According to incomplete statistics, at least 50 players have entered the market, with over 2 billion yuan in funding already poured in. Founded in June 2015, Xiao e Weidian pivoted to unmanned shelf shopping in the second half of last year. At that time, there were few players in the field, and founder Rong Guang predicted, "After a period of time, many people will pay attention to this business model." Rong Guang probably didn't expect that this period would be rapidly compressed by the frantic influx of capital, giants, and players. According to incomplete statistics from Phoenix Tech, at least 50 players have entered the market, with over 2 billion yuan in funding already poured in. On December 19, Alibaba, together with Midea Group, launched "Xiao Mai Gui," officially entering the unmanned shelf field. On the same day, Cheetah Mobile, seemingly unrelated to unmanned shelves, confirmed its entry, with its "Bao Bian Li" operating since early November and having deployed 5,000 points. New retail became a keyword this year, and among all offline formats, unmanned shelves are closest to users, leveraging the natural advantage of the last 100 meters. This holds an irresistible allure for both new entrepreneurial opportunities and giants exploring new offline traffic. Suddenly, the "thousand-shelf war" resembles the thousand-group-buying war, the O2O wave, and the sharing economy trend. These similar histories share a common narrative: a lack of good investment targets, leading to frenzy and irrationality under the bubble. We want to pour several responsible buckets of cold water on this hot trend to cool the industry down, and also make a few irresponsible predictions and judgments. Cold Water 1: Low barriers make it hard to build a moat. Among all current startup projects, unmanned shelves may have the lowest technical barrier. The business model is simple, making it difficult to establish competitive advantages. If we set aside smart cabinets, starting an unmanned shelf project only requires a shelf, some products, and channel negotiations to get going quickly. In Chinese entrepreneurship, how do you quickly gain a leading edge? For unmanned shelf projects, the current approach is to raise funds, quickly seize points, and sign exclusive agreements. The practices of various companies are similar, with little product differentiation, and everyone is charging forward in unison. The environment is basically "you have money, I have money, and he has backing." For entrepreneurs still considering entering, if you lack money, people, and first-mover advantage, stay calm and don't squeeze in. Cold Water 2: Products are mainly standard items with high similarity. Although most unmanned shelf companies claim smart product selection and real-time dynamic updates via backend big data, according to Phoenix Tech's research, the product update rate may only be around 10%. Most products are still daily FMCG standard items, such as water, ham sausages, biscuits, and puffed snacks. Ding Yiding, an investment manager at Panda Capital, has intensively reviewed unmanned shelf projects. He told Phoenix Tech that the SKUs that can be placed on unmanned shelves are very limited, usually around a hundred, and the dispersion of SKUs is not that large. That is, although there may be slight differences between the top product lists preferred by employees of different companies, the differences are not significant, meaning most unmanned shelves have high product similarity. Based on this, many unmanned shelf companies are trying differentiation, such as introducing cold cabinets, hot cabinets, fresh food, and internet-famous snacks. This greatly increases supply chain requirements, and when points are not dense enough, operational costs become high. According to Phoenix Tech's investigation, many small players are already experiencing slow restocking, such as shelves being empty for a week without replenishment, indicating that supply chain and logistics capabilities are completely inadequate. Balancing the number of points and product differentiation is a hurdle that unmanned shelf companies must overcome. Cold Water 3: High product loss rate and the broken window effect. Most companies that publicly disclose their product loss rate claim it is around 5% or even lower. There is an obvious issue: how is this percentage calculated? Different companies may have different interpretations and definitions, making direct horizontal comparison impossible. An industry insider told Phoenix Tech that some unmanned shelf companies may have a loss rate as high as one-third. Rong Guang of Xiao e Weidian thinks of control strategies through scenarios: "If you are in a place with high personnel turnover or an open environment, there will be significant loss. I see many players placing shelves in elevators, hallways, and corridors, which will become a business that challenges human nature. You must have a closed scenario with a fixed group of people for consumption." Cold Water 4: Initial costs are rising. Yan Limin, founder of Guoxiaomei, once calculated: the cost of a single shelf is 300-400 yuan, products about 600 yuan, and BD cost 100 yuan, totaling an initial cost of about 1,000 yuan per point. However, an industry insider said that initial costs are now rising. He recalculated for Phoenix Tech: 1. Equipment cost: A regular shelf costs 300-500 yuan, while cold and hot cabinets cost at least 1,000 yuan each. As competition intensifies, the proportion of points using cold and hot cabinets will increase, so hardware alone costs at least 1,000-1,500 yuan per point, not including smart equipment. 2. BD cost: This includes basic salaries and performance bonuses for BD staff. To compete for points, many players offer cash or rebates to companies. In first-tier cities, the cost to set up a point starts at 1,000 yuan, and many companies' BD costs per point reach 2,000 yuan. 3. Initial product stocking: The display cannot be too sparse; it must be rich, with at least dozens of SKUs and hundreds of items, plus drinks in the fridge, which adds another 1,000 yuan. 4. Logistics: Transporting equipment and products to the company costs at least 200 yuan. According to this list, the initial cost per point starts at 4,000 yuan. Cold Water 5: Deploying ineffective, low-quality points to quickly increase numbers. In communication with some unmanned shelf BD staff, they told Phoenix Tech that competitors' BD teams proactively approach them, offering to buy their discarded points. These are typically points that have been proven to be low-quality or cover few employees, having little positive effect on operations. The only benefit is helping them meet performance targets and quickly increase the number of points. Wen Zhaohui, founder of Qizhi Kaola, discussed their handling of low-quality points: for example, if a certain loss rate is reached, they proactively remove the shelves. Regarding competitors' attitude of "if goods are lost, so be it," Wen Zhaohui refrains from commenting, "But we don't have that much money to lose." Cold Water 6: Profit expectations are not optimistic. The profitability of unmanned shelves themselves may not be significant. Many entrepreneurs tell a story: unmanned shelves can reach office white-collar workers earlier and faster than convenience stores in office buildings. After capturing this traffic, profitability can extend beyond selling products to other areas, such as Focus Media advertising and new product distribution test channels. How can Focus Media advertising work? First, directly advertise on the shelves, but because offices are not commercial scenarios, it may face obstacles from corporate administration. Second, pop-up ads after users scan to pay, which might be more reliable than the first method, but the damage to user experience and the conversion rate of ads remain to be seen. What about as a test channel for new products? On this point, Ding Yiding believes that users don't buy from shelves to try new things; they usually satisfy immediate, impulsive consumption needs. In this case, the key is whether new products can sell on the shelves. He thinks opportunities may exist in categories without strong brand effects, such as salads, fruit cuts, and bread. After pouring so much cold water, here are a few "irresponsible" predictions based on Phoenix Tech's communication with multiple players and investors. Prediction 1: A wave of closures among unmanned shelf companies will occur in the first half of 2018. From the peak of shared bikes to the wave of closures took about a year. In the unmanned shelf field, this process may be compressed further, with closures expected in the first half of 2018. The characteristics of those exiting include: first, few points, especially few high-quality points; second, low operational efficiency; third, lack of supply chain support. Prediction 2: Industry mergers will increase, with resources concentrating in the top two players. Currently, mergers have already occurred in the unmanned shelf field, such as Guoxiaomei with Tomato Convenience, and Xingbianli with 51 Snacks. Next year will see a free-for-all, and more industry mergers will emerge in the "thousand-shelf war." Si Jianghua, co-founder of Xingbianli, once told Phoenix Tech that before New Year's Day or the Spring Festival, the track will quickly consolidate to two or three players. "Whoever can first reach a scale of about 300,000 points will basically hold an absolute advantage." Prediction 3: The imagination space for purely unmanned shelf companies is limited. Companies that only do unmanned shelves may not sustain expansion and attrition wars. More giants are entering, and their advantage lies in the synergy with their core businesses, sharing manpower, logistics, warehouses, and having inherent BD and procurement resources. Wen Zhaohui rejects the label of "office unmanned shelves" and prefers to position his company as a "near-field retail" company, solving users' consumption needs within 100 meters and satisfying them within 3 seconds. Based on this, he and his team think more about what medium to use to deliver goods and products to corresponding scenarios, which could be closed, semi-open, or open. -END-