A summary of the background of the case
As the first Internet apparel brand, Handu Yishe has Korean, European and American, and Oriental brand clusters, with an annual GMV of nearly 1.8 billion. Handu Yishe E-commerce Group, which is also one of the largest Internet brand ecological operation groups in China, has established a product competitive advantage of "many styles, fast updates and high cost performance" by virtue of its unique flexible supply chain management model - "IOSSP", and strives to build a fashion brand incubation platform with global influence, and is expected to complete the layout of more than 50 brand clusters based on clothing categories in 2020 and achieve a transaction scale of more than 10 billion.
Challenge:How does Handu Yishe use AI & BI to take the efficiency of flexible supply chain to the next level?
Scheme:Zhiyi Technology "AI BI Intelligent Commodity Decision-making System"
效益：Handu Yishe applies the "AI BI intelligent commodity decision-making system" in the IOSSP flexible supply chain, and the annual inventory turnover rate can reach more than 6 times, and the quarterly sell-out rate can reach 90%~95%. The selection efficiency is increased by 30%, and the product selection planning decision based on the "labeled" data source is used to control 60% of the subjective risk.
Application and implementation of the scene of Handu Yishe
Handu Yishe has ranked first in the industry comprehensive ranking for many years on major domestic e-commerce platforms. Women's clothing has won the "triple crown" of the whole year, Double 11 and Double 12 in the history of Tmall. Relying on the unique flexible supply chain, we can flexibly deploy marketing planning, product planning and supplier production to achieve a competitive strategy of "more money and less quantity, production and sales". In order to quickly respond to market demand and support operations, Handu has to open 500-1000 models every day, but most of the time the product planning and selection decision-making process is mainly based on manual selection and subjective judgment. Handu, which has "Internet genes", always believes that the application of "science and technology" can have a more efficient model. In 2014, Handu began to optimize and adjust the organizational structure of the product production center, put forward higher requirements for its own quality, and maximize the potential of more than 240 existing high-quality suppliers, and set up a special technical "selection forest" internally, focusing on fashion picture data collection and analysis, capturing market trends, and carrying out data transformation in all links of the supply chain, assisting decision-making through objective analysis, and reducing the risk of subjective factors.
As one of the pioneers in the industry to apply the enterprise middle platform system, after years of deep cultivation in the field of technology and data application, Handu has further improved the single product operation system (IOSSP) with the product group as the core, and the whole process of data-based and refined operation management system, established the competitive advantage of "many styles, fast updates, and high cost performance", and effectively solved the most headache inventory problem in the clothing industry, which can ensure that customers can provide more product choices with high cost performance.
However, on the other side of the remarkable results of supply chain management, Handu is facing the continuous increase in technical investment and labor costs. More than 500-1,000 models are opened every day, which means that designers need a lot of design inspiration and market data to assist in decision-making. In order to support the needs of the product design department, the "selection forest" should be collected from multi-platform fashion sources, including overseas platforms, paid platforms, large e-commerce platforms, etc. Image source collection is just the beginning, and problems come one after another: 1. The data sources are huge, and about 10,000 pictures every day cannot be classified into the warehouse because there is no detailed label data (including category, collar type, sleeve type, skirt length, pattern, etc.); 2. The labor cost is too high, and the design team has to spend a lot of energy to screen out the styles suitable for each product line from the millions of orders of magnitude source data, which is time-consuming and labor-intensive; 3. The cost of sample clothing is uncontrollable, and the opening of a large amount of money every day means a huge cost of sample clothing, how to control it? These problems have allowed Han Du to gradually perceive the pressure in the process of quickly analyzing the trend and capturing the market trend. In 2018, the Handu "Selected Forest" project team began to look for external mature products and teams, hoping to alleviate the current bottleneck in the form of cooperation.
Handu compared a number of mature big data products and AI companies on the market, and their functions and services could not fully solve the problems they faced. In March 2019, Zhiyi Technology's AI trend products were officially unveiled at the CHIC exhibition, and Handu began to contact Zhiyi, after inspecting the AI trend products, Handu was pleasantly surprised to find that it not only highly matched in data source collection, but also had a relatively complete label system, through machine deep learning, in terms of color, silhouette, process, accessories, patterns, fabrics, these attributes, AI artificial intelligence map recognition accuracy has reached more than 80%. And compared with competing products, the key attributes of the above clothing determine the designer's judgment benchmark for style, reflecting that Zhiyi's understanding of the business also exceeds that of competing products.
After in-depth communication with Handu, Zhiyi found that AI trend standard products can solve the "label system problem", but cannot completely solve the three problems mentioned above: data marking, selected styles and sample cost. To this end, Zhiyi has tailored an intelligent commodity decision-making solution for it, fully supporting the application and practice of Handu in commodity scenarios, design scenarios, ordering scenarios, human efficiency scenarios, and review scenarios, comprehensively upgrading the technical application and artificial intelligence landing of the digital middle platform, and realizing full automation in information flow. At present, the system has been enabled, and the design department of Handu, which has a strong demand for image sources, can receive the fashion pictures that have been identified and labeled every day, with an accuracy rate of up to 87%, which has basically solved the problem of picture labeling; And through the in-depth understanding and targeted customization of the Handu style, the pictures obtained by the designer have been intelligently screened and deduplicated, which greatly improves the accuracy and utilization rate of the style; At the same time, it also makes the opening more accurate and saves a lot of sample costs.
Not only that, after the three pain points of the enterprise have been effectively alleviated, the system's commodity digital assets are supported by the algorithm engine of the intelligent middle platform, extending more marginal effects: 1. Display the purchase, sale and inventory status of the commodity in a visual way, and match the mobile phone terminal, update and synchronize in real time, so that designers can obtain information quickly and accurately, and at the same time, the feedback to the market is also keen, and the adjustment of the design direction is more timely; 2. It allows the enterprise decision-making level to comprehensively control the progress and effect of nearly 300 development teams, and can intuitively manage and deploy human efficiency, which further expands the depth and breadth of data applications.
Handu's internal control department compared competing products in terms of cost and implementation, and also compared labor. It sounds like the lofty "AI artificial intelligence & BI business intelligence" can not only improve efficiency after landing, but also be flexible in the way, and charge in depth according to customized needs, so that the overall management and development costs can be greatly optimized. At the same time, Zhiyi provides 7*24.hour after-sales service, and there is a special person responsible for follow-up response, which also reassures them a lot.
韩都衣舍在IOSSP柔性供应链中应用知衣的智能商品决策系统后，年度库存周转率可以达到6次以上，当季售罄率可以达到90%~95%。并且选款效率提升30%，以数据分析为基础的选品企划决策，把控60%主观风险。整个产品端反应更灵敏，产品在上新15天后，即按照数据“爆、旺、平、滞”四类。不同级别的产品，产品小组根据市场行情进行商品营销策略的确定和实施。
选款森林负责人表示，在知衣“AI + BI 智能商品决策系统”的助力下，让韩都突破瓶颈，在数字化运作上，再上一个台阶，未来希望在更多技术能力应用层面深度合作。
“AI+BI智能商品决策系统“
The whole scene of the design end of the landing clothing enterprise
不仅韩都，97%以上的服装企业认同，为了满足快速多变的市场需求，提升款式研发的速度是关键。从趋势、大流行到爆款，每一个款式背后都是图片和数据，而结构化标签化的数字资产是一切商业智能的基础与开始。知衣科技智能商品决策系统，通过AI人工智能与BI商业智能全面构建企业设计端全场景决策体系。面对企业个性化述求，智能商品中台通过建立算法模型，进行智能款式推荐，匹配智能化呈现工具，为企业实现零成本线上预审。同时，一线人员也能直接参与，多方反馈即时沟通，加快选款决策速度，降低运营成本，缩短选款周期，进而挖掘人员价值，驱动设计创新，提升人均效能，构建快反能力。
Customized style algorithm recommendation,Set the recommendation algorithm according to the customized conditions, compare the market and enterprise data in real time, and give style data analysis.
Online selection and ordering, detailed style description and feedback on the mobile phone, and systematic review and personnel management on the PC.
Intelligent design tools, including intelligent fitting, search by drawing, search by drawing, search by model, search by model function module.
数据化系统复盘，可细化到单款的数据描述，大屏可视化目前的销售情况和客户情况。
Zhiyi Technology——
Expert in explosive style mining and product planning solutions
Zhiyi Technology is a technology company serving the apparel industry with artificial intelligence technology, and is committed to becoming an expert in the apparel industry's explosive style mining and product planning solutions, solving problems such as fashion trend discovery, clothing style recommendation, production and sales balance, etc., and helping many apparel brands and industry businesses to make business decisions based on "data", reduce risks, and improve operational efficiency and market share. AS A NEW GENERATION OF ARTIFICIAL INTELLIGENCE APPLICATION IN THE INDUSTRY, IT HAS RECEIVED INVESTMENT FROM A NUMBER OF WELL-KNOWN INVESTMENT BANKS, AND THE FOUNDER MR. ZHENG ZEYU (FORMER SENIOR ARTIFICIAL INTELLIGENCE ENGINEER OF GOOGLE'S E-COMMERCE DEPARTMENT) AND HIS CORE TEAM TAKE IMAGE PROCESSING, MACHINE LEARNING, AND DATA MINING AS THEIR CORE CAPABILITIES TO HELP THE DEVELOPMENT OF THE INDUSTRY.
I want to know more about it in detail
How does Handu apply artificial intelligence to upgrade supply chain efficiency?
Scan the QR code to add us as a friend and answer one-on-one.
Sweep the advisor online
