#Phase 5: Performance indicators, who should define them? #
Looking back at the previous issue of " "Automation" is easy to talk about but hard to do ", we talked about how to create "automation", which is also a necessary condition for truly improving R&D efficiency. From single-point automation to improve efficiency, to full tool chain automation, and then to platformization of the R&D system, these are the three steps of automation practice.
In this process, how to establish a set of performance index system for the R&D system that not only has clear business goals but also can positively guide the R&D team is the key to the success of platformization. In this issue, we will talk about the issue of "efficiency metrics".
Interviewee A: Manager Li, agile coach of a cloud vendor in a foreign company
Interviewee B: Manager Su, R&D efficiency manager of an Internet bank
Q: Some R&D friends have reported that there are too many performance indicators, and some indicators do not know the meaning. Wondering where the metrics should come from? What is the logic behind defining the metrics?
A: I think so. With the development of enterprise business, the energy efficiency in the enterprise system is required to be higher and higher, and the method of judging energy efficiency is becoming more and more complicated. In the R&D system, the development of software engineering has enabled scientific methodology to improve R&D efficiency. In the process of implementing the methodology, "how to control the correctness of the process" is the fundamental logic for defining performance indicators.
The definition of R&D performance indicators should be based on the company's business goals, as well as the specific tasks and responsibilities of the R&D department. Different goals correspond to different results, some to improve delivery quality, some to reduce project R&D costs, the data of the corresponding key result indicators are quite different, and these data are only comprehensive results. For the definition of raw data, it is necessary to divide different levels of evaluation indicators for the structure and work functions of different teams, so as to ensure that projects, teams, and individuals can all follow the results.
Defining R&D performance indicators is a dynamic process. We need to conduct regular assessments and adjustments based on actual business conditions. Through regular performance assessments and feedback, we can determine whether the weights and goals of indicators need to be adjusted.
B: Ideally, performance indicators need to be established by the performance growth team and the business team. But most of the actual situation is that after the efficiency growth team is established, it seeks the evaluation of the business team, and then it is accused of unscientific definition and not in line with business goals, which leads to resistance. So, our internal practices are:
First of all, our efficiency growth team will set the indicators, let the indicator data run for a period of time, and do a good job of data accumulation.
Then, based on the accumulated data, we evaluate the effectiveness of the business to see whether the business team has achieved the relevant goals, or what hinders the achievement of the goals. In order to ensure the accuracy of the data, we need to split the dimensions of the data in order to better analyze the data rules.
Finally, regarding the performance evaluation results, we will communicate with everyone in the business team in a timely manner, so that they have a clear understanding and demand feedback under this indicator system, and let the business team know that the indicators are to help them rather than evaluate them , prompting it to take the initiative to join the definition of the indicator system to ensure that the indicators are of practical use to the business to help improve performance.
Q: When the R&D system is established, many performance indicators are accompanied. When these indicators are applied, they will become the KPI numbers pursued by the team, so they only care about the good-looking numbers, and the performance has not been substantially improved. Is there any unreasonable performance indicators? Condition? How to treat this situation?
A: In fact, unreasonable indicators are common. Therefore, we continue to optimize performance indicators, constantly measure, improve, measure again, and improve again. The entire indicator system is actually an intuitive reflection of the operation of the R&D team, just like the dashboard on the car, giving the driver an intuitive data to determine whether the speed of the car is fast or slow, how much fuel, how much electricity, and where there is a fault Wait, help the driver to judge the condition of the car.
Many times, the team is under the pressure of KPIs and has to pursue some indicators too much, but from the actual effect, it is not helpful to the business goals, which is completely deviated. We also use the car as an analogy. If our goal is the speed of the car, there are many ways to achieve this goal: we can change the engine, change the tires, modify the body, change the suspension system, and even change the driver. However, the indicators pursued by the team or individual may become: the speed of fuel consumption, the efficiency of the valve, and the temperature of the radiator, and these data can be tampered with or falsified. Just like many years ago, we regarded the number of lines of code, working hours, etc. as important indicators of the R&D team, and there is a lot of water in it. But now we replace unscientific indicators with more scientific indicators such as valid code, thousands of lines of code defects, and the number of required deliveries.
B: I think so too. Unreasonable indicators are hard to avoid. The R&D team is a department that is difficult to quantify, and the complexity is extremely high. Unreasonable indicators are generated during the process of establishing a quantitative system. After verification of unreasonableness, they also need to be eliminated quickly. Bill Gates also said: Measuring the progress of programming by the number of lines of code is like measuring the progress of aircraft manufacturing by weight. This is to say that unreasonable indicators and business goals are out of focus.
On the issue of pursuing good-looking KPIs, these indicators are often process indicators. My suggestion is to give business teams autonomy and let them decide whether to use them, rather than blindly assessing them. The purpose is to allow business teams to participate in indicator construction. The purpose of defining process indicators is to play a role in finding and solving problems, and it can also reduce the phenomenon of chasing data. If an indicator cannot guide the team to improve quality, there is no need to measure it.
In this issue, we chatted with two guests about "defining performance indicators" and "unreasonable performance indicators". The most profound sentence: "If an indicator cannot guide the team to do quality improvement, there is no need to Measurement." It is regarded as an indicator of KPI, if it cannot play the role of business goals, it is an unreasonable indicator.
The three steps to establish the performance index system form a closed loop:
-
The first batch of indicators is defined by the effectiveness team, which is used to accumulate preliminary data.
-
Through the achievement of business goals, analyze the data to evaluate the performance.
-
By evaluating the performance results, communicating with the business team, confirming whether there are problems with the performance indicators, improving and optimizing, and allowing the business team to participate in the establishment of the performance indicator system.