01 Excellent CIOs pay more attention to the methodology of combining strategy and IT
When it comes to CIO, what comes to your mind? For the management and implementation of the company's internal information technology projects? Do not! This is past tense. Today's CIOs have strategies in their eyes, technologies in their hands, and can carry business on their shoulders, even leading the development and transformation of the entire company. In the past, our expectations for CIOs might be for the company's internal technology or for some implementation projects and management of IT, but now everyone's expectations for CIOs are getting higher and higher, and they even hope that CIOs can lead the development and transformation of the entire company. I have interpreted CIOs at three different levels: our traditional CIOs focus more on the technical level, but many excellent CIOs have expanded from the technical level to the business level; and outstanding CEOs are more capable of Focus on the methodology of combining strategy and IT.
If IT is separated from strategy, the CIO may face the problem that the company's decision-makers will feel that the money spent is not on the knife, and the things being done are not directly related to the strategy. On the contrary, this CIO is very accomplished. More than 40 years ago, BCG Bruce Henderson proposed the growth share matrix, which distinguishes the company's business products from the two dimensions of market share and high growth rate. And this management tool is still very suitable for today's business environment. For example, if the product share and growth rate are both high, then we define this product as a "star product"; and if a product has a relatively high market share, but the growth rate has slowed down, it means that this is a product. A mature product is also a high-profit product; for some products that may have a high growth rate, but the market share is not particularly high, they may face success or failure, which we call "potential products"; finally, the growth rate is low and the market Products with a low market share, which we call "declining products", are products that need to be eliminated. Growth share now requires faster adoption and a greater focus on strategic experimentation to accommodate an increasingly unpredictable business environment. Since market share is no longer a strong indicator of performance forecasting, the matrix requires a new measure of competitiveness. Successful companies need to explore new products, markets and business models more frequently, continually update their strengths through disciplined experimentation, and do this more systematically to avoid wasting resources. Companies are now required to invest in more "problem products" in a faster and more economical way than their competitors, and systematically develop potential "problem products" into "star products." At the same time, companies need to be prepared to respond to changes in the market, monetize "star products" faster, withdraw "cash cow products" from the market, and maximize the information value of "thin dog products". From this perspective, IT is the "product" of the CIO. In order to increase the market share and growth rate of the product, the CIO needs to consider the corporate strategy, personnel, technology, etc. Higher value capture "occupancy" -- that's what we call a "digital strategy".
The construction of China and Taiwan is an important means for CIOs to support the internal strategy and product implementation of the enterprise.
02 What is Zhongtai?
In the era of digital wave sweeping almost all industries, Zhongtai has become one of the important drivers of digital transformation, and some people will blindly equate digitalization with Zhongtai. In fact, Zhongtai does not mean digitalization. In addition, it should be reminded that the construction of China and Taiwan is not suitable for start-up enterprises, but should be in the stage of rapid development of enterprises, such as enterprises in the expansion period, breakthrough period, and mature period. And, if you don't make a strategic move to digitization, you're very likely to make mistakes. Remind this again.
The so-called middle office, from a business point of view, is to abstract and refine the same or similar parts from several vertical business processes, and turn them into a shared independent business unit to provide services to the outside world. The emergence of China-Taiwan has solved the various problems exposed by the enterprise in the process of the rapid development of the company's business and the continuous expansion of the organization.
For example, Alibaba's digital middle office:
The predecessor of Alibaba's middle office is actually the Alibaba Sharing Division. Why is China Taiwan derived? How did Alibaba Zhongtai come into being? Mainly based on Alibaba's huge and complex business. For example, Alibaba has an e-commerce business, and there are first-hand e-commerce and second-hand e-commerce under the e-commerce business, such as Tmall and Xianyu. If each business unit operates independently, there will be a business department and an IT department. Repeat similar transaction work business. So at this time, enterprises must consider: how to turn the trading center into a shared center to support different trading scenarios - this is the initial concept of Zhongtai. The same is true for order splitting, creating order payment, modifying price, closing order, deleting order, confirming receipt, automatic delivery, order inquiry, and so on. If you continue to dig deeper, you will find that the underlying operation and maintenance logic, data logic, and security logic of these businesses are all the same. We pull these out separately to make a trading center. This is Zhongtai.
03 When will you be a middle station?
So when are you going to be in the middle? My suggestion is: Before introducing the middle stage, in addition to considering whether it is suitable for the middle stage at this stage, the company should first torture itself and ask itself the following questions: Is it inefficient? Can it be reused? Is the business in a mature stage? Do you understand that what China and Taiwan are facing is the problem of "organization", that is, the problem of "people"? Zhang Yong, chairman and CEO of Alibaba, once mentioned a concept: if a company is rushing to become a middle-stage company, it will die.
The reason why there is a high probability of failure in blindly being in the middle stage is that the middle stage is not a simple IT problem, it faces more organizational problems, and organizational problems are people's problems. Blindly importing the middle platform will eventually face three major problems: difficult development, difficulty in learning, and difficulty in compatibility. If these three difficulties are not solved, the entire business change will take a long time, and the business middle platform will become heavier and heavier, and finally lead to this business. The state is discarded when no one is using it.
01 Difficult to learn from: Foresight of future business development. There are not many rules to follow in the new changes of enterprises, and there is not much history to learn from. It may happen today that a scenario that will not happen will happen the next day, resulting in a lack of business processes and system construction. It did not happen, resulting in a waste of construction, and as a result, the opportunity cost of development was paid.
02 Compatibility is difficult: Compatibility with existing businesses. It is not difficult to write a bunch of code for a certain business requirement. What is difficult is to make a general model to be compatible with existing and known business scenarios; to build a system from scratch actually requires It's easier than changing an existing system, and adding a piece of code to ensure that the original logic does not have problems is more difficult.
03 Difficulty in development: The pressure of rapid business development. The advantages of the middle platform are fast implementation and low cost, which is conducive to the rapid development of business. This is a prerequisite, the new business and the existing business should be similar enough, or even the same. If there is a big difference between the new business and the existing business, the transformation time will be longer and the cost will be higher.
The combination of these three difficulties leads to three possible adverse consequences: 1. The change takes a long time, and the more and more heavy it is, 2. The middle stage is abandoned, and the smaller the middle stage, 3. The "pseudo" middle stage, the more and more stupid
Therefore, to be a middle-office, you should first think about two questions:
1. Is the business stable?
Stability does not mean that the business does not change, but refers to whether the business changes within a predetermined range, and whether the business changes have rules to follow. For example, the sales management center mentioned above often opens or closes the sales qualification of an insurance product for a certain channel, which is stable. But if there are often unpredictable constraints on sales eligibility, it is unstable. Stability and instability, in fact, depends on the person who designed it and the goal of the design, different levels of abstraction and induction methods, and there will be differences in the judgment of stability or instability.
2. Is the business developing rapidly?
Rapid expansion of new similar business lines, or frequent changes in multiple business lines, are considered rapid development. For example, new third-party platforms are often added to sell insurance products, and the sales process is consistent with the existing sales channels, which is in line with the characteristics of rapid business development. If there are only one or two sales channels, for example, some companies mainly use offline agent channels, and no matter how fast their development is, it does not make much sense to have a sales management center. If you want to dismantle a subsystem, it is not a middle platform. Its purpose is to decouple and reduce the size of the system, not to be common and general.
As we said earlier, China and Taiwan are more faced with organizational and human problems. Therefore, to be a middle and Taiwan should first consider setting up a corresponding organization according to the business process after the business process is solidified, and finally become a middle and Taiwan. , that is, import from the entire business model to people, then to the organization, and finally to the business state. In addition, the middle stage is not suitable for every stage of every company. In the period of independent business expansion and breakthrough, "it must be done with independent regiments, independent divisions, and independent brigade establishments", otherwise it will become a bottleneck; but at a certain stage of development, there are too many hills, and it is necessary to "shut down and transfer, It is necessary to merge similar items. Ask for efficiency in management, and cancel repetitive construction.” For example, Suning Technology’s organizational structure model: in Suning Technology’s organizational structure, each BU or R&D center is an independent business, and it is also an R&D center. , the entire business has a very high degree of overlap with the R&D department:
Another practical case is Suning Retail's "Smart Retail Brain" concept, called "Retail as a Service". With so many people talking about digitalization, is there a retail company with technological capabilities that can empower the entire retail industry? So we propose RaaS. You may not know much about Suning's current R&D strength, which is relatively strong, which is rare in the retail industry. We gathered them not only engineers, but also retail experts. Separately pull out the infrastructure of artificial intelligence and big data to make a service; at the PaaS level of the platform, we pull out some common business, such as membership, transaction, supply chain, etc., to make a platform. In terms of AI, there are three basic computing platforms, which allow our computing engineers to develop many applications on them, and also allow us to have more than 80 AI retail scenarios and 150 AI skills that can be invoked . Create a big data middle platform that can carry many SaaS applications, such as finance, logistics, supply chain, marketing, and various user-based SaaS services:
Therefore, the definition of the business platform should be an individual full of vitality. It can carry all the business flows and logic of your entire company, and it can also precipitate the business data. When you need data, there is business data that can be used. , and this middle platform should generate business value for the enterprise. Basically, the fact that a company can become a middle-stage company indicates that the company has passed the infancy and start-up stages, and is in a relatively mature stage or a stage of rapid development.
04 How to do business in the middle?
We say that business middle-office should be based on business battles. So how do we start from the direction of a battle, and then go to informatization, and connect the whole point, line, and surface? The following questions should be considered:
- How to judge the value of existing data, plan and continue to think deeply about how digital transformation can create value for enterprises
- How to formulate a strategy and route for the landing of the data center
- How to accurately evaluate the ROI of the data center
- How to design the corresponding organizational structure and divide rights and obligations
- How to make the right technology selection, and grasp the balance between purchasing products and independent research and development
When an enterprise encounters some business opportunities, from the perspective of digitalization, it should be considered from two levels. The first is the business opportunity of the enterprise, and the other is the practicality of technology. These two must be fully integrated. First of all, based on the business opportunities of the enterprise and the direction of the battle, the enterprise structure design and business activity formulation are carried out, and finally the construction direction of the business middle platform is exported. In addition, the business middle office is a living entity that carries business logic, precipitates business data, generates business value, and evolves with the continuous development of the business. Designers should also consider some principles, and set these principles in the beginning to avoid useless work later:
- Service loose coupling principle
- interface-oriented implementation
- Decoupling of asynchronous events
- Service provider location decoupling
- Version loosely coupled
- Service Dependency Principle
- Valuable Domain Model
- Minimum dependencies between services
- Capability entities are hierarchical
- Delayed reliance on tech components
- Service Design Principles
- Optimize remote calls
- remove redundant data
- Design rough and subtle service interface
- Identify and design common service interfaces
- Isolate changes within a service
- Service interface first
- The service interface is downward compatible
- Service Naming Guidelines
- Service Granularity Principle
- The Stateless Principle of Services
- Service Operation Design Principles
- Service Binding Principle
There are many experts writing similar principles in the market, so I won't say much here.
When an enterprise develops business opportunities and battles, it should combine business battles with information technology battles. In short, establishing the relationship between strategic objectives and information support provides a strong guarantee for the realization of strategic objectives. The business frame and IT
Here is Suning as an example: Suning has a product called Retail Cloud, which mainly focuses on the "sinking market" dominated by counties and towns, and consumer goods dominated by 3C. Driven by various measures to benefit the people, the county has launched a A new potential for town consumption. Suning's retail cloud is to empower and export retail capabilities through digitalization, allowing traditional brick-and-mortar store managers in townships to obtain successful experience and capabilities from the system when they join, so as to reduce the risk of franchise operations. The approach here is to combine the business center, data center and artificial intelligence to make a common state, enabling external output:
In addition, when we talk about the middle office, we may refer to the business status, or we may refer to the data status. Taking Alibaba Cloud as an example, Alibaba Cloud mainly promotes the data center, not the business status. The reason is that Alibaba Cloud currently cannot meet all business scenarios, and the business scenarios of each industry are different. And data can be used universally, so the data center has become an easier plan to promote in the enterprise.
- Data Repetitive Development
- High cost of data usage
- Waste of storage and computing resources
- The problem of business data silos is serious
- Data standards are not uniform
- Data utilization efficiency
However, in the actual operation process, there are still many enterprises facing the failure to promote the data center. What's the main reason for that?
- Difficulties in establishing and harmonizing data standards
- changing data needs
- Data security management is complex
- Technology selection is difficult
- Data correctness is difficult to determine
- Data Permission Management
- Diverse data needs
- Data management is complex
- Data is expensive and difficult to quantify
If the data center is finished in one sentence, in fact, its essence is "data warehouse + data service intermediary software.
05 Decentralization
Based on the above, I personally prefer the concept that everyone thinks, that is, the decentralized data state.
The practice of Silicon Valley companies such as Twitter, Facebook, and Airbnb is that the big data department provides enough useful tools to empower business departments to share data capabilities. The situation is different for some companies. They extract a certain capability and are responsible for it by a special group. In fact, this concept can be seen in many excellent leading Internet companies, such as Ali, JD.com, Suning, etc. in China.
What is the difference between a decentralized data state and a centralized data state? From a business perspective, the centralized data status usually requires the data middle-office business team to sort out the business logic of the entire company, which is very challenging for the data team. Decentralization is different. The data middle-end business team is mainly a tool provider, and the specific development capabilities will be in charge of the data R&D engineers of the business department. The advantage of this is that the data middle-end business team must ensure the functions of the entire middle-end and stability, and the business department is not only a user of the data center, but also a contributor. In the construction of the data center, in addition to the traditional big data team, the active participation of the business department is also required. Because the common data capabilities are related to the business, and the process of developing iterative operations requires coordination and communication with various business departments and IT departments, it is necessary to make overall arrangements for the participants when building a data center. This is also a problem that we often encounter in the process of data center planning. The following lists the personnel generally involved in the construction of the data center and their main responsibilities:
- Head of business department: In-depth understanding of business processes and priorities, able to map business scenarios to data, and guide the modeling process.
- Business System Architect: Understand the system architecture and technology framework of the enterprise. l Business process engineer: very familiar with business processes, usually the link between the technology department and the business department.
- Data platform engineer: usually has a system engineer background, responsible for building and operating a data platform, installing and operating various big data components, and ensuring the efficiency and stability of the data platform.
- Data development engineer: With the background of data warehouse skills, understand the business, responsible for modeling, data cleaning and writing ETL programs.
- Data application development engineer: With the background of application development, develop data applications serving business departments.
- Data middle-office architect: fully grasp the functions of the data platform, provide data support and requirements for the company's products, be responsible for the integration of the company's products and data platforms, the architectural planning for connection with business systems, and the promotion and control of the company's data standards .
- Data analyst: with a background in statistics, able to generate reasonable and accurate business intelligence reports from data.
- Data Scientist: With the background of machine learning, provide data analysis products and results based on machine learning and artificial intelligence.
- Data Product Manager: Responsible for planning and coordinating the development process of data capabilities within the company, sometimes this role is taken on by a data architect.
The data center business team leads the technology formulation, and the business departments jointly develop the application layer. This advantage has been reflected in the core capabilities of many leading companies.
All in all, when we are a business middle-office, we need to think about the business state from the beginning of design. A successful business state should design some principles from the first day. This principle will help the business middle-office not go astray when it develops later. , there is a datum direction to follow. If you want to evaluate how much money it takes to evaluate the middle office, this is a question that many people will ask me. I generally have three dimensions that I can introduce to you. You can answer this question from these three dimensions:
1. Estimate according to actual needs. Taking RT-Mart as an example, a rigorous R&D process can be introduced.
To put it simply, enterprises must define their own needs before they can enter R&D. Otherwise, it may lead to changes in project requirements, new budgets, and other risks. This dimension is to let everyone think about where the real needs of the middle and Taiwan are. . And the rigorous and efficient R&D process can evaluate the cost through people and days.
Advantages: According to the specific needs of enterprise development, it can fully explain how much it actually costs to be a middle office.
2. Calculated according to the annual IT budget, and estimated by companies with annual revenue of tens of billions.
Let's make a calculation. Suppose a company with an annual revenue of 10 billion has a net profit of 100 million, and the annual IT budget generally falls within 2-7%. Suppose this company is an enterprise that attaches great importance to digital construction, and its IT budget is as high as 100 million. 7%. It is equivalent to investing 7% of the 10 billion revenue, that is, 7 million in the annual digitization project, and the middle station is the key project of the year, accounting for 30% of the budget, then 30% of the middle station budget, that is 2.1 million. But the problem is coming. If it is 2.1 million, China and Taiwan can only outsource, and due to budget problems, the failure rate is high. After all, this is a growing company with tens of billions of revenue, so it is estimated that some moderate adjustments should be made. It is not possible to formulate plans according to this dimension, but the evaluation method of this dimension has certain reference value.
Advantages: According to the specific needs of enterprise development, it can fully explain how much it actually costs to be a middle office.
3. According to past experience and the current situation of the enterprise, estimate the cost of a self-developed data center.
In the construction of the data center, in addition to the traditional big data team, the active participation of the business department is also required. This picture is to calculate how much the cost should be based on the actual situation. As a result, it may be estimated that the annual expenditure is 80 million. Assuming that the average salary is 400,000, the annual labor expenditure is about 42 million, plus hardware and operation and maintenance costs, about 80 million.
Advantages: Evaluation and judgment based on past experience, there is a certain reference value.
Finally, a reminder that the data center is difficult to build and even more difficult to maintain.
Finally, returning to the CIO's needs for building a digital middle platform, my suggestions are:
- Better estimates, incorporating digitization into strategic planning communications
- Carry out sufficient research and analysis according to the input and output of digitalization and the value it may bring
- Obtain consensus from board members and senior management
- Finally, according to the actual needs assessment, put into implementation
- Hold a meeting to summarize the project
The reason why I do this sharing is because too many friends have asked me about the construction of Zhongtai, so I will do a systematic sharing. Finally, whether it is digitization or the issue of China and Taiwan, you are welcome to communicate with me, Wang Junjie's personal WeChat account, jawinjawin.
