在全球化进程中，供应链网络日益复杂，企业与企业的竞争业已逐渐演变为供应链与供应链的竞争，在应对高度复杂性和更高端到端可见性等诉求之下，供应链控制塔受到更多关注和应用。接下来将对供应链控制塔出系列文章，基于SCOR模型体系，系统讲述控制塔在不同的分析阶段、使用的不同的技术能力，不同的应用环节和对象、建立控制塔的渠道并结合案例等等这些方面综合讲述供应链控制塔。
本篇文章为开篇，我们从分析方法在不同阶段的体现做展开介绍，因为这其实是我们做任何数字化项目的建设的基础路径。
分析阶段深度与不同的技术场景
1.1 五个分析阶段
在大数据，智能化时代之下，分析工具日益增长的可用性等，将基于数据的分析带到新的阶段。托马斯 达文波特和珍妮 哈里斯在其《分析竞争》一书中（下附书籍链接），引入一个矩阵的思考维度，横轴将分析方法本身分为五个阶段，即描述性分析、诊断性分析、预测性分析、规范性分析和认知性分析五种。纵轴将不同的分析阶段对应到其在应用中的描述性的信息，将某一现象从基本的描述到原因追溯，到预测，再规定，最后到认知层面。
这五种分析方法通常是分阶段实现的，可以解决不同层面不同角度的问题，不过我们不能简单地说哪种分析方法比其他方法更好，无论在哪个层面，其都有对应的应用场景以支持决策。
1.2 举个供应链场景为例子
假设在某一个供应链业务部门，其日常会遇到的问题典型的例子就是供应短缺，满足不了客户需求。我们第一个要问的问题是，发生了什么？这是个描述性的分析，是哪个产品出现了交付不及时的问题；接着，下一个问题是为什么会交付不及时造成短缺？这个分析就涉及诊断性分析了；我们需要往深了查询，是哪颗子物料或者组装流程或者供应商排产比预期节奏晚了吗？当诊断出这个问题后，我们基本就可以判断由于这个问题的产生，后续的影响，接下来这种短缺会持续多久，预计何时会解决，这就到了预测性分析的层面了。然后，我们能做什么，有什么资源可以来改善这个预警到的问题呢，从而尽可能最优地平衡供需？于是，在这种优化措施的干预之下，我们会看到一个“修正”，一个我们期望的供应改善结果，这里称其为规范性分析；最后，达到智能程度的认知性分析，即试图学习到目前的供应链问题，采取的改善措施，这些路径所给予的信息、经验。基于此，可以发展比如基于AI的预测模型，从数据出发，到信息整合和相当程度的洞察。
上面结合一个供应场景的问题简要说明了当我们分析问题时，在企业想快速看到问题到寻找方案解决问题，以及回顾做lesson learned的时候，这其中所遵循的不同分析阶段的基本规律。可以想见，达到人工智能层级的状态，是我们一直说的数字化供应链所能实现的最优局面。而供应链控制塔在不同分析阶段的主要功能也会有所侧重，并且可以在不同的应用场景下将多个分析阶段的功能做融合。
1.3供应链控制塔在各分析阶段中的应用场景
这几个阶段是基于使用的不同的技术能力做区分的，这里简单介绍如下6个典型的控制塔应用可能性：
- 数据实时智能连接: 实施获取数据用于商务过程监控，流程监控等
- 问题分析: 使用预测性、描述性的分析方法使得数据本身从反应型朝着预测型转变
- 影响分析: 理解某个预警信号的影响，从数字化本身的生态系统到公司业务中的供应链
- 场景建模: 对于不同的场景，提供合适的模拟模型做仿真，给定某个输入后预测可能的响应情况
- 协同响应: 在整个生态系统中互相连接，互相协作
- AI智能: 通过AI/机器学习等技术实现高度自动化
那么，从供应链控制塔出发，在不同的分析阶段之下，数字供应链所涉及的对应的技术趋势有哪些？下篇文章我们做展开介绍。
翻译：
In the process of globalization, supply chain network is becoming increasingly complex, and the competition between enterprises has gradually evolved into the competition between supply chain and supply chain. In response to the demands of high complexity and more high-end to end visibility, supply chain control tower has received more attention and application. The basic definition and coverage of supply chain control tower have been briefly introduced before, and the link is attached. The following will be a series of articles on the supply chain control tower, based on the SCOR model system, the system tells the control tower in different stages of analysis, the use of different technical capabilities, different application links and objects, the establishment of control tower channels and combined with cases and so on these aspects of the supply chain control tower. Welcome interested colleagues to exchange and discuss.
This article begins with the presentation of the different phases of the analysis method, as this is the fundamental path we take to build any digital project.
Analyze phase depth and different technical scenarios
1.1 Five stages of analysis
In the era of big data and intelligence. The increasing availability of analysis tools will bring data-based analysis to a new stage. Thomas Davenport and Jenny Harris introduced the thinking dimension of a matrix in their book Analyzing Competition (the link of the book is attached below). The horizontal axis divides the analysis method into five stages. Namely descriptive analysis, diagnostic analysis, predictive analysis, normative analysis and cognitive analysis. The vertical axis corresponds different stages of analysis to the descriptive information in its application. From the basic description of a certain phenomenon to the cause tracing, to prediction. Then regulation, and finally to the level of cognition.
These five analysis methods are usually implemented in stages and can solve problems at different levels and perspectives, but we cannot simply say that one analysis method is better than the others, and there are corresponding application scenarios to support decision making at each level.
1.2 Take the supply chain scenario as an example
Suppose that in a supply chain business department. The typical example of daily problems encountered is supply shortage, unable to meet customer demand. The first question we have to ask is, what happened? This is a descriptive analysis, which product is not delivered on time; Then. The next question is why is there a shortage due to late delivery? This analysis involves diagnostic analysis. We need to further inquire about which sub-material or assembly process or supplier’s production scheduling is later than expected? Once the problem is diagnosed, we can basically determine the cause of the problem, the subsequent impact, how long the next shortage will last, and when we expect it to be resolved, and that’s predictive analysis.
And then, what can we do, what resources are available to ameliorate this alarming problem, so as to balance supply and demand as best as possible? So, under the intervention of this optimization measure, we see a “correction”, a result of the expected improvement in supply, which is called normative analysis; Finally, to achieve intelligent level of cognitive analysis, that is, to try to learn the current supply chain problems, take measures to improve the information and experience given by these paths. Based on this, it is possible to develop predictive models such as AI-based models, from data to information integration and a fair degree of insight.
In combination with a problem in a supply scenario, the above briefly explains the basic rules of different stages of analysis when we analyze a problem, when enterprises want to quickly see the problem, find a solution to solve the problem, and review the lesson learned. As you can imagine, reaching the AI-level state is the optimal situation that we have been saying digital supply chain can achieve. The main functions of the supply chain control tower in different analysis stages will also be focused on, and the functions of multiple analysis stages can be integrated in different application scenarios.
1.3 Application scenarios of supply chain control tower in each analysis stage
These stages are distinguished based on different technical capabilities used. The following six typical control tower application possibilities are briefly introduced:
Real-time intelligent connection of data: Implement the acquisition of data for business process monitoring, process monitoring, etc
Problem analysis: Use predictive and descriptive analysis methods to transform the data itself from reactive to predictive
Impact analysis: Understanding the impact of a warning sign, from the ecosystem of digitization itself to the supply chain in a company’s business
Scene modeling: For different scenarios, provide appropriate simulation models for simulation, and predict possible responses after given a certain input
Collaborative response: Connecting and collaborating with each other throughout the ecosystem
AI Intelligence: High degree of automation through AI/ machine learning and other technologies
Then, from the perspective of supply chain control tower, what are the corresponding technological trends involved in digital supply chain at different stages of analysis? In the next article, we’ll do an introduction.
本文由数字化转型网（www.szhzxw.cn）转载而成，来源：供应链图书馆；编辑/翻译：数字化转型网宁檬树。
免责声明: 本网站(http://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。
