Improve quality and efficiency, reveal the "value growth flywheel" of large software teams｜Review of the live broadcast

On August 24th, the third course of ONES R&D Management Master Course officially started. Feng Bin, co-founder and CTO of ONES, shared "Methods and Practices for Efficient Collaboration of Large Software Teams". Using the concept of the "value growth flywheel", he dissects how large software teams can achieve rapid growth in business value.

The following is the core content shared by Mr. Feng Bin.

Problems in the process of software R&D collaboration

Let's first analyze the problems that often arise in the process of software team collaboration.

The figure below shows the reasons behind the problem. Different colors represent different reasons. For example, dark blue represents quality problems, and green represents changes in requirements.

On the whole, how to improve the target matching rate, improve work efficiency, ensure quality, and reduce rework rate are the issues we are more concerned about when doing R&D management. To sum up in one sentence, we hope to generate business value better and faster while ensuring efficiency and improving efficiency.

The Value Growth Flywheel of Large Software Teams

So, how to ensure efficiency while producing business value better and faster?

Software teams, especially large software teams, in the process of collaboration, to ensure the efficiency and value output of the team, three important management methods are needed—target management, organizational management, and end-to-end digital project management. Known as the "value growth flywheel" of large software teams. Next, I will give an in-depth explanation of the "value growth flywheel".

Target management:
The team needs to form a consistent cognition and consensus on the goal and focus on the goal. If the goal deviates too much, no matter how good the execution is, it will be difficult to produce value. So how should managers set goals? There are several steps that cannot be ignored:

Think globally, look down on the present, and plan for the future. For example, in an agile process, we not only express what to do, but also express business value.
Set a five-year direction, make a three-year plan, and implement a one-year goal. We not only look at short-term goals, but also adhere to long-termism.
Prioritize the goal of "increasing business value", and plan the goal of "reducing cost" later.

Organization management:
The typical organizational structure is divided into "functional" and "business". Different organizational designs reflect the organization's resource allocation and team collaboration methods. Functional type, such as front-end group, back-end group, test group, product group, is the characteristic of functional organization structure; business type is more concerned with the closed loop of business, for example, a function requires product, R&D, testing, operation and maintenance, etc. Only the collaboration of multiple teams can successfully release. In order to achieve the same goal, these roles are put into one team, which is the business organization structure.

It should be noted that the "functional" and "business" organizational structures are not static. Depending on the specific situation of the team, there may be two organizational structures at the same time. When the team size increases by more than 50%, managers should make a decisive decision and adjust the organizational structure; at the same time, the design of the organizational structure is also grayscale, and the two structures can be mixed. In the communication with many customers, I found that most of the team organizational structure design is grayscale.

Strengthening team culture is equally important. The key is to clearly define team values and be able to implement them from top to bottom. The same values and small differences in behavior are the basic guarantee for collaborative efficiency.

End-to-end digital project management:

The project management of large teams often encounters problems such as unaligned goals, lack of team innovation, and cumbersome processes, which need to be solved by establishing an end-to-end digital project management solution.

Target alignment:
Round nine to one, and do the task with the highest priority. We often joke that the world is so big and there are so many needs, it’s impossible to finish it all. When we really align our goals, we will find that many lower-priority requirements are not necessarily needed. Therefore, we must focus on the goal and find the highest priority task to do. Only by focusing on the highest priority goals can the maximum resources be invested. The hardest thing about "focusing on goals" here is that even if there are 10 things that are valuable, we need to find the 1 most valuable thing, discard the other 9, and do the task with the highest priority.

Using quantitative priorities, all tasks are sorted. In practice with the ONES project management tool, we order tasks by priority. After sorting, the top task is the highest priority and should be completed first. In this case, the priority of the task can be shared with the team members and aligned.

Team innovation:
Provide members with contextual information and provide the ground for innovation. Only when we share information, so that each member has contextual information, they have the opportunity to think about how to better solve the problem through this information.

Information transparency brings a sense of security and stimulates initiative. When communicating, it is unavoidable that there will be a lot of new news. Therefore, if you don’t know, you won’t dare to say it; if you don’t dare to speak, you won’t dare to think; if you don’t dare to think, you won’t be able to make innovations. The sense of security brought by information transparency is very important. When our information is aligned, the sense of security will naturally deepen.

The precipitation of information and knowledge improves the efficiency of collaboration. Through the software of online collaboration, after depositing our knowledge, we can read it at will, which helps to improve the efficiency of collaboration.

Process standardization:
Workflow standards are difficult to implement. The reasons can be summarized as follows: high execution costs, process standards limit innovation capabilities, and the cycle for team members to develop new habits is too long. So how to solve it? There are the following ideas:

The razor principle: Streamline processes and shave off unnecessary processes.
The lower limit of the process standard guarantee, such as 70%, does not require precise control.
Solidify process standards into tools and automate them.

So how exactly do we implement the standard?

• The first step is to establish a standard document, which is the foundation.
• The second step is to create a template based on the tool and solidify it into the tool. The following figure shows the workflow of the ONES system when processing a certain type of work order. As you can see, the flow of each state requires the fulfillment of different roles and quasi-exit standards before reaching the next link. This is the so-called solidification of the standard process into the tool.

• The third step is process automation and large-scale application. For example, after submitting the code, the task status is automatically updated; the status of the parent task is automatically updated according to the status of the subtask, etc., which can greatly improve the execution efficiency of the standard process.
  

keep improve:
Quantitative feedback is the premise of effective review and improvement, and sufficient project data is the premise of quantification. Through the one-stop R&D management tool, statistics and analysis of the performance data generated by the entire R&D process can more objectively and scientifically measure team performance, review and improve tasks, and prioritize them into the task list to achieve continuous circulation Improve.

The following figure is a very classic PDCA cycle diagram. When we go to review related improvement tasks, we can import it into the to-do list; determine its priority in the to-do list, and finally form a loop to solve the problem.

To sum up, the three "flywheel blades" of "target management", "organization management" and "end-to-end digital project management" are interdependent and mutually advancing relationships. When the three blades are promoted at the same time, the business value of a large team can be rapidly increased and business growth can be achieved.

- Q & A -

Q1: The teacher mentioned goals when talking about the value growth flywheel. So for large software teams, should the team's goals be formulated bottom-up or top-down?

Teacher Feng Bin: Different decisions should be made in different situations. For example, when there are many determinants, it is more efficient to determine the goal from the top down; if there are many uncertain factors, the top management has ideas, and at the same time, the grassroots employees also provide feedback and suggestions according to their own ideas, and finally The two sides are aligned again. Therefore, a bottom-up approach may be more effective.

In fact, in many software teams, you will see a combination of the two methods. Just like OKR, managers have their own thinking and pass it down, and grass-roots employees also have innovative ideas. It is a better practice to align and integrate both sides.

Q2: Our company is a large group with a lot of application software, and we always buy new software. How to solve the problem of isolated information islands in this process?

Teacher Feng Bin: One of the reasons for this problem is that the coverage of these systems is relatively small, and it can only solve a small problem. In addition, the development ability or callback ability is not strong enough, and so on.

How to solve it? two methods. First, we have to find a complete set of solutions. At the beginning, the one-stop R&D management tool made by ONES was valuable, because in the R&D process, many problems were connected. Therefore, a one-stop solution can better bring information together and achieve the level of digitization; second, the product must have a certain degree of openness and rich API interfaces. With these tools as a foundation, you can solve the problem of data silos and manage these data, model and collect valuable data.

Q3: Some time ago, our company undertook a software project, and the development team also developed it according to the required specifications, and completed the realization of all functions. But the customer said in the process of acceptance that this is very far from the product they want. How should this problem be avoided?

Teacher Feng Bin: We should give customers a usable product as soon as possible through a similar agile method, and then communicate and revise after getting feedback. In the process of developing this usable product, priority also plays a big role. What to do first and later must be sorted. In fact, no one can accurately match the initial requirements with the final product. It is only through every step of iteration, communication, and revision that a valuable product can be delivered.

The first season of ONES R&D Management Master Class is still going on. Courses such as "Industry Trends and Five Improvements of R&D Efficiency Measurement" and "Practice of Improving Efficiency in Top 500 Enterprises" will be launched soon.