The station is to "let people hear the call of artillery fire," the face of tide construction in full swing in Taiwan, only to explain "how to use the data" problem, have the need to further answer "to how data", "data how to do it "questions.
Why do we need intelligent data model
In Taiwan, 2019 one of the most fiery words.
If the data in the table compared to a modern aircraft carrier, enterprise data management, no doubt now alone aboard the aircraft carrier or a Big Mac, no convoy, is not equipped with cluster operations, but no *** nuclear submarines.
The question now is, we built the "central power stations", but found the lack of release of its huge production capacity "light bulb."
We had plenty of single cluster of thousands of servers-scale infrastructure operators force, just allow the reports to run faster or eliminate data silos? This is undoubtedly the biggest waste of resources to count the massive force, calling the market can count these huge forces released data applications.
What kind of system to match the huge calculation power? What kind of system can truly effectively to solve real business problems? Intelligent data model!
What is the Intelligent Data Model
What is intelligent data model? He and traditional information systems What is the difference? Smart and reflected in what areas?
It should explain the difference under the "information system" and "intelligent systems", the nature of "information system" is editing the database, if a system is the core of decision-making and relies a lot on artificial human interaction to complete the task, then that information systems . The "intelligent system" is to rely on highly automated machines to complete the "data cleansing - to locate the problem - business decisions" and a series of operations to mission as an input to the processing result is output.
According to this standard, commercially-shaped form ××× intelligent smart systems are just borrow the name of fish.
Intelligence of the intelligent system illustrated in Figure 17, L0 to L4 means the higher degree of intelligence.
Four stages of the development of the standard of enterprise management, enterprise information technology projects must complete the upgrading L0 to L4.
Intelligent Data Model - Smart tune replenishment system
Construction of enterprise information management business is actually an upgrade projection. Of course, I say that it is quite abstract.
Take the most basic commodity operations tune up goods sectors, for example.
1, L0 stages: in the early stages of business, area store replenishment, transfer goods between stores, not a very serious problem, one column a few tables, can get a little thought, you can only achieve standardized management ;
2、L1～2阶段：随着规模的扩大，当门店达到上百家，这时候就必须组建商品部门，去协调上百家门店之间的商品短缺和区域间不平衡的情况，这里就需要流程化管理。半自动化或者自动化管理系统可以辅助商品运营人员形成企业独有的运营风格和策略；
3、L3～4阶段：数据的传输效率会随着节点（人员）的增加边际递减，随着规模的进一步扩大，想依靠加人头来管好线下的整盘货，就变成了一个人员臃肿、效率低下、成效难以衡量的问题，期间伴随着高昂的人员培训成本和核心员工离职的风险。智能化系统也就应运而生。
我们近期服务的某集团正处于流程化管理向自动化管理过渡阶段，商品运营部门人员多达20余人，这20余人平均每周需要投入四天时间去处理商品的补货、调货数据。
这20余人训练有素，需要人肉去判断500家门店商品的盈亏状态，在供大于求的时候优先满足哪种类型的门店，应该满足多少？供小于求的时候？一个熟练的投放专员在每次补调货的时候需要同时考虑十多个衡量指标。
现在某集团预备在未来一年内开启×××模式，门店预计扩张到4000家，一个熟练的商品专员平均培养周期最少2年。
门店扩大十倍，商品专员也相应扩大十倍？
智能调补货系统，集采了天气、区域、位置等外部数据，结合行业先进的调补货经验，上线后预计可达到：
1、采用深度学习算法，原先需要十多人协作完成的补货数据，机器仅仅在几分钟内就完成了全部补货过程，即使千余家门店的调补货计算量也不在话下，无需辛苦招人培训，还要随时警惕竞争对手来挖角；
2、原本需要整个商品部反复拉扯的补货问题，系统自动按照毛利最优解决方案，完成了全部的调度工作，预计可将商品平均周售罄率提升到60%～70%，双周售罄率稳定提升到80%以上，区域间调拨次数降低30%以上（18年某集团仅补货物流成本可达250万）；
3、业务人员的脑力被极大释放，珍贵的核心员工只需管理好算法模型的优化方向和数据补充，有更多时间去思考商品本身的运营策略，商品部也由一个成本部门升级为利润部门。
巨舰还是得配利炮
The wave of cloud computing reduces the cost of hardware acquisition, directly spawned ecological prosperity in Taiwan today.
Today, China's small and medium enterprises can build their own low-cost data in the table, have the opportunity to correct the data from the source to start the construction of direction, but the prevailing demand and business development in Taiwan, is bound to Forced upgrading business reception.
Taiwan under the strategy, all aspects of the current reconstruction of enterprises in fact are worth the upgrade again.