Performance measurement is not an unfamiliar topic for DevOps R&D teams. All R&D teams want to improve the efficiency of the R&D team through performance measurement and open up the closed loop of R&D management.
Regarding software R&D effectiveness measurement, you can follow the "Software R&D Efficiency Measurement Specification" standard. In the E³CI software R&D effectiveness measurement framework, E³ refers to the combination of Efficiency efficiency, Effectiveness effect, and Excellence, C refers to Cognition, and I stands for Improvement Improve. E³ is the core of R&D efficiency. The combination of cognition and continuous improvement supports R&D efficiency and improvement, namely: efficiency = cognition + improvement. The E³CI metric set runs through the entire R&D process, covering the end-to-end R&D process, including the definition of R&D performance from all dimensions of delivery value, delivery efficiency, delivery quality, delivery cost, and delivery capability.
The goal of R&D performance measurement is to help DevOps R&D teams understand their own work efficiency and quality, discover problems in a timely manner and take corresponding measures to improve, so as to achieve efficient and high-quality product delivery, meet business needs, and realize business value.
- Identify bottlenecks: By measuring key indicators in the R&D process, bottlenecks and shortcomings in production efficiency can be identified in a timely manner, so as to optimize the R&D process in a targeted manner.
- Monitor progress: R&D effectiveness metrics can help teams monitor project progress in real time, identify problems and risks, and take timely measures to resolve them.
- Improve efficiency: By measuring workload and work efficiency, the team can identify inefficient links in the work process and take corresponding measures to optimize and improve work efficiency.
- Improve quality: By measuring indicators such as product quality and test coverage, product quality problems can be found, repaired and improved in time, thereby improving product quality.
Simple "5 steps" to complete R&D performance measurement
To efficiently improve the efficiency of R&D operation and maintenance, and do a good job in DevOps system construction, it is impossible to achieve without reliable measurement. The essence of performance measurement is to evaluate the speed and quality of value flow. Therefore, it is imperative to do a good job in R&D performance measurement.
So how to measure the effectiveness of the R&D team? Then focus on indicators and divide them into five major steps.
- Determine indicators: The team and enterprise managers should select appropriate indicators for measurement according to project requirements and team conditions.
There is a structured indicator system in the industry, which corresponds to different indicators from different stages such as requirements, design, development, testing, release, and operation and maintenance. But it needs to be reminded that the establishment of an indicator system is not as complete as possible, but according to the different stages of the team and the different problems encountered. The most important thing is to solve the problems through indicators, and then carry out the design and system of indicators build. Effectiveness measurement can use various indicators, which can be divided into quality, efficiency, cost, customer satisfaction and so on.
- Set goals: The team and business managers should jointly set goals, such as the delivery frequency of each iteration, product quality indicators, etc. Goal setting must follow the SMART principle: Specific (specific): the goal should be specific and clear, able to clearly communicate the desired results, and avoid vague and ambiguous descriptions; Measurable (measurable): the goal Should be measurable, can be visualized using data and metrics, track progress, avoid subjectivity and unquantifiable goal setting; Achievable: Goals should be achievable, taking into account reality and availability Resources, to ensure that goals are achievable, challenging and motivating; Relevant (relevant): Goals should be relevant, consistent with business needs and strategies, able to have a positive impact on the achievement of business goals, and avoid meaningless goals Setting; Time-bound: Goals should be time-bound, with clear deadlines and timelines set to ensure that goals are achieved and tracked in a timely manner.
- Collect data: The team and business managers need to jointly collect data, such as the number of defects per iteration, delivery time, etc. There are many ways to collect data. You can use some tools to obtain the data required by the indicators, or you can directly connect to the database through scripts to obtain the corresponding data. When the Zhongan team built the performance measurement cockpit, they used the Jizhi BI tool to pull data from the data source, and then processed it for visual configuration.
- Analyzing data: Teams and business managers need to analyze the collected data to find problems and directions for improvement.
The improvement of software R&D efficiency is complex and affected by many factors, and there is a correlation rather than a causal relationship between factors and results. Even if we find an association between two sets of data, it doesn't mean that one set necessarily leads to the other. For example, if a team's "code technical debt ratio" indicator is high, it generally means that many problems in the code have been temporarily shelved, and the cost and technical risk of continuous maintenance in the future will be high. , it is very likely that the indicator of "lead time" will continue to increase, that is, there is a correlation between the two sets of indicators. But this is not an inevitable causal relationship. Although there is a lot of technical debt, it is very likely that other factors such as personnel capabilities and unannounced overtime temporarily cover up the problem and counteract this trend on the surface.
5. Feedback and improvement: The team and enterprise managers should feed back the analysis results to the team members, discuss the direction of improvement and optimize together. R&D managers should conduct down-to-earth analysis by measuring the index data of the broader market. First, analyze and interpret the data from multiple perspectives to obtain effective insights; then combine other related indicators and investigation methods to investigate the root cause, locate performance bottlenecks and optimize opportunities; Ultimately, these insights will be turned into clear, executable, and verifiable improvement plans to standardize the R&D process and establish a good R&D culture. Performance improvement cannot rely on staged sprints. To achieve effective and sustainable performance improvement, it is necessary to integrate the practice of measurement and improvement into the daily R&D process, continuous tracking, and continuous improvement.
In the context of the digital age, information technology is the key to driving the development of enterprises, and R&D efficiency has become the core competitiveness of enterprises. Adhere to data-driven, through the correct performance measurement method, can make R&D performance quantifiable, analyzable, improveable, and enhanceable.