以下是张北平的演讲实录（略有删减）
新零售很重要的元素要数字化，要用数据“武装”自己。中国新零售现状目前是“红海”。虽然新零售由马云提出包括京东腾讯都提出相关的概念，小米没有提新零售，但是小米做了相关的实践。
站在整个行业或者产业链角度，新零售并不是一家企业能做得了，不论是上游还是下游还是甲方乙方，都要站在整个行业生态链的角度去看，而小米之家在过去几年连续被入选新零售的新物种。
新零售要从技术角度出去思考，技术的重要性对于新零售业务有很大的驱动力。不能单纯脱离业务谈技术，技术能够助力新零售变革，特别是大数据、物联网、人工智能，通过数字化方式帮助传统企业进升级比如人工智能+新零售目前是新风口。
新零售本质上还是围绕“人货场”，所以怎样把线下的零售所有业务、所有流程数字化，也是过去电商为什么能够在中国风起云涌，迅猛发展的原因，因为它有红利、数字化的场景。对于线下来讲，线下如何与线上融合，这是新零售最重要的课题。借助技术和手段，AI、人工智能、物联网，能够把“人货场”真正结合起来。
1、小米新零售的模式
现在小米的模式叫“铁人”三项——硬件、新零售、互联网平台。小米现在在做多样化，但它的多样化并不是存在的而是想整合，相互辅助。比如顶层围绕手机和3C数码展开的硬件，这也是区别于腾讯、阿里、京东，小米是以产品为导向的企业，是重产品的企业。
新零售包括小米之家和小米商城。小米与传统零售品牌商不一样，从路径来看，零售商是从线下拥抱电商做线上然后再做融合，而包括小米和其他互联网公司更多是原来做线上。
小米最早成立之初是有电商平台——小米商城。在过去三到五年间开始做线下发力、渠道下沉，未来还有很长的路要走。
2、小米之家的新零售系统
首先，产品如何交付到消费者手中。一是产品和产品相关的服务；第二，供应链，如何把产品交付到渠道不管是线上还是线下；第三，与顾客的交互，不论是在线下门店还是在电商与用户的交互和体验。
围绕这三方面，小米之家要有自己的收银系统、基于货品的管理包括促销、优惠活动、商城、会员等。现在已经基本上打通了小米商城和会员体系。进销存是另外一个维度，包括后台的系统和供应链体系也有打通。
3、移动收银、商业智能、无人零售、大数据应用、视觉识别等在小米之家的应用
第一、移动收银
小米提到新零售，很重要的是如何提高效率。小米之家有很大的用户量，与其它不同，小米之家的品类会较大，基本上能保证顾客每个月去到店里要能保证有新品、新的惊喜，所以，用户黏性比较高，有大量的IOT的数据。
这其中小米之家很看重收银效率，如果效率低，用户体验一定糟糕，糟糕的话转化率就很低。所以，我们担心的并不是客流量的问题，而是进店以后如何把客流量转化为销售。所以，通过移动端设备加上，后台大数据的相关算法，会做相关的分析，提升效率。
第二、商业智能（BI）
小米之家拥有专业的数据团队，同时它有更全域的数据，而且是是基于场景和需求发出，每周都在迭代，风雨无阻。结合小米云平台的大数据能力，做更多的分析。
第三、无人零售
如果去深圳还是去旗舰店，如果店里面没有货，可以通过扫码的方式用物流去发货。现在自主研发一个平台，做了许多的尝试，基本上能保证线上的体验和相关产品品类，在大屏上都是能够浏览到，这是对于用户和商品的触达。
同时小米之家也做了部分无人零售的尝试，有人脸识别和RFID，顾客进店的购物，包括离店的整个闭环。RFID是一个非常成熟的方案，但是目前也不能简单用它，还是要结合自身的业务场景，应该不是完全无人。
小米之家最主要的点，如果真的是无人，他不像很小的便利店，小米的产品还是比较重的。包括它的品类比较多，所以需要有人对消费者进行解释，只有了解他才有购买的欲望。
第四、夯实基础
大数据支撑，结合开源的平台和互联网的方式。通过做持续交互、运营优化，保证后台的高可用性、弹性扩展、松耦合，支持业务多元化的发展。夯实基础还有一些方面，全新的UI设计，速度缩短了200毫秒，PHP升级到7.0，性能预计提升1倍。
同时开发运维一体化，保证运维能够与开发结合起来。问题一旦出现后，能够快速给到开发，能够快速迭代去修复BUG，保证产品不断完善。
第五、大数据提升销量
所有市场都是服务于业务比如开店选址，小米之家现在处于迅猛发展的阶段，所以拓展很重要。一是开店，一是卖货。如何开好店，要保证店是一个好店。经营店的能力是内功，前提是选好店。结合内部数据去做深入的分析包括智能商品推荐、个性化营销等等
个性化营销或者是精准营销是新零售很好打法的一种体现，怎么做好营销而不是盲目做顾客进店之后的营销结合的还是大数据。
结合小米用户的画像，做好用户分群。通过用户画像和触达分析出：第一，消费者是谁；第二，怎么样把产品或者营销信息，触达到消费者。小米之家通过地理围栏和基于用户画像筛群把信息触达到消费者。
另外，营销活动推广，通过年龄、地域、兴趣、行为，圈定了目标消费客群，通过短信营销的方式。
MIUI系统，我们通过小米的企业的数据画像，再通过APP的方式做Push。你在小米之家购买过，有过消费记录的客户，基于他进行消费交易数据、订单数据做定向推送。
第六、人工智能视觉识别
AI就是视觉识别、机器识别，现在是风口。万变不离其宗基于过程和场景数据。小米之家内部有交易数据但是是远远不够的。所以怎么结合技术，获取到场景的数据。比如识别场景，比如小米优品的店，生活方式的各个区域有摄像头然后去识别。
虽然看似很简单但是做起来也有存在一些问题，比如很多人脸识别只是基于人脸但是对于货品没有办法通过视觉抓取或者单纯靠视觉。顾客进店有这么多的产品顾客选了什么样的产品，拿起放下哪个产品，我们把人的数据和货的数据匹配在一起，这是非常难的。
新零售无论是不是取代传统零售的升级，这不是一个终结，而是新的开始，对于整个行业是一个很好的引领。
翻译：
On July 21, 2018 China Enterprise Internet Summer Summit, “Reconstructing Human Goods Yard · Detonating New retail” hosted by Cui Niuhui, was held in Nader Freedom Hotel in Hangzhou. Zhang Beiping, CIO of Xiaomi Home, shared and communicated with the guests on “Exploration and practice of Xiaomi Home new retail”.
The following is a transcript of Zhang Peiping’s speech (with slight abridgement).
An important element of new retail is digital, armed with data. The current situation of China’s new retail is “red sea”. Although the new retail was proposed by Jack Ma, including JD.com and Tencent all put forward the related concept, Xiaomi did not mention the new retail, but it did the relevant practice.
Standing in the perspective of the whole industry or industrial chain, new retail is not a single enterprise can do, whether upstream or downstream, Party A and Party B, should stand in the perspective of the whole industry ecological chain, and millet home has been selected as a new species of new retail in the past few years.
New retail should think from the perspective of technology. The importance of technology has a great driving force for new retail business. Technology can not be separated from business. Technology can help the transformation of new retail, especially big data, Internet of things and artificial intelligence. Through digital means, traditional enterprises can be upgraded, for example, artificial intelligence + new retail is a new trend.
New retail essentially revolves around “people and goods yard”, so how to digitize all the offline retail businesses and all the processes is also the reason why e-commerce can surge and develop rapidly in China in the past, because it has the dividend and digital scene. For offline, how to integrate offline with online is the most important subject of new retail. With the help of technology and means, AI, artificial intelligence, the Internet of things, can truly combine the “human goods yard”.
1. Xiaomi’s new retail model
Now Xiaomi’s model is called the “triathlon” — hardware, new retail, Internet platform. Xiaomi is now doing diversification, but its diversification does not exist, but wants to integrate and assist each other. For example, the top layer focuses on mobile phones and 3C digital hardware, which is different from Tencent, Ali and JD.com. Xiaomi is a product-oriented and product-focused enterprise.
New retail includes Mi Home and Mi Mall. Xiaomi is different from traditional retail brands. From the perspective of path, retailers embrace e-commerce from offline to online and then do integration, while Xiaomi and other Internet companies are more original online.
Xiaomi was first founded as an e-commerce platform – Mi Mall. In the past three to five years started to do line distribution force, channel sinking, there is a long way to go in the future.
2. Mi Home’s new retail system
First, how the product is delivered to the consumer. First, products and product-related services; Second, the supply chain, how to deliver the product to the channel whether it’s online or offline; Third, the interaction with customers, whether in offline stores or in e-commerce interaction and experience with users.
Centering on these three aspects, Mi Home should have its own cash register system and management based on goods, including promotion, preferential activities, shopping mall, members and so on. Now we have basically got through the small rice mall and the membership system. Invoicing is another dimension, including the backstage system and the supply chain system.
3. Mobile cash register, business intelligence, unmanned retail, big data application, visual identification and other applications in Mi Home
First, mobile cash register
Xiaomi mentioned new retail, it is important to improve efficiency. Millet home has a large number of users, different from other, millet home category will be larger, basically can ensure that customers go to the store every month to ensure that there are new products, new surprises, so, user viscosity is relatively high, there are a large number of IOT data.
Among them, Mi Home values the efficiency of the cashier. If the efficiency is low, the user experience must be bad, and the conversion rate is very low if it is bad. So, we’re not worried about foot traffic, but how to convert foot traffic into sales once we’re in the store. Therefore, relevant analysis will be done through mobile terminal devices and relevant algorithms of background big data to improve efficiency.
Second, Business Intelligence (BI)
Xiaomi Home has a professional data team, and it has more comprehensive data, and is based on the scenario and requirements of the issue, iterative every week, rain or shine. Combined with the big data capability of Xiaomi Cloud platform, do more analysis.
Third, unmanned retail
If you go to Shenzhen or go to the flagship store, if there is no goods in the store, you can use logistics to deliver goods by scanning code. Now I have independently developed a platform and made many attempts. Basically, I can ensure that the online experience and related product categories can be browsed on the large screen, which is a touch for users and products.
At the same time, Mi Home has also made some attempts of unmanned retail, including facial recognition and RFID, and the whole closed-loop of customers’ shopping in and out of the store. RFID is a very mature scheme, but it can not be used simply at present, or should be combined with their own business scenarios, should not be completely unmanned.
The main point of millet home, if it is really no one, it is not like a small convenience store, Xiaomi products are still relatively heavy. There are many categories including it, so someone needs to explain to consumers. Only by understanding them can they have the desire to buy.
Fourth, strengthen the foundation
Big data support, combined with the open source platform and the Internet way. Through continuous interaction and operation optimization, high availability, elastic expansion and loose coupling of the background are ensured to support the development of diversified services. A new UI design, 200 milliseconds faster, PHP upgraded to 7.0, and performance expected to double.
At the same time, the integration of operation and maintenance is developed to ensure that operation and maintenance can be combined with development. Once the problem occurs, it can be quickly developed and quickly iterated to fix the BUG to ensure the continuous improvement of the product.
Fifth, big data to increase sales
All markets serve the business, such as store location. Xiaomi Home is now in a stage of rapid development, so expansion is very important. One is to open a shop, the other is to sell goods. How to open a good store, to ensure that the store is a good store. The ability to operate the store is internal work, the premise is to choose a good store. Combine internal data to do in-depth analysis including smart product recommendations, personalized marketing, etc
Personalized marketing or precision marketing is a good embodiment of the new retail, how to do a good job of marketing rather than blindly do the marketing after customers enter the store or big data.
Combined with the portrait of Xiaomi users, do a good job in user segmentation. Through the user portrait and touch analysis: first, who is the consumer; Second, how to reach consumers with products or marketing messages. Mi Home reaches consumers through geo-fencing and a screen group based on user profiles.
In addition, marketing activities promotion, through age, region, interest, behavior, delimit the target consumer group, through the way of SMS marketing.
In the MIUI system, we Push through the data portrait of Xiaomi’s enterprise and then through the APP. You have purchased in Millet home, have a consumption record of the customer, based on his consumption transaction data, order data do directional push.
Sixth, artificial intelligence visual recognition
AI is visual recognition, machine recognition, and now tuyere. All changes depend on process and scenario data. There is transaction data inside Mi Home, but it is far from enough. So how to combine technology to get the data of the scene. For example, the identification scene, such as Xiaomi Youpin store, the lifestyle of various areas have cameras and then identify.
Although it seems very simple but there are also some problems to do, such as a lot of face recognition is just based on the face but for the goods there is no way through visual grasp or simply rely on vision. There are so many products that customers come into the store and what products they pick up and what products they put down, and we match the people data with the goods data, and it’s very difficult.
Whether new retail is an upgrade to replace traditional retail or not, it is not an end, but a new beginning, which is a good lead for the whole industry.
本文由数字化转型网（www.szhzxw.cn）转载而成，来源：牛透社；编辑/翻译：数字化转型网宁檬树；审核：数字化转型网默然。
免责声明: 本网站(https://www.szhzxw.cn/)内容主要来自原创、合作媒体供稿和第三方投稿，凡在本网站出现的信息，均仅供参考。本网站将尽力确保所提供信息的准确性及可靠性，但不保证有关资料的准确性及可靠性，读者在使用前请进一步核实，并对任何自主决定的行为负责。本网站对有关资料所引致的错误、不确或遗漏，概不负任何法律责任。
本网站刊载的所有内容(包括但不仅限文字、图片、LOGO、音频、视频、软件、程序等) 版权归原作者所有。任何单位或个人认为本网站中的内容可能涉嫌侵犯其知识产权或存在不实内容时，请及时通知本站，予以删除。
