equity research·industry research·e-commerce e-commerce industry research 1 / 8 东吴证券（香港） 请务必阅读正文之后的免责声明部分 [table_main] on-demand retail industry report 3: market outlook and competition deduction from a channel comparison perspective outperform (maintain) investment thesis ◼ why do we want investors to pay more attention to on-demand retail? on-demand retail is a retail mode based on on-demand delivery systems, which matches the supply and demand of local goods online and provides hourly home delivery service. thanks to relatively clear demand & business model and multi-party efforts, on-demand retail enjoys rapid growth and is currently in an important development phase. with slowing growth of overall online retail, we see great opportunities in on-demand retail industry. ◼ demand: potential of rmb1tn in 2025. the motivation of on-demand retail consumption upgrades from product dimension of traditional e-commerce to "product time space" dimension. penetration potential remains for categories whose demand is not fully satisfied by less time-efficient traditional e-commerce or whose offline scenario does not have an absolute advantage. as per our calculation, the ceiling of on-demand retail users is close to 550-640 million, higher than food delivery but lower than integrated e-commerce. we expect the market size to reach rmb1018.2bn in 2025, and the penetration rate of offline channels to be about 3.3%, with a cagr of 31% in 2022-2025. specifically, we expect the penetration rates of categories with short expiration & low requirements on commodity form, other high-frequency & low-value categories, low-frequency & large price band span or policy control, and other categories to be 15%/13%~14%/8%/5%. ◼ supply: trends of platform mode and supply diversification. the early explosive growth of on-demand retail platforms is lower than other platforms, and we believe the growth difference is related to the extent to which platforms optimize the matching efficiency of supply and demand. the lack of effective supply restricts the current industry development, and future increase in supply will drive up penetration rate in offline channels. 1) supply mode: self-supply mode mainly meets the demand of fresh food related categories, while platform mode becomes the main force to expand industry supply. platform mode's channel characteristics of high turnover and regional precision marketing should promote brands to increase investment (we expect its market size to grow from rmb227.6bn in 2021 to rmb709.7bn in 2025, taking 70% of the market share in 2025 vs. 65% in 2021). 2) growth path: considering the low concentration of offline retail industry and the strategic tendency of leading retailers, we believe the contribution of leading retailers is relatively limited while small- and medium-sized stores with dense distribution, large number and digital possibility, as well as vertical stores are likely to become the main supply increment on platforms. in addition, "lightning warehouse" with both user traffic and efficiency advantages develop rapidly, and we expect its supply growth to accelerate with more successful merchant cases. ◼ competition deduction: industry-wide benefits in the near term with more focus on supply and efficiency in the long term. 1) traffic acquisition stage: advantages in brand recognition help instashopping to effectively convert meituan users (instashopping/ meituan users = 41% vs. jddj/jd users = 11% in 2021), resulting in lower customer acquisition cost. due to differentiated supply, instashopping and jddj share different positioning, but we expect them to both enjoy the industry-wide benefits, given low user penetration rate. 2) traffic retention stage: a) instashopping has a better supply expansibility. instashopping and jddj are both breaking through their comfort zone of supply, and we believe instashopping enjoys better supply expansibility. diversification of meituan's traffic sources is also an important incentive for merchants. we estimate instashopping and jddj’s gmv to reach rmb438/140.1bn in 2025, accounting for 62%/20% of the on-demand retail market respectively. b) instashopping has obvious advantages in efficiency. as per our estimate, jddj’s average revenue per order is higher (higher aov), but thanks to the positive externality of food delivery, instashopping has a better ue, showing its efficiency advantages. considering their existing supply structure and development strategy, as well as instashopping's greater elasticity in aov and monetization rate, we expect instashopping to first achieve op break-even and enter a virtuous circle after business self-sufficiency, with optimized fees and higher user stickiness. we expect instashopping to achieve ue of rmb1.58 (1.6% of aov) in 2025, contributing rmb6.9bn to total operating profit, and jddj to achieve pro