In the first two lectures, we completed the first round of enterprise-level divergence and convergence in the D4 model through Discovery combined with Define. Standing at the height of the enterprise, based on the enterprise vision and internal and external environment, we determined the final platform-based enterprise architecture through top-down strategic decomposition and bottom-up current situation investigation, and then applied enterprise architecture analysis methods to answer questions such as whether to build a middle platform, what types of mid-platform buildings are needed, who should build first and who should build later.

Starting from this lecture, we will enter the second round of divergence and convergence of the D4 model. From the perspective of a mid-platform product, let’s take a look at how to design and implement it, which is the two stages of Design and Delivery in D4. Similarly, this lecture starts with Design. We will use the design and implementation of a Central Product Platform as a sample to explain the ideas and key points of this stage.

Okay, at this point, we still need to borrow the example of Geek Real Estate.

We assume that after the first round of enterprise-level research and architecture design, Xiao Wang and his team found an urgent need to build Geek Real Estate’s own Central Product Platform to achieve the company’s strategic goal in 2022, which is to transform from heavy asset business to light asset service business, and better provide multi-scenario services for Geek Real Estate’s client base, that is, IT practitioners.

So the next question is, how should we build the Central Product Platform? And where should we start building the Central Product Platform?

In this class, let’s discuss the specific design and launch issues of the middle platform together. In the current example, it is the design process of Geek Real Estate’s Central Product Platform (hereinafter referred to as Geek Middle Platform).

During the design phase, the questions we need to answer are: what is the vision of the geek middle platform? What does it include? Where does it start to be implemented? After completing the design phase, we should have a clear idea of how to start mid-platform building, which can allow the team to start developing the middle platform MVP. If you are not sure what MVP is, don’t worry, it will be introduced later.

# Determine the vision of the middle platform product

By now, you must know that the first step is to clarify the vision and goals of this Central Product Platform, because this is the foundation of our entire mid-platform building process and the original intention of mid-platform building.

Do you remember what we mentioned before? The construction of the middle platform is a road full of thorns and twists and turns. We will encounter various problems and choices during the process. A good vision for the middle platform can help us always maintain our original intention and not easily lose our direction.

So how can we develop a good vision for the middle platform? When I go to a company to do consulting related to the middle platform, the first question I ask the person in charge of mid-platform building is: You said you want to do the middle platform, what is the vision of mid-platform building?

Most of the answers I received were positive, after all, no one would admit that their project doesn’t even have a goal or vision. But once asked what the specific vision is, it usually starts with a long narrative, from the company’s history to the current situation, from industry discussions to recent speeches by the company’s management. In short, my summary is: if the leader says we want to do the middle platform, we will do it.

I think this is quite dangerous. After all, as we mentioned earlier, the vision of the middle platform, as the starting point of all problems, needs to be simple and clear enough to become a tool that guides the direction of our mid-platform building process like a Compass, and also to achieve the “look at the vision when in doubt” mentioned earlier.

Here is a simple and useful tool that we often use to help us diverge and converge our product vision. Of course, it is also applicable to the middle platform, which is the Elevator Pitch.

As the name suggests, the elevator speech assumes a scenario where we encounter the company’s management on the elevator and can explain to them clearly what we are doing, such as the middle platform, within a limited time. This requires us to be sufficiently restrained. Therefore, the form of the elevator speech mainly limits several key factors of the product, such as who the user is, what problems they solve, and what the differentiation characteristics are. Each element is very important. If you cannot answer or the team’s answer is different, it reflects to some extent that everyone has not yet reached a consensus on the vision of mid-platform building, which will lay the foundation for the emergence of various problems in the future.

Don’t underestimate such a small elevator speech. It feels like answering a few questions is enough. From my actual experience, every discussion on planning the mid-platform product vision will exceed the time limit. At first, I thought it might be such a simple thing that could be done in an hour. Later, because the discussion was too intense, the time expanded to half a day. Now, we usually reserve half a day or even a whole day to fully diverge and converge the mid-platform product vision.

But even though we have spent so much time, I still think it is very necessary. It is worth spending more time on discussing the vision, because it will affect the future direction of our mid-platform design and construction. Once formed, it will become the foundation of our mid-platform building. The importance of this step cannot be overemphasized.

Here is a question that needs to be mentioned. The value and difficulty of vision lies in full convergence But I have seen many companies’ visions, although they also use elevator speeches, but the result is full of walls. Although it looks shocking, the effect is actually greatly discounted. After all, we cannot repeat the content of a whole wall in one elevator time. As the saying goes, addition is easy, subtraction is difficult.

# Determine the scope of business consolidation

After the vision of the middle platform product is determined, the next step is to sort out the business architecture of fine grain, extract commonalities, and identify the specific needs of the middle platform product.

However, for the same enterprise-level reasons, the problem we often face is not knowing where to start. After all, the middle platform has enterprise-level attributes. Even if it is business sorting, it will involve the entire Line of Business of the enterprise, and it is an end-to-end and full-process sorting work. The workload is often daunting.

Do all Lines of Business need to be sorted out? Do all stages of end-to-end need to be sorted out? What level of granularity does business sorting need to reach? These are frequently faced issues.

Generally, the situation of different companies is different here. If the overall scale of the company is not particularly large, it is also possible to do end-to-end business sorting for the entire business. The advantage of this is that it is more comprehensive and less likely to be missed. It can have an overall picture of the business status of the enterprise, and the identification of common middle platform needs will also be more accurate.

However, many enterprises, especially those with the highest demands in the middle platform, are often large conglomerates with multiple Lines of Business. For enterprises of this scale, it is basically impossible to do end-to-end sorting of the entire business, otherwise just coordinating personnel will be a troublesome problem. At this time, we need to first determine a scope, and then carry out business sorting and development within this scope.

So how should this scope be defined? As the saying goes, when in doubt, look at the vision. This is where the mid-platform vision we discussed earlier comes in handy.

Taking Geek Real Estate as an example, the vision of mid-platform building is to quickly achieve lightweight transformation of the business through the ability to reuse mature business capabilities and reuse them in other new types of light asset Line of Business, such as long-term rental apartment business (specific elevator speeches will not be elaborated here).

Currently, Geek Real Estate focuses on serving the IT community, and its real estate and property businesses have matured with a very sound IT infrastructure. There is a strong demand for long-term rental apartments targeting this group, but as a new Line of Business, long-term rental apartments start from scratch in terms of investment expansion, design, inviting tenders procurement, engineering construction, decoration and renovation, customer management, and property operation.

Returning to the issue of business organization scope, with a clear vision, we can make our business organization more focused.

First of all, from the perspective of business lines, not all Lines of Business need to be sorted out. Let’s take Geek Real Estate as an example. Its subsidiaries have a wide range of interests, and Line of Business focuses on investment, culture, education, etc. Because these Lines of Business have long-term independent development and are not directly related to the vision and goals of Geek Real Estate’s mid-platform products mentioned earlier, they can be skipped directly.

Then comes the end-to-end business Value Chain. Through analysis, since the company’s strategy is to serve light assets, it is not applicable to the construction of projects in the real estate industry, that is, the part of building buildings and houses, and the new business based on the company’s light asset strategy investment, so there is no need for further sorting.

Based on the vision of the middle platform construction, we can converge into a focused area in the horizontal end-to-end Value Chain and vertical business lines.

Further business sorting and development in this focused area will be more efficient and focused.

# Finishing of fine grain business

After determining the scope of sorting, the next step is the specific business sorting work.

Generally, everyone’s approach is based on existing processes or systems to sort out the existing business processes, such as the four flows often mentioned in the field of e-commerce, which specifically refers to the sorting work of information flow, business flow, capital flow, and logistics. The tools used are mostly flowcharts , which are very mature tools.

However, this kind of sorting based on existing processes sometimes has some limitations. Simply put, it may deduce pseudo-middle platform requirements based on past business debts. What does this mean? It means that for some long-term and mature businesses, the existing business methods may not be the best way, but rather the product of long-term development and compromise with various peripheral and IT restrictions in the past.

For example, Line of Business, which usually takes a long time, will have a very complex approval and review process. However, the review and approval process is actually just a solution. The problem to be solved may be process fraud or other internal risks and problems. Complex approval processes are often the product of process control in the information age or even earlier paper age. Since information technology only digitizes offline processes, it has continued to this day. In order to highly replicate and restore the process of the paper age, various strange approval logics, such as countersigning, have emerged with Chinese characteristics.

However, if we return to the problem domain and rethink the problem itself, we will find that in today’s digital age, where Big data and AI are widely used, there may no longer be a need for process approval. Through new technologies, we can use real-time audit analysis or risk identification to discover some anomalies in business processes, and even completely automate and eliminate the involvement of intermediaries, fundamentally solving these problems and truly unleashing the power of digitalization.

This kind of thinking method of returning to the problem domain itself, returning to the origin of the problem, and redesigning the solution based on the user request under the current business and technical conditions can avoid the business debt trap mentioned above, and is also mostly used in innovative business. The design process, and this way of thinking also refers to and borrows a lot of design thinking methods.

Therefore, unlike the traditional business process sorting based on the status quo, we have adopted a large number of methods based on design thinking , combined with User Experience Map (User Journey Map) and Service Blueprint (Service Map) in the process of business sorting. Returning to the business itself, starting from the problem domain, user-centered, User Experience design and business service blueprint sorting are also very good in terms of effect.

Of course, it’s not about which method is better. Different tools and methods have different applicable scenarios and advantages and disadvantages. If they can be flexibly mastered, they can even be used in combination to fully leverage their respective advantages.

By sorting out the business architecture within the scope, combined with the analysis of cross-scenario general capabilities, we can have a certain in-depth understanding of the business overview of the focus area, and identify some business data, business functions, and business processes that can be reused in different Lines of Business. These general business data, business functions, and business processes are processed and analyzed to form the specific requirements of the middle platform. For these requirements, we describe them through User Story (User Story).

# Determine MVP

Through business sorting, we have identified and organized a large number of Central Product Platform requirements. It is worth noting that the so-called middle platform product requirements at this time are more derived from qualitative abstraction. Considering that the requirements of the middle platform are often not as clear as those of the front-end business system, strictly speaking, the requirements at this time still belong to a series of high-risk assumptions.

From the mid-platform building projects I have actually participated in, it is often the case that the mid-platform requirements initially envisioned deviate greatly from the original expectations after the actual development is completed. This means that the longer the mid-platform building cycle, the greater the risk and cost of the assumed requirements.

Therefore, in the process of building the middle platform, we introduced the MVP principle (Minimum Viable Product) in Lean Startup, and also implemented the Start Small in the overall principle I mentioned earlier.

In order to achieve MVP and ensure that the MVP produced is usable, we adopt an end-to-end vertical segmentation method for business requirements, combined with multi-dimensional scoring and sorting of requirement priorities, to ultimately determine the scope of MVP requirements. The final form of implementation may be the first iteration of the User Story list.

Now that we have a specific User Story list, can we start developing it? Don’t worry, there are two things we suggest considering before starting mid-platform building, which are operational plans and metrics.

# Operation prerequisites: Develop iteration plans and access plans

First, let’s take a look at the operation plan. For the middle platform, it may be the promotion of middle platform products, the front-end (user) access plan, and the operation support after access.

What I have seen in the past is that it is often too late to start considering these issues in the middle and later stages of mid-platform building. Moreover, in many cases, the front-end will not stop and wait for mid-platform building to be completed, and then start their own business evolution after integration, so it is often the front-end and middle-end parallel.

This process is very similar to branch development in our development, modifying the same piece of content, so it will be very painful when merging.

Therefore, it is crucial to consider the operation plan, especially the access plan, in advance, and try to use a reasonable access plan (I have seen many companies call this access plan the on-board plan, which is very vivid) to avoid some risks, which is also crucial for the construction of mid-platform products. The specific operation strategy will be introduced to you in the next article.

# Metric Prerequisites: Define Validation Metrics

Another thing that needs to be clear before the mid-platform product construction starts is the measurement index. I don’t know if you remember this, but we have already talked about the four questions that need to be clear before mid-platform building. The specific reason why it is necessary to design the measurement index before development has been explained before. Here, I will share some design ideas for the measurement index with you.

First, let’s take a look at the commonly used indicators. I have divided the most common product measurement indicators on the market into four categories: strategic, user, market expansion, and cost reduction and efficiency improvement. Based on each dimension, many commonly used measurement indicators can be found.

As for the middle platform, the difficulty lies in the fact that there is still a front-end layer between it and the market and users, so the measurement is not as clear as the front-end system that directly contacts users. However, we say that the middle platform is empowering the front-end business, and it cannot be separated from the business. Simply measuring some technical aspects, such as interface stability and interface quantity, does not conform to the positioning and value of the middle platform for business support empowerment.

Therefore, when considering the measurement indicators of the middle platform, we generally take strategic value and business value as the starting point, start to break down and deduce, and use multi-dimensional and multi-level indicator design to examine the effect of mid-platform building according to the different concerns of stakeholders. It should be emphasized here that although the dimensions and perspectives are different, we must ensure that the mid-platform building goal reflected by all indicators must be the same.

As for the specific implementation method, it is difficult to give a standard answer because each middle platform product is different. In the practical operation process, it is recommended that you think more from the perspective of others and ask more “how” questions. I believe you are familiar with the 5Why analysis method. Here we can slightly modify it and use 5How to drive the design of verification indicators.

For example, if I stand from the perspective of the company’s management:

• So how do I judge the results of mid-platform building? If the answer is to empower new businesses,
• So how do I verify the effectiveness of empowering new businesses? If the answer is to look at the launch speed of new businesses,
• So how do I verify the launch speed of the new business? If the answer is to look at the time from project approval to launch of the new business,
• Then how do I verify…

You see, people often say that measurement is difficult. In fact, everyone has a scale in their heart, but we just haven’t expressed it clearly. Through 5How, we can constantly ask questions and discover this scale in everyone’s heart to better guide the construction and promotion of our mid-platform products.

# Summarize thinking

This lesson takes the construction process of a Central Product Platform as an example. Taking the vision of the Central Product Platform as the starting point, it introduces the tools for elevator speeches. Then, based on the mid-platform building vision, it clarifies the scope and priority of business sorting.

After clarifying the entry point and scope of business sorting, we can carry out user-centered business process sorting and service design based on design thinking, combine domain-driven design for business model sorting, and finally identify the middle platform requirements of common business data, business processes, and business models through cross-Line of Business comprehensive analysis.

Afterwards, by introducing the MVP principle from Lean Startup, end-to-end vertical segmentation is adopted based on business priorities for mid-platform requirements, and the starting point of mid-platform building is finally determined, which is the specific construction content of the first phase.

Finally, with the analysis and design of the product operation mechanism and measurement indicators in the middle platform, we can officially initiate the project and enter the delivery stage for specific implementation.

In the next lecture, we will officially enter the final stage of D4, which is the delivery stage of Delivery. Let’s take a look at the ideas and methods we use in the delivery stage, as well as the issues we need to pay attention to.

Finally, I’ll leave you with a few thought-provoking questions:

• What is your business sorting process like? How are the middle platform requirements analyzed?
• Have you adopted the MVP principle? How do you avoid the hypothetical risk of mid-platform building?

Please write down your thoughts in the comment section and discuss with me. You are also welcome to share today’s content with your friends and learn together. See you in the next lecture!

In the previous two lectures, I took you through the development process of the middle platform from a time perspective, and also introduced various middle platforms that have appeared on the market from a spatial perspective.

You must be a bit confused by so many types of middle platforms now. Can these middle platforms be called middle platforms? It feels no different from the platforms you have been working on before, right?

This is also a question that I have been constantly asking myself in my heart for the past few years. As the last article of the concept section, I will try to take you out of the dimension of time and space to analyze why companies need to build a middle platform. When we talk about the middle platform, what are we really talking about?

# Why do companies need to build a middle platform?

Let’s first look at why companies need to build a middle platform. To answer this question, we need to first solve another question, which is why companies need platformization?

Why do companies need platformization? Here’s my answer:

Because in today’s internet era, users are the center of the business battlefield. In order to quickly respond to user needs, leveraging the power of platforms can achieve twice the result with half the effort.

The logic behind this is simple: constantly responding, exploring, exploring, and leading users’ needs is the key factor for enterprises to survive and continue to develop.

Those companies that truly respect users and even adjust or overturn themselves to respond to users will survive and develop in this user-centered business war. Conversely, those companies that are complacent about past achievements and hope that users will continue to follow them as before will be eliminated by users.

It’s cruel, but this is the most basic corporate survival rule of this era.

The reason why platformization is important is that in this user-centered modern business war, it endows or strengthens the most core ability of enterprises: user responsiveness . Platform Thinking encourages enterprises to constantly abstract and precipitate their core underlying capabilities. Through platform packaging, it can better empower front-end business and help enterprises cope with the uncertainty of front-end business and end-user requests with the certainty of the underlying layer.

From the results, it can be seen that in today’s uncertain business battlefield, the real trendsetters are mostly those enterprises that have platform thinking and have started to focus on platform construction early on, which also proves the importance of platform for a company to some extent.

Okay, now we understand the first question: why do enterprises need platformization? However, platformization is not a new concept, and many enterprises have been working hard and accumulating in this direction for many years. So why has the relatively new concept of “middle platform” emerged in recent years? For enterprises, why can’t the traditional "front-end + back-end (or platform) " platformization architecture meet their requirements?

This brings us to our core question: Why do companies need to build a middle platform?

Similarly, give me the answer first:

Setting up functional teams is the next stop of platformization. It is a process in which the platform continuously evolves its own governance, breaks through technical boundaries, gradually embraces business, accommodates business, and has stronger business attributes. The middle platform focuses on empowering front-end business, truly born for the front-end.

In order to explain my point more clearly, it is necessary to first define what front-end and back-end refer to in the context of today’s discussion.

• Front-end: A front-end business platform composed of various front-end systems. Each front-end system is a user touchpoint, mostly directly used by enterprise end-users, and is the intersection between enterprises and end-users. For example, websites, mobile apps, WeChat official accounts, Mini Programs, etc. directly used by users belong to the front-end category.
• Backend: A backend support platform composed of backend systems. Each backend system generally manages a type of core resources (data + computing) of the enterprise, such as financial systems, product systems, customer management systems, warehouse logistics management systems, etc. These systems constitute the backend of the enterprise. (After chatting with many friends in the Internet industry, many Internet companies do not have the concept of backend, and more directly use the concept of platform, such as dividing it into the front-end layer and the platform layer, but the position and function are similar to the backend in traditional enterprises. Here, I directly use the concept of backend to represent it.)

After defining the front-end and back-end, let’s come back to the question of why enterprises need to build a middle platform.

We see that many companies’ back-end systems were not primarily designed to serve business innovation in the front-end system, but rather to achieve electronic management of back-end resources and solve the efficiency problem of enterprise management. These systems are either software packages purchased at a high price in the past, requiring a large amount of service fees to be paid annually, and some versions are already very old, making customization difficult; or they are self-built at a high cost, with long-term disrepair and patches all over them, making it difficult to change and basically impossible to change. They are also the so-called “legacy systems” of enterprises.

It can be said that for most enterprises, the existing backend either cannot be used by the front-end, is not user-friendly, or cannot keep up with the pace of front-end business development. In summary, it can be summed up in two words: “slow” and “expensive”. The response to business is slow, and it costs a lot of money to change small functions.

Some people may say, you can’t bring up legacy systems, we can create a new backend system, the whole 2.0, won’t the problem be solved?

This is a problem-solving approach. However, even for newly built back-end systems, because the back-end management is often the key core data of the enterprise, considering restrictions such as enterprise security, auditing, compliance, and laws, such systems are often unable to be directly used by the front-end system, or are subject to various restrictions and cannot change quickly, and cannot support the rapid business innovation needs of the front-end.

At this time, the front-end and back-end are like two gears with different speeds. The front-end needs to respond quickly to the needs of front-end users, emphasizing rapid iteration and innovation, so the faster the speed, the better. The back-end, on the other hand, faces relatively stable enterprise core back-end resources, often with outdated and complex systems, and even subject to compliance constraints such as laws, regulations, and audits. Generally, stability is the top priority, and the slower the speed, the safer it is.

Therefore, with the continuous development of enterprise business, the problem of “unbalanced matching” of “front-end + back-end” gear speed has gradually emerged.

With the continuous development and growth of enterprise business, the cost and risk of back-end modification are increasing. This drives us to try our best to maintain the stability of the back-end system, but at the same time, we also need to respond to the continuous needs of users. What should we do? The most natural way is to directly stuff a large amount of business logic (business capabilities) into the front-end system, because the front-end is close to the user and the response is relatively fast.

However, this also comes at a cost. The consequence is the introduction of duplication, which also leads to the continuous expansion and bloating of the front-end system, forming a “chimney-style single application” of large mud balls. The result of this situation is that the “user responsiveness” of the front-end system is gradually worn down, the Customer Satisfaction Score gradually decreases, and the competitiveness of the enterprise also continues to decline.

Gatner released a report in 2016 titled “Pace-Layered Application Strategy” to address this issue. The report proposed a solution, which is to divide the enterprise’s application system into three levels based on “pace” (which coincides with the three levels of front-end, middle-end, and back-end), and adopt completely different strategies for each level.

The Pace-Layered Application Strategy also provides theoretical support for the inevitability of the “middle platform”.

In this report, Gatner proposed that the systems built by enterprises can be divided into three categories from the perspective of Pace-Layered: SOR (Systems of Record), SOD (Systems of Differentiation), and SOI (Systems of Innovation). Each type of system has a different rate of change, so it will also have different lifecycles, suitable for different technical architectures, different development processes, and even different investment models.

Going back to the two-layer architecture of the backend and frontend mentioned earlier, if mapped to this model, it is actually a two-layer application architecture with only SOR (backend) and SOI (frontend). In this case, we also need to try our best to maintain the stability and reliability of the backend (SOR) system, and the frontend system (SOI) should be small and beautiful, and iterate quickly. Therefore, the problem of “gear matching imbalance” mentioned earlier naturally arises, and it feels like we can’t have our cake and eat it too.

What to do? The SOD in Place-Layered is the answer. It provides stronger stability than the front-end (SOI) and higher flexibility than the back-end (SOR), finding a wonderful balance between stability and flexibility.

The middle platform is SOD. With this layer of “middle platform”, we can “sink” the stable and universal business capabilities in the already bloated front-end system to the middle platform layer to help the front-end lose weight and restore its responsiveness. We can also “extract” the business capabilities in the back-end system that need to change frequently or be directly used by the front-end to the middle platform layer, giving these business capabilities stronger flexibility and lower change costs. Alternatively, we can directly transform the back-end by setting up functional teams, accelerating the back-end through configuration, self-service, visual operation, and other forms, thus providing the front-end with more powerful, faster, and easier-to-use “capability artillery” support.

The middle platform is like a set of “gears” added between the front-end and the back-end, matching the speed of the front-end and the back-end, serving as a bridge and lubricant between the front-end and the back-end. It is born for the front-end, easy to use by the front-end, and smoothly directs back-end resources to users, supporting enterprises to better respond to users.

Okay, now we can summarize and answer the question of why enterprises need to build a middle platform. It’s simple. In the process of platformization, enterprises need to build their own middle platform in order to continuously provide better services to front-end business, support continuous response to users, and increase user responsiveness.

# Define the middle platform

After talking about the ins and outs of the middle platform and various forms of expression, as well as my understanding of enterprise platformization and setting up functional teams, you may still feel that it is quite abstract, so I think we must try to define the middle platform.

Why do I need a definition? It’s not for any other reason. I think I need to give myself a simple ruler so that I can easily remember the characteristics of the middle platform and use it to quickly judge whether a so-called middle platform meets my understanding of the middle platform. Therefore, my definition of the middle platform is:

Enterprise-level capability to reuse the platform.

It’s simple, are you a little disappointed? But in order to find a definition that I think is reliable, I spent almost two years. During this period, various definitions have emerged, but at least so far, I think only this definition is the most appropriate, simplest, and accurate. It can explain almost all the problems I have encountered about the middle platform, such as:

• Why are there so many mid-platforms?
• What is the difference between setting up functional teams and platformization?
• What is the difference between setting up functional teams and service-oriented?
• How to build the middle platform?
• ……

These 9 characters may seem simple, but what’s important is the interpretation of the value of “middle platform” behind them. Let me break them down for you one by one.

1. Enterprise level

Enterprise-level defines the scope of the middle platform. It does not mean that an enterprise can only have one middle platform, nor does it mean that a middle platform can only contain one enterprise. Enterprise-level represents the problems that the middle platform handles at the enterprise level, which means at least including multiple Lines of Business or serving multiple front-end products (teams). If a middle platform only supports one Line of Business or product line, it is not a middle platform, even if it uses service-oriented or Big data technologies.

Enterprise-level is very, very important. It made me realize that mid-platform building is not a technical issue, but a problem that needs to be elevated to enterprise architecture. When doing mid-platform building, we must step out of a single Line of Business and look at the business panorama from the perspective of the entire enterprise.

After thinking this through, my understanding of the middle platform has undergone a qualitative change, and I finally understand why using a system service-oriented approach to do the middle platform at the beginning would face so many problems, such as the most common issues of organization and management, as well as interest redistribution.

We can also see a trend in the rise and explosion of the mid-level platform, that is, more and more enterprises, whether they want to improve their own Operational Efficiency or meet the needs of business innovation and development, have accumulated the ability of enterprise global perspective and cross-Line of Business to an unprecedented strategic height.

2. Capacity

When it comes to the middle platform, the most commonly heard word is “ability”. This may be because the word “ability” is simple enough yet has enough tolerance and breadth.

The ability defines the main objects that the middle platform carries.

The abstraction of ability explains why there are so many types of middle platforms, and also explains why the middle platforms of each enterprise are different. The reason why different enterprises can coexist is because their core capabilities are different, which can meet the needs of users at different levels, that is, what we often call differentiated competitiveness.

3. to reuse

To reuse defines the core value of the middle platform and carries the evolution process from platformization to setting up functional teams mentioned above. Traditional platformization has not paid enough attention to “reusability” and “ease of reuse”, but more attention to how to eliminate duplicate capacity building, also known as “deduplicate”.

However, the proposal and rise of the middle platform reflects a trend of paying more attention to the user experience of front-end business. People shift their focus from the internal construction of the platform to the support of front-end business by focusing on “reusability” and “ease of reuse”. There is a shift from the perspective of governance to empowerment, from “deduplicate” to “to reuse”.

Deduplicate “and” to reuse “often appear together and are mentioned together, but they are talking about completely different things, with different purposes and difficulties.” Deduplicate “talks more about looking backwards and being technology-driven, while” to reuse "talks more about looking forward and being business-driven and user-driven. I think this shift in perspective is the key to understanding the concept of the middle platform, so

• “To reuse” is the goal that the middle platform pays more attention to.
• “Reusability” and “ease of reuse” are important metrics for measuring the quality of mid-platform buildings.
• Business responsiveness “and” business satisfaction "are important criteria for assessing the progress of mid-platform building.

This also explains why many Internet companies refer to the process of platform governance as the set up functional teams transformation process through business abstraction, configurability, and white screen transformation.

4. Platform

The platform defines the main form of the middle platform. Different from the traditional way of piecing together application systems, the middle platform realizes the flexible reuse of enterprise capabilities through the identification and platformization of finer grain capabilities, better supports front-end business, and meets the needs of quick response and reuse for business.

Although the definition of " enterprise-level ability to reuse platform " seems simple, after such a long time of practice and thinking about the middle platform, I think the meaning behind this definition is currently the most appropriate interpretation of the value of the middle platform.

• " Enterprise-level " defines the scope of the middle platform, distinguishing between service-oriented and microservices of a single system;
• " Capability " defines the main carrier of the middle platform, and the abstraction of capability explains the existence of various middle platforms;
• " To reuse " defines the core value of the middle platform. Traditional platformization has not paid enough attention to ease of reuse and front-end User Experience. The proposal and rise of the middle platform allow people to shift their focus from the internal design of the platform to the support of the front-end business through reusability.
• " Platform " defines the main form of the middle platform, which is different from the traditional way of piecing together application systems. Through the identification and platformization of finer grain capabilities, it realizes the flexible reuse of enterprise capabilities and better supports front-end business.

# Summarize thinking

Today we started with why enterprises need platformization and discussed why enterprises need to build a middle platform. Combining the content of the previous two lectures, I gave my opinion. I don’t know if it has inspired you, but let me summarize it again in the end.

User-centered continuous large-scale innovation is the core goal of mid-platform building. The user response ability and large-scale innovation ability of enterprises are the core embodiment of the comprehensive competitiveness of enterprises in the Internet era. Platformization, including setting up functional teams, is only a means to help enterprises achieve this goal, not the goal itself.

The construction of the middle platform (whether it is a technology mid-platform, Central Product Platform, or organizational middle platform) is fundamentally to solve the dilemma of enterprise responsiveness, to make up for the contradiction between the front-end that needs rapid changes for innovation-driven and the back-end that needs relatively slow change cycles for stable and reliable driving. It provides a middle layer to adapt to the speed problem of the front-end and back-end, connects and smoothly links the front-end requirements and back-end resources, and helps enterprises continuously improve user responsiveness as a whole.

Therefore, what the middle platform is is not important at all. The most important thing is to find ways to continuously improve the responsiveness of the enterprise to users. Platformization or setting up functional teams just happen to be on the right track.

Finally, we gave a definition to the middle platform, which is " enterprise-level ability to reuse platform ". With this definition, we can not only use it as a ruler to measure whether a middle platform is genuine, but also have a clear idea of how to build a middle platform.

Because if “the middle platform is the platform for enterprise-level capabilities to reuse”, then “set up functional teams” is also "the process of using platform-based thinking and means to sort out, identify, precipitate and to reuse enterprise-level core capabilities ".

Okay, from the next lecture, we will officially enter the landing section. Let’s take a look at how we plan and implement the middle platform.

Finally, I’ll leave you with a few thought-provoking questions:

• Does your company really need a middle platform?
• If you were to define the middle platform in one sentence, how would you say it?
• Is the middle platform the solution or the problem itself in your company?

As of now in 2019, if there are any hot topics in the IT industry, then the middle platform will definitely occupy a place. It’s like any company or news related to the middle platform will attract attention for a while. If something is not related to the concept of the middle platform, it seems to fall behind the times.

The middle platform has become one of the focuses of attention for a while.

But what exactly is Zhongtai? The answer is still unclear and vague, and many people are still pondering whether this concept has its own significance and whether it is another hyped concept, just a flash in the pan.

I sorted it out. In the past half year of 2019, not counting the ones I have read, I have collected no less than 300 articles related to the middle platform. Recently, for this course, I reviewed all the content again and found that each author has their own perspective and viewpoint on the middle platform in each article. Even now, the industry is still unable to completely unify and there are still different opinions. However, it is gratifying that there is more and more consensus in many aspects, and the concept of the middle platform is slowly fading away its mystery and coming to us.

And this is also the reason why I am interested in taking this course. In recent years, I have been working full-time on mid-platform related work. I have had the privilege of personally participating in the mid-platform construction projects of many large-scale group enterprises as an architect and product manager. Therefore, I hope to take this opportunity to organize and summarize what I have seen and learned, and try my best to restore a panoramic view of the mid-platform for you, helping you understand the essence of this concept and trend.

I have always followed a viewpoint that in order to understand a concept or technology, one must first go back to history, go back to the time of its birth, walk through its development process, and explore the background and reasons for its emergence.

As the saying goes, one must know the reason why things happen.

Today’s content serves as the beginning of the main course. I want to take you back to 10 years ago, to the starting point of the birth of the middle platform. Let’s walk through the birth and development of the middle platform together with the pace of time. Let’s see where the middle platform comes from and where it will go.

# 2008~ 2015 Keywords: Pregnancy

The rise of the middle platform is a trend. However, the concept of the middle platform was first noticed by everyone and must be considered as the middle platform strategy proposed by Alibaba. I don’t know if you have noticed this, but there is an interesting point, which is why I said that the middle platform was first noticed by everyone because of Alibaba, instead of directly saying that the concept of the middle platform was created by Alibaba?

Since our column is called “Explaining the Middle Office”, we can say a little more here. Many people believe that the term “middle office” was created by Alibaba. However, as of now, there are still some different opinions and views in the industry on this issue, because there have been front-end, middle-end, and back-end divisions in banks for a long time. Interestingly, when Alibaba introduced the Alibaba dual middle offices at the 2019 Alibaba Cloud Ali Cloud Aliyun Summit in Shanghai, the English translation also used the translation of “middle office” in banks. However, there is currently no official information from Alibaba to confirm whether this concept is related to the middle office concept in the banking industry.

But for you and me, whether the term Zhongtai was created by Alibaba or indeed refers to an existing concept in the banking industry, I think it is not that important. We can focus more on what this concept is and how it can help us in the context of IT.

As for Alibaba’s middle platform strategy, it is generally believed in the industry that it started with Jack Ma’s visit to Supercell in 2015. However, after following the information shared by Alibaba and actually understanding it with many colleagues from Alibaba, I learned that this visit can only be regarded as a prelude. Alibaba’s process of setting up functional teams had already begun before this. Therefore, to truly understand the background and reasons for the emergence of Alibaba’s middle platform, we need to go back even earlier, at least to 2008.

Because in 2008, with the adjustment of Alibaba’s strategy, Tmall emerged. However, due to its own characteristics compared to Taobao, there was a problem of duplicate construction between Tmall and Taobao at that time, which is now commonly referred to as the chimney-style system architecture.

The chimney-style system architecture has caused a lot of duplicate construction and resource waste. What should we do? The most natural idea is to integrate the duplicate organizations and systems. Therefore, the Alibaba Sharing Business Unit was officially born, responsible for platform transformation of the public parts in various front-end systems. After a painful exploration, it took the opportunity of Juhuasuan’s outbreak to truly establish the important position of the Alibaba Sharing Business Unit and sow the seeds of Alibaba’s large-scale platform strategy.

# 2015 Keywords: Alibaba middle platform strategy born

History is like a train traveling between ridges, everything is moving forward in a predetermined direction at a leisurely pace.

The new species of Zhongtai is also constantly nurturing with the passage of time, waiting for an opportunity to come.

In 2015, this opportunity finally arrived.

Next is the story that everyone is talking about: In 2015, Jack Ma led Alibaba executives to visit Supercell, the world’s most successful mobile game company located in Finland. You may feel unfamiliar with this company, but when it comes to the games developed by this company, I believe you have heard of them. Famous games such as “Clash of Clans”, “Island Raiders”, and “Cartoon Farm” are all produced by this game company.

However, what touched Jack Ma and Alibaba’s executive team at that time was that the company that gave birth to so many globally popular games had less than 200 employees. Each team responsible for a game had an average of only 5 to 7 team members. The team had sufficient freedom to decide what kind of product to develop, and then the public beta version would be launched as quickly as possible for the market to judge and verify the quality of the product. Once the product failed, it was quickly abandoned. At this time, there would be no punishment, but the team would raise a glass to celebrate and immediately make adjustments to continue to quickly find new directions.

Hmm, yes, this is the typical routine of Lean Startup.

However, in order for this mechanism to operate normally, there must be a prerequisite, which is that the construction time of the product must be short enough, and the cost of trial and error must be low enough. Only in this way can the team ensure that through continuous learning from failures and continuous iterative adjustments in a large number of trials and errors, the correct direction can be found as soon as possible, and the innovative and successful Progress Bar can move forward quickly.

What supports this mechanism is the common and universal game materials and algorithms that Supercell has accumulated over 6 years in the game development process. Based on these basic materials and algorithms like Lego building blocks, it is possible to support several small teams to quickly develop a new game in a few weeks, just like building blocks.

This approach touched the visiting Alibaba executive team, and this concept coincides with the “thick platform, thin application” architecture direction that Alibaba and the industry have been trying and conceiving for many years. It was this visit that strengthened the Alibaba management’s determination to adjust the organizational structure and accelerated the formal birth of Alibaba’s middle platform strategy.

Shortly thereafter, on December 7th, 2015, Zhang Yong, then CEO of Alibaba Group, stated in an internal letter: “Starting today, we will fully launch Alibaba Group’s 2018 middle platform strategy, and build a more innovative and flexible’big middle platform, small front desk 'organizational mechanism and business mechanism that is in line with the DT era.”

With this, Alibaba’s middle platform strategy is officially born, and the previous “thick platform, thin application” has also become “big middle platform, small front desk”.

Interestingly, when Alibaba’s mid-platform strategy was first proposed in 2015, it didn’t cause much of a stir in my impression. At that time, people were still talking about Hua Qian Gu and “My Skateboard Shoes”, while the internet circle was talking more about Internet + and O2O. Little did they know, a new battlefield had already begun to breed, and the seeds had been planted.

# 2017 Keywords: Emergence

With the birth of the mid-platform concept in 2015, after a silent but turbulent 2016, we gradually began to hear more and more voices about the mid-platform in the community in 2017. Alibaba and Didi Chuxing coincidentally began to share their experiences in mid-platform building, and books related to the mid-platform also appeared on the market.

The collective voice of internet giants has brought the concept of middle platform back into everyone’s view after two years.

As far as I know, many companies, whether they are internet giants or some strategic enterprises, are starting to re-examine and attach importance to this emerging new concept at this point in time.

Most of the earliest companies to start their own mid-platform planning and construction also began at this time. It was also at that time that I became familiar with the mid-platform, paid attention to and studied it, and actually participated in the actual mid-platform projects of each customer.

However, at that time, everyone faced many problems and difficulties. Companies like Alibaba or Didi Chuxing shared their middle platforms from the perspective of their own development process and problems. There is no doubt about this, but when everyone looks back at their own mid-platform building, each company’s situation is different, and each company’s problems are also different. What should their middle platform look like? How to build it? Many of these questions cannot be directly answered in books or sharing, and can only be explored bit by bit by themselves.

At this point, some brave pioneers began their respective mid-platform exploration journeys, although this was destined to be a path full of thorns.

Looking back now, some of the explorations at that time ultimately failed, while others are still ongoing. However, in any case, the explorations and experiences gained and lost by these pioneers have cleared many obstacles for our current understanding and construction of the middle platform. Without these explorations and reflections, there would not be the current experiences and consensus.

# 2018 Keywords: Full Outbreak

In mid-2018, Zhongtai had been accumulating momentum for a long time and finally ushered in a comprehensive outbreak.

This is something I have personally experienced. I remember it very clearly. It was at noon on September 9, 2018, when my colleagues were all taking a nap on their desks. Finally, I took the opportunity to post an article that I had been holding back for more than three months to my official account as usual. This article is about my thoughts on the middle platform over the past year. In my opinion, it is an ordinary article.

What happened afterwards far exceeded my imagination. My article was reposted by various community accounts in the afternoon, and the reading volume quickly exceeded 10,000, 20,000, 30,000, 40,000… For me, who usually only has a few hundred views when writing articles, I was a bit at a loss. At that time, I felt that all my friends, friends of friends, and friends of friends were reposting this article from the platform.

That was also the first time I felt the heat of the middle platform. It had been accumulating momentum for a long time and finally ushered in its own outbreak.

But only this one big discussion is not enough. You may still remember what happened afterwards. Come, let me quickly review it with you.

• On September 30, 2018, Tencent announced its largest organizational change in seven years, establishing the Cloud and Smart Industries Group (CSIG). At the same time, Tencent established a new technology committee and announced that it will build a technology mid-platform in the future.
• On November 26th, 2018, Alibaba announced an organizational upgrade. The Alibaba Cloud Ali Cloud Aliyun business group was upgraded to the Alibaba Cloud Ali Cloud Aliyun Intelligent Business Group, fully integrating mid-platform intelligence with Alibaba Cloud Ali Cloud Aliyun.
• On December 21, 2018, JD.com Group Human Resources released an announcement regarding the organizational restructuring of JD.com Mall. The announcement stated: “Under the new organizational structure, JD.com Mall will be customer-centric and divided into front, middle, and back offices. The middle platform will provide professional capabilities for front-end business operations and innovation.”
• ……

It is precisely because of the dazzling collective operation of this wave of internet giants that they have all spoken out for the relay of the middle platform, officially pushing the middle platform onto the stage and ushering in a comprehensive outbreak.

# The fog of 2019 still exists

The popularity of the middle platform exploded at the end of 2018 and continued into 2019, showing no signs of fading, but instead increasing.

However, despite the outbreak, it does not mean that the previous problems have been solved. The problems and confusion still exist, and have not become clear due to the popularity of the concept. Instead, as more and more companies follow suit, the problems increase instead of decrease.

• What is the difference between the middle platform and the platform?
• How many types of middle platforms are there? Which ones are not? Is there a construction order?
• How to build the middle platform? Where to start? How to calculate the end?
• Does the middle platform need organizational adjustment? How to adjust?
• How to verify the construction effect of the middle platform?
• ……

And these issues have been troubling me for the past few years. Now, after such a long time of exploration, research, and learning from predecessors, I have some of my own thoughts and summaries. Now, I am sharing these contents with you to explore the answers to these questions and help you clear some obstacles in your mid-platform building.

# Summarize thinking

Today, I reviewed the development process of the middle platform with you. Finally, let’s see if we can answer the question at the beginning: Why is the middle platform so popular?

Just as many trends are not formed by a single force, I think the popularity of mid-platform is at least due to the combination of the following four aspects.

1. The model effect of Internet companies. This is beyond doubt. At present, the model and benchmark effect of Internet companies, especially major companies, is still very strong. What’s more, the attitude of Internet companies towards the middle platform is so highly consistent that it is rare in the past. The construction effect is also really visible to everyone, which makes people envious.

2. Why are Internet companies so unanimous in promoting the mid-platform? There is a deeper reason behind it. This year’s industrial Internet is booming. In the mid-to-late stage of consumer Internet, the battlefield on the consumer side is becoming increasingly white-hot. Internet companies have turned their attention to the supply side in order to pursue sustainable growth. This is why ToB is also unusually popular this year. The cloud and mid-platform strategy is a very good entry point for Internet companies to enter traditional industries. Therefore, we see more and more Internet companies participating in the process of cloud migration and digital transformation of traditional enterprises, bringing their own technologies and practices to traditional industries. Throughout the process, the mid-platform is indeed a good breakthrough point and weapon.

3. As the saying goes, it takes two to tango. This timing does indeed coincide with the transition of some industries from systematization to platformization. Through years of informatization construction and accumulation, the information systems that should be built in the enterprise have also been built, such as ERP, CRM, etc. The enterprises that started informatization construction earlier also began to experience pain points such as chimneys and data islands mentioned earlier in their internal systems. And the enterprises that started informatization construction relatively later also want to take a shortcut through this mid-platform wave and achieve it in one step. At this time, a whirlwind of mid-platform struck, and every enterprise felt the need and pain points of doing mid-platform, and began their own mid-platform planning and construction.

4. Finally, I think there is another very important and fundamental reason, that is, the overall economic trend in the past two years has not been very good. Uncertainty and unpredictability are constantly impacting various enterprises and even industries, and the fear and anxiety of enterprise managers about the future development of enterprises have doubled. At this time, Internet enterprises use the mid-platform strategy to precipitate and reuse their capabilities, use certainty to cope with uncertainty, and have the ability and ideas of rapid trial and error and rapid innovation, which allows traditional enterprises to see a breakthrough direction. When the economic situation is good, everyone is either busy expanding their business quickly and doing what they can quickly, or they are stuck in their mature business, harvesting on time, without strong motivation to change. But when the economic situation is severe, the pressure and fear are becoming more and more serious. In order to maintain the survival and sustainable development of the enterprise in the future, or to prepare for the next round of improvement, the mid-platform strategy has become the choice of more and more traditional enterprises and industries, following the example of the Internet industry.

In short, the so-called favorable timing, geographical location, and human resources have brought together various factors to give birth to this wave of mid-platform hotspots. Is the popularity of mid-platform just a flash in the pan? Or will it become another important milestone in IT development like Cloud Service and microservices? It is still uncertain, but I personally believe that the latter has a very high possibility.

Finally, I also want to know why you are interested in the middle platform. Why do you think the middle platform is so popular? What is your biggest concern about the middle platform now?

Looking forward to fully discussing with you in the column. You are also welcome to share today’s content with your friends. See you in the next lecture!