not use refrigerated transport between regional warehouses and shops. Front-end warehouses tend to purchase many of their products directly from the source and operate their own logistics chain with only a regional (city) warehouse and a front-end warehouse between the farmer and consumer. As such, the supply chain is much shorter than for supermarkets, which in China usually have to deal with various layers of producers and wholesalers. Products, therefore, have a shorter total transport time, and less fresh produce is lost. MissFresh, for instance, claimed to be able to limit spoilage loss to 2.5%.
In China, regular supermarkets have a catchment area of 500 metres to 1 kilometre. Front-end warehouses extend this to 3 kilometres through home delivery. A company can ensure good coverage with multiple front-end warehouses in a city. If the population density is high, the warehouses need to be closer together to guarantee fast delivery and to keep delivery profitable. Due to the high population density in Chinese cities, MissFresh delivery drivers could deliver around 60 to 70 orders daily, and a courier could take 6 or 7 orders per trip.
Population density is often lower outside city centres, so the delivery time is longer. The front-end warehouse model no longer works if the population density is too low. Hence, companies using this model tend to focus on China's 30 to 40 largest cities. In addition to the higher population density, consumers in these cities have a busier life, more disposable income and are more interested in the convenience of grocery delivery. But as we will see, even those cities can prove to be a challenge for the model.
Indeed, let's not get too enthusiastic about the supposed advantages of front-end warehouses. The model has proven difficult to operate unless there is enough scale (per warehouse). Spreading resources too thinly can lead to disaster, as we will see.
In addition to the cold chain, inventory management is challenging for these platforms, just like for supermarkets. Which products should you put in those front-end warehouses? On the one hand, you must avoid stocking too many products and throw away as few unsold perishable products as possible. On the other hand, you could have too few products in stock and lose customers because of frequent out-of-stocks.
The players in this sector use artificial intelligence and big data to try and solve these problems. For example, based on available data about consumers in the catchment area, Meituan Maicai's product range is adapted to consumers' needs and preferences. Dingdong also uses big data to predict future orders and minimise waste. In their apps, they give individual customers recommendations based on data to increase the order value and offload leftover stock.
Front-end warehouses compared to other grocery delivery models. For more on community group buying, read our earlier report.
Because of the investment costs, front-end warehouse players usually started in a few cities and, if possible, in places where the competition was not yet present. A few years ago, you could see different players popping up in various parts of China.
The model snowballed since it was introduced in 2015, and according to iResearch, its market was RMB 33.7 billion five years later. At that time, it was expected to grow to RMB 306.8 billion by 2025, and Missfresh and Dingdong held 23% and 39% market share respectively. [1]
Direct sourcing
It is well-known that only selling fresh produce is not profitable because of low prices and fierce (price) competition. Like other fresh-food e-commerce companies, front-end warehouse companies try to shorten the supply chain to minimise procurement costs and spoilage. They cut out wholesale suppliers and middlemen through direct sourcing by buying goods directly from farmers. They sometimes even set up self-operated farms and processing factories.
For more popular product categories, they adopt a series of supply chain strategies from procurement to fulfilment, working backwards in the supply chain when the customer and order scale are large enough. With the accumulated data, they can predict order volume per day and season, allowing them to cooperate with suppliers to obtain more stable supply and lower costs.
Some large platforms choose direct sourcing from the origin for high-volume products like seafood, potatoes and vegetables. These commodities usually are not processed and shipped straight to the transit or regional warehouses. The elimination of middlemen reduced costs, and shortening the delivery time improves product quality. No direct source will be interested if the sales quantity is too small. Unlike secondary suppliers, source suppliers usually only deliver full trucks. Therefore, no fresh-food e-commerce company can directly source all of its assortment.
Still from MissFresh promo clip.
Products purchased from the place of origin need to be cleaned, selected and sometimes processed. Therefore, it is sometimes necessary to establish origin warehouses close to the source to improve the efficiency of packaging, sorting and delivery. While direct sourcing means lower procurement costs, it often increases investments in the required origin and processing warehouses.
Direct sourcing is a trend you don’t just see with front-end warehouses but has also played an essential role in Hema’s recent profitability. In the future, whichever fresh food e-commerce company can promote direct sourcing faster will get better development opportunities.
Marketplace platforms like JD Daojia and Meituan Shangou, discussed in Need for Speed: Instant Retail, don’t engage in direct sourcing. Their speciality lies in quickly getting goods from A to B, not in controlling the supply chain. They aren’t retailers; they are service providers.
Let’s look at the leading players that have dominated the front-end warehouse sector in recent years.
MissFresh
MissFresh's story is one of initial success, followed by a crash and burn. It tells us a lot about Chinese internet business and is therefore worth re-evaluating.
Still from MissFresh promo clip.
Founded in 2014 by ‘boy wonder’ Xu Zheng, MissFresh (每日优鲜, Meiri Youxian) was the first player to pioneer the front-end warehouse model. [2] Each warehouse had refrigerated storage capacity, one manager, 5-10 sorting staff and 10-20 delivery staff, bringing goods to the consumer in 30-60 minutes in a 3 km radius. Through customer subsidies, it hoped to develop habits of frequent online shopping with its customers. [3] MissFresh claimed that delivery from a front-end warehouse was about ten times more efficient than from a store and used this cost advantage to compete with supermarkets on price.
MissFresh was backed by Tencent with a $450 million funding in 2018. In that year, MissFresh sold RMB 10 billion worth of products. According to Trustdata, the company had a 50% share of the online market for fresh products in 2019, followed by Hema with approximately 26%. [4]
At the start of 2019, MissFresh had set a GMV target of 20 billion and full profitability in the same year. However, its IPO prospectus showed that 2019 only saw RMB 7.59 billion GMV and a 2.9 billion loss. [3] As we will see, it would never reach profitability…
MissFresh claims to source 93% of fresh products and 80% of all products directly from their source and have a loss rate of only 2.5%. [1] It had a minimum spend of RMB 39 for free delivery. In 2020, its customers ordered 8.2 items per order and in Q2 2021, it saw an average order value (AOV) of RMB 96.1 and an average number of daily orders per warehouse of 417. [5]
On average, MissFresh customers would use the app four times a month. Older customers bought slightly more frequently than young people: fifty and sixty-year-olds bought every one or two days, while thirty-year-olds sometimes bought only once every two weeks. The latter group does less home cooking, buys more goods at once, and stocks up.
Still from MissFresh promo clip.
The number of SKUs MissFresh offered grew as it added new product categories; when it launched, it had only 800. The limited number of SKUs gave the company a good negotiating position with its suppliers, being able to buy in large volumes. Of the SKUs that MissFresh supplied, about 70% were the same throughout China, and 30% were regionally selected and purchased. Eventually, MissFresh settled on 4,000 SKUs, about a tenth of what many supermarkets offer on their shelves. [6] It could, however, also deliver another 20,000 SKUs the next day. [1]
Fewer variants and, as a result, larger volumes provided benefits in the cold chain. Different products have to be transported at different temperatures. For example, apples are sensitive to cold, while durians should be transported frozen. In other words, they cannot go to a warehouse in the same truck. Larger volumes, however, ensure well-filled trucks and, therefore, lower logistics costs.
Jack Yang, at the time, Chief Growth Officer at MissFresh, shared some of the company’s strategies in a Tech Buzz China Live Cast. “MissFresh was also driven by penetration of mini-programs among 500 million users. We were creating a lot of gamification mechanisms, such as giving cash rewards to users who shared a red envelope with their groups. WeChat allowed that and in some ways encouraged us to give money away to groups to grow the penetration of mini programs.”
“Because labour was still relatively cheap in China, we relied on many foot soldiers, in-person salespeople, to drive our growth. For example, we hired tens of thousands of foot soldiers to go from apartment to apartment to do in-person sales. We asked people to join a WeChat group, and if you joined, we gave you a 5 yuan coupon. We created a new group each time a group reached about 200 to 300 people. We used bots to moderate these groups. We did a lot of these relatively manual things. And because the labour is cheap, were still able to scale these operations to 12 cities and have coverage of millions of users.”
By mid-2020, MissFresh operated over 1,500 front-end warehouses, serving 25 million monthly active customers. When it opened a new warehouse in a city it was already active in, it claimed to break even in three months. In new cities, it would take six to nine months.
MissFresh app.
Funding dries up
Until 2018, MissFresh's financing went smoothly, with investors, among which Tencent, as mentioned, hoping the front-end warehouse model would bring new vitality to the challenging fresh food industry. As early as 2018, Alibaba, Meituan, and JD.com all intended to acquire MissFresh but stopped pursuing this as they considered the firm’s valuation too high. [7]
MissFresh tried various diversification strategies to grow its business, including a Pinduoduo-like social commerce platform. Seeing Luckin's initial success, MissFresh bought hundreds of commercial coffee machines worth tens of millions of yuan when it planned to sell coffee and meals. Most of these machines were left to gather dust in warehouses. Indeed, these experiments were failures, and MissFresh lost RMB 3 billion in 2019, almost all the funding it had raised in the previous year. [3]
After receiving RMB 10 billion in funding since its 2014 founding, investors started losing interest in MissFresh. After 2018, even Tencent scaled down its investments in the company, never leading another funding round again. The only reason Tencent continued to make smaller investments probably was that MissFresh would lose other investors' trust without Tencent’s endorsement. MissFresh had approached JD.com for funding, but the company kindly declined because it didn’t see improving unit economics. [3]
By 2019, US venture capitalists had stopped investing in MissFresh. The company was running out of cash, but the sudden explosion of business during the pandemic temporarily saved it. When many Chinese had to stay in their apartments during strict lockdowns, grocery delivery was one of the few ways to get your hands on food, and it renewed interest among investors in the sector. During Spring Festival 2020, daily orders tripled compared to the previous month, and the AOV rose from RMB 80 to RMB 120. Gross profit increased to 25%, three times the previous year's. [3]
MissFresh got two more investments totalling RMB 5 billion and continued to use most of this funding for customer subsidies. When these stopped, the daily volume of orders dropped again. MissFresh received another funding from Qingdao’s local government on the condition it would move its HQ to that coastal town in Shandong. [3] The same happened when the city of Changshu offered an investment, giving MissFresh two main offices.
With no more private and government investments left to draw from, MissFresh could only go public. While it planned an IPO, MissFresh’s front-end warehouses dropped from around 1,500 to 631 in number. The company claimed that small warehouses of 100 square metres had been merged into bigger 300 square metres ones and that there was no effect on scale. [8] Meanwhile, MissFresh’s main competitor, Dingdong, had expanded to 1,136 warehouses.
After the 2020 funding, MissFresh claimed to have reached nationwide positive cash flow for six months. However, the prospectus and financial reports that were issued later showed a negative cash flow in the first half of that year at RMB -745 million. It wasn’t the first time MissFresh had shown questionable financial reporting. In 2016, MissFresh had claimed to have broken even in Beijing. A journalist later found out that it was due to adjustments in financial calculations; user subsidies and marketing expenses for Beijing had been moved to the cost centre of headquarters. [3]
MissFresh kept intentionally downplaying market risks, omitting RMB 830 million in short-term loans and RMB 248 million in convertible bonds, together 30% of their liabilities, from the first version of their prospectus. The U.S. Securities and Exchange Commission (SEC) forced MissFresh to make these risks more visible in later versions. [3]
Going ‘Smart’
While it was preparing for its IPO, MissFresh tried new initiatives like ‘intelligent fresh markets’ and ‘retail cloud’, the investments of which put even more pressure on its operating results.
Still from MissFresh promo clip.
The concept of Intelligent Fresh Market was based on renting vegetable markets from governments and subcontracting spaces to vendors with a profit margin. It also provided the vendors with paid SaaS tools for bookkeeping, online marketing, CRM and business planning (far too fancy a service for the average vegetable seller if you ask us). At the markets, it would also set up online shops with pick-up of goods featuring FMCG that were usually only available in supermarkets. Vendors would get a share of the revenue of such products. MissFresh claimed to operate 34 fresh markets in 15 cities in June 2021, at a net margin of 10-15%. [5]
Retail Cloud, launched in cooperation with Tencent Smart Retail, involved packaging MissFresh’s technical capabilities as SaaS for supermarket chains in lower-tier cities. It was meant to help SME supermarkets create and operate online stores and campaigns in mini-programs and optimise their logistics. MissFresh would get a share of the incremental profit. At the time of the IPO, MissFresh was testing the service with just one client. [5] One year later, by mid-2022, MissFresh claimed the intelligent fresh market business had expanded to 20 cities and that 50 retail companies were using its retail cloud. [36]
MissFresh also set up a next-day delivery service with JD.com and made half its products available on JD.com and JD Daojia. But all these new initiatives only made up 5% of revenue by Q3 2021. [3] They did look fancy in the IPO PR though … and the company said it would invest 40% of the funds raised in intelligent fresh markets and retail cloud to help it penetrate lower-tier cities.
Let’s leave MissFresh for a moment and look at its biggest competitor. Until 2018, MissFresh had acquired most of its customers through social channels, and because of the lack of direct competitors, the acquisition costs weren't high. But that all changed with the arrival of Dingdong Maicai that same year.
Dingdong Maicai
Another party in the front-end warehouse market is Dingdong Maicai (叮咚买菜), founded in 2017 by wholesale veteran Liang Changlin. Dingdong burned RMB 300 million in savings from Liang's previous entrepreneurial ventures in its first year. One hundred fifty investment institutions rejected funding requests, and if it weren't for RMB 45 million in funding in May 2018, Dingdong would not have seen the end of that year. Further funding rounds proved much easier, and Liang even got offered more capital than he needed. [8]
Still from Dingdong promo clip.
While MissFresh expanded around the country, Dingdong initially stuck to Shanghai and the surrounding area. It also stuck to its product category of fresh fruit and vegetables while competitors expanded their categories to increase average order value. [8] Instead, Dingdong emphasised economies of scale and only offered around 2,000 SKUs, half of what MissFresh offered for instant delivery. It focused more on fresh products and kitchen consumables, serving as a substitute for fresh markets. It gave Dingdong a higher single-product concentration and better bargaining power with suppliers than most other retailers. [5]
Like MissFresh, Dingdong delivered within a radius of 3 kilometres within 30 minutes. But unlike MissFresh, it provided free delivery without a minimum spend requirement. In 2021, customers typically purchased fewer items per order than with MissFresh but bought more frequently. Dingdong's AOV was RMB 56.8 (MissFresh had RMB 96.1), and the average daily order per warehouse reached 913, more than twice that of MissFresh. [5]
Still from Dingdong promo clip.
While MissFresh focused on fruit until early 2019, Dingdong focused on fresh vegetables. Dingdong also established self-operated farms and pork processing factories to reduce supply chain costs. It launched 600 private labels in 2021, 60% of which are self-processed or produced, which could increase profit margins by 10-30% compared to brand suppliers. Meanwhile, MissFresh's attempts at private labels were mostly still produced by suppliers and had much lower cost reductions. [3]
Dingdong customers were, on average, slightly older (40-60 years) than those of MissFresh (30-50 years). The core of its users consisted of urban households aged 26-45. Customers are broken down into the following segments: [9]
high-quality consumption focussed, aged 30-50, accounting for 60%;
health focussed silver-haired group over 50 years old, accounting for 20%;
young people under 30 years old with emerging demand, accounting for 20%.
Dingdong invested in infrastructure: 1,000 front-end warehouses, 60 city sorting centres, ten food R&D facilities and processing plants, focussing on quality control and R&D. [8] In Q1 2021, Dingdong processed 900,000 orders per day and generated RMB 1.5 billion in monthly sales in 27 cities. [10]
Unlike MissFresh, which used some franchised warehouses that paid a RMB 60,000 deposit, Dingdong was fully self-operated. Thus, Dingdong had higher costs, and MissFresh could initially expand faster, but MissFresh had less control over quality. It tried to fight Dingdong in Dingdong's home base of Shanghai, but in Q1 2020, Dingdong overtook its competitor in daily orders, with better performance efficiency, repurchase rate and other KPIs to boot. [3]
Like MissFresh, Dingdong benefited from the 2020 pandemic and demand for online grocery delivery, surpassing MissFresh and even Alibaba’s Hema (Freshippo). [8] In the first half of 2020, Dingdong's orders in Shanghai increased by 300% and its number of users by 200%. In Q3 2020, Dingdong expanded to Beijing, Tianjin and other cities. Xiamen and Chongqing followed, and more than 20 second and third-tier cities. However, most warehouses outside first-tier cities could not meet their KPIs. The average daily orders lay between 300 and 600, and warehouses required substantial subsidising to maintain volume. [8]
According to their IPO prospectuses, in 2020, MissFresh and Dingdong had average order values of RMB 95 and 57, respectively. MissFresh was operating in 16 cities and Dingdong in 36. MissFresh held 23% market share and Dingdong 39%. [5]
Dingdong app.
Two IPOs and the fall of MissFresh
At the end of 2020, MissFresh released its listing plan, closely followed by Dingdong. Both companies remained unprofitable and needed new funding to support their development ... or even survival. Both planned to go public on June 29 2021, and both hoped to raise $300 million.
On June 24th, MissFresh suddenly broke off its IPO roadshow and announced it would go public the next day at a valuation of $3.2 billion, a move intended to snipe Dingdong. MissFresh raised $273 million on Nasdaq with a share price of $13. But on the day of the listing, shares plummeted 26% to $9.60, making some funds that had planned to invest in Dingdong pull out. Dingdong reacted by lowering its funding target to $80 million. [3] Its shares sold at $23.50, valuing it at around $5 billion.
After their IPOs, MissFresh and Dingdong had cash to survive for 1.5 and 1 year, respectively. The two needed to quickly become profitable or find extra funding before money would run out again. With their poorly performing IPOs, finding new investors would prove difficult. [11]
Please note: the rest of this article is only available to paid subscribers.