Research and development efficiency improvement practice of large insurance groups with 10000 + code base and 3000 + R & D personnel

This article sums up 7354 word , Suggested reading time ：14~15 minute .

‍

Read this article and you will get ：

1、 The overall improvement of R & D efficiency in the financial industry

2、 Pain points of improving R & D efficiency in the financial industry ;

3、 Experience in the practical process of improving R & D efficiency ;

4、 Research and development efficiency improvement system methodology ;

‍

Preface ： This case comes from A large insurance financial services group , It has many subsidiaries . In recent years , Cloud native 、 big data 、 Cutting edge technologies such as artificial intelligence are widely used in product innovation in the insurance industry 、 marketing 、 Management and other aspects . The R & D scale of the group has also grown rapidly , Currently reached The code base 10000+, R & D team members 3000+.

‍

01 Overview of R & D efficiency improvement project

‍

Project objectives

2020 In, the group proposed , Focus on business , customer-oriented , Through continuous improvement of R & D efficiency , Fast 、 High R & D quality continuously delivers the closed loop of effective value .

‍

Project strategy

In order to be on a large scale 、 The group with complex levels should maintain the unity of research and practice ideas , Avoid making wheels repeatedly , The strategy adopted by the group is ：

‍

• At the Group organizational level , Coordinate the construction of R & D efficiency management system , Output influence to front-line R & D teams
• At the level of front-line R & D team , Work closely with the organization level , Precipitate excellent solutions , The overall situation will be driven by the pilot , Realize the double improvement of R & D efficiency and R & D quality

‍

This case covers several key practices in the early stage of the group's R & D efficiency construction ：

‍

• Organizational restructuring ： Ensure that the organization and front-line R & D team are in the same direction and step
• Pilot research ： Dig deep into problems in the front-line R & D team , Verify commonalities at the organizational level
• Construction of R & D efficiency measurement system ： Improve the opacity of R & D process and products , Implement a quantitative management system applicable to the whole organization , Ensure the implementation of the lower limit
• Explore and improve practice ： Deep pilot , Explore the direction in the real practice of the front-line team , Improve the upper limit of R & D efficiency through systematic improvement

‍

initial results

‍

After careful research 、 Continuous verification 、 Steady progress , The group has achieved phased results in the early stage of R & D efficiency construction ：

• Establish R & D efficiency measurement system and improvement strategies , R & D performance data is automatically deposited on the data platform , And present scene charts to different roles , Realize organization level research efficiency digital governance
• R & D efficiency ： The team reflects that the degree of collaboration has improved ; From the data point of view , Per capita R & D efficiency （ Code equivalent ） And the on-time online rate of demand shows a continuous upward trend , Demand throughput tends to be stable .

Demand delivery trend analysis （ The code equivalent index comes from the code analysis system ）‍

• Research and development quality ： Take the improvement department as an example , comparison 2021 Annual data , One time smoke passing rate increases 9%, The proportion of serious defects decreases 33%, Defect density （ Weighted total number of defects / Code equivalent ） Reduce 57%.

Improve the quality and efficiency trend correlation analysis of the Department

02 Pain points of improving R & D efficiency in the financial industry

‍

When starting R & D efficiency improvement , There are mainly the following pain points ：

‍

R & D efficiency

• The workload of R & D personnel is difficult to quantify , Lack of data for management
• The granularity of demand varies , The reliability of demand related R & D efficiency data is low
• Long demand lead time , The on-time online rate is low

‍

Research and development quality

‍

• All links of R & D are subject to delivery time squeeze and cost pressure , R & D quality control is only carried out in the final test link , The cost and pressure are huge , Leave hidden dangers in R & D quality
• Some R & D teams suggest that the R & D quality is unstable
• The average proportion of serious defects is high
• Some details of the requirements implementation are incomplete , Affect user satisfaction
• According to the root cause analysis of previous production problems , Code design and version release class roots account for a large proportion
• The construction progress of R & D quality management of each subsidiary is different , Lack of a perfect and unified R & D quality management system （ Including role configuration 、 Make a plan 、 Supervise and implement specifications 、 Assessment and evaluation mechanism ）; Although all subsidiaries have R & D efficiency metrics , But there is a lack of unified norms .

‍

03  Practical experience of improving R & D Efficiency ;

（1） Preliminary construction

‍

Before implementing this round of improvement , The construction of R & D efficiency of the insurance group mainly focuses on the construction of tool chain and the measurement of testing links .

In terms of tool chain , Self built + Introducing tools , Build coverage requirements 、 Design 、 Development 、 test 、 The whole process tool chain of the release link . The high degree of instrumentalization provides basic support for the availability and effectiveness of R & D efficiency data .

Test link measurement aspect , Previously, we mainly focused on the following two aspects ：

‍

• Test link R & D efficiency ： Task closure rate 、 throughput 、 Task closing duration 、 Task completion rate on schedule
• R & D quality trends , Include ：
• One time smoke passing rate
• Number of defects 、 Defect severity distribution

‍

After a period of implementation , Although the quality of software research and development has been improved to a certain extent , But it is difficult to promote the overall continuous improvement through testing a single link . In order to strengthen the research and development process, research and development quality , The group began to implement the test left shift , The tester is QC towards QA The transformation of , Cooperate with the R & D team to establish the R & D quality management baseline throughout the whole process .

‍

In order to coordinate the construction of R & D quality management system within the organization , Avoid making wheels repeatedly , The group established a research and development quality management group directly under the Group .

‍

meanwhile , R & D quality is closely related to R & D efficiency ： One side , The low quality of R & D will cause the R & D team to be busy fighting the fire or repeated redoing , Reduced R & D efficiency ; On the other hand , The improvement of R & D quality cannot be at the expense of R & D efficiency . therefore , The R & D efficiency management group, also directly under the group, began to coordinate the measurement and improvement of R & D efficiency .

‍

Both the R & D quality management group and the R & D efficiency management group belong to the organization level R & D efficiency special team .

‍

（2） Organization building ： Double layer structure , Ensure that the direction and pace are consistent

As mentioned above , The group has set up an organization level R & D efficiency special team . In order to ensure the implementation of R & D efficiency practice in daily R & D , Establish organization level + Team level double-layer structure , The R & D efficiency side is managed by the R & D Efficiency Management Group + Team level R & D efficiency leaders are jointly responsible , The R & D quality side is controlled by the R & D Quality Management Group （ Organization level QA） + Team level QA Share responsibility for .

The figure below takes the R & D quality side as an example , Introduce the division of labor of double-layer structure ： The R & D quality management team undertakes research 、 As a whole 、 To guide the 、 Support expert responsibilities ; Team level QA Then undertake the implementation of R & D quality system in its team 、 Control process R & D quality 、 Responsibility for continuous improvement of practice .

The two work together , On the one hand, it enables the special team of R & D efficiency to maintain a relatively objective and neutral external perspective ; On the other hand, it ensures that the strategies related to R & D efficiency are communicated to all teams in a timely manner , Specially assigned person is responsible for implementation , The goals of all parties are aligned , Act in unison .

Schematic diagram of double-layer structure on the R & D quality side

‍

（3） Pilot strategy ： Point face combination , Verify commonalities

‍

Through investigation and research , The group adopted the starting from pilot 、 Point face combination 、 Gradually promote the construction strategy of R & D efficiency .

‍

（4） Key pilot + General research , Location problem

‍

In the design idea of the pilot program , It is divided into two categories: key pilot projects and general research .

‍

The role of the key pilot is to identify problems . The organization level R & D efficiency special team will cooperate with the front-line R & D team , In depth interview and research . As a result of the discussion , Select key pilot samples according to the following conditions ：

‍

• Key business ： The pilot team itself is in a strategic position , Or the system is the core system within the Group , Business needs continue to enter , High requirements for R & D efficiency
• Excellent performance in R & D efficiency ： Conduct measurement pilot in the team with better R & D efficiency , Such teams are generally more open-minded , Can give an objective view of the measurement system 、 Constructive feedback
• There are demands for improvement in R & D quality ： From the data analysis, it is observed that there is a downward trend in the near future or a large fluctuation in the long term , After preliminary analysis , There are certain R & D quality risks in the R & D process
• Subjective will ： The team pays attention to , The superior is willing to do quantitative management

‍

The role of general research is to verify whether the problem is universal . therefore , The sample selection should cover all types of teams and systems ： Construction 、 Different stages of maintenance ; agile 、 Steady state different R & D modes ; Size version 、 On demand version 、 Emergency version and other different distribution types .

‍

（5） In depth interview and research , Check the current situation of R & D efficiency

‍

In key pilot projects , Conduct in-depth research through the following methods ：

‍

• It is necessary to extract several samples of different sizes , View the key activities and phased work products of the whole R & D process . Take the R & D quality side as an example , Work products include demand communication / Disassembly record 、 Business / System requirements document 、 Design document 、 Test execution record 、 Production problems 、 Defect list and other inputs / Output document . According to the inspection items formulated in advance , Check whether there is any defect in the work product .

‍

Take the R & D quality side as an example , Various work products

• On Project Management 、 demand 、 Development 、 test 、 Interview relevant members in the release session , Around the previously extracted sample demand , Ask members to demonstrate the process and key work , And invite members to give open feedback on suggestions on R & D quality .
• Observe the communication of business needs pertinently 、 review 、 Assembly line operation and other key R & D activities

‍

Based on the above research actions , Sort out the research and development process of the pilot team 、 Business side characteristics 、 Team work habits 、 Personnel structure and other information , Summarize the excellent practice and improvement opportunities of each link , And reach a consensus with the pilot team .

‍

After collecting information from the pilot team , Analyze which problems occur frequently in multiple pilot sites , And verify whether these problems have commonalities through general research . In the Research Report , Present the current situation of the problem with quantitative data , And give the problem details and corresponding excellent practice examples .

‍

04 R & D effectiveness measurement system ： Step by step from less to more

‍

（1） Select a few indicators , Meet common needs

‍

After collecting high-frequency and influential key issues from pilot and research , The organization level R & D efficiency special team began to gradually establish a R & D Efficiency Measurement System , Start with a few indicators , Take quantitative data feedback as the starting point of management , Drive improvements .

‍

R & D efficiency

‍

The invisibility and unpredictability of R & D efficiency is the most fundamental problem at present , This problem directly leads to the lack of basis for management , Difficult to carry out .

‍

Because the practice of each R & D team in the demand link is different , It is difficult to take the demand related indicators as the R & D efficiency indicators uniformly used by the organization . The traditional indicators of workload such as the number of lines of code are vulnerable to noise , The effectiveness of measurement is difficult to guarantee .

‍

The R & D efficiency management team has conducted research , The code equivalent of simeyi is selected to measure the coding workload of developers . Compared with the number of code lines and other indicators , Code equivalent can effectively count the logical workload contained in the code , Exclude code styles 、 New line habit 、 notes 、 Formatting operation and other factors interfere .

‍

Research and development quality

‍

The whole process of R & D quality assurance is the most critical task at present . In order to cooperate with the practice of testing left shift , On the basis of existing indicators , Added defect density index , Drive the team to achieve an appropriate balance between output and R & D quality .

‍

（2） Special promotion , Strive for consensus

‍

Getting the R & D team to embrace measurement is not easy . When strange concepts appear in the measurement system , It is more likely to arouse doubts .

‍

In order to win the support of R & D teams of subsidiaries “ Code equivalent ” Consensus on such new indicators , The organization level R & D efficiency special team carried out special promotion ：

‍

• First step , Introduce the R & D efficiency measurement concept to the core backbone members of the subsidiary 、 The principle and advantages of code equivalence , Get feedback , Further promotion by key members , At the same time, special Q&A link
• The second step , Cooperate with the capacity data and special analysis reports regularly published online , Conduct offline research ; Combined with the actual situation of the R & D team , Find specific problems , And promote the improvement of management measures
• The third step , Summarize the practical experience in the promotion process , Further promote and apply to all subsidiaries ; At the same time, in the whole group 1024 Programmer's Day activities , Apply multi-dimensional indicators of R & D efficiency and R & D quality , Conduct comprehensive evaluation and analysis ; Based on R & D efficiency data , Invite internal and external experts to discuss and write code

‍

Through the promotion strategy from point to surface 、 Open communication attitude and positive operation actions , Gradually obtain the recognition of all R & D teams .

‍

（3） Establish organizational baseline , Ensure the implementation of the lower limit

‍

The main goal of quantitative management is to reasonably raise the lower limit , Reduce the problem of low R & D efficiency caused by subjective factors . The quantitative management practice of the group is to carefully select indicators , Reasonably set the baseline , And bring it into daily management ; Each front-line team can set the internal baseline according to the actual situation , The lower limit of internal baseline should not be lower than the lower limit of organization level .

‍

While ensuring the implementation through the organization level baseline , Give the R & D team room for flexible adjustment , Encourage the team to maintain a positive attitude while , Respect objective constraints 、 Step by step .

‍

R & D efficiency

‍

After obtaining the consensus of the team , The R & D efficiency management team selects code equivalent as the workload indicator , And set the R & D efficiency baseline of the R & D link .

‍

There are many R & D teams in the Group , Their workload is affected by the business stage 、 Business types 、 The number of Posts and demand cyclical fluctuations and other complex factors , Setting the baseline across the board is bound to be unreasonable . So how to respect the objective fact that the R & D team is diverse , The practice of R & D efficiency baseline can be effectively promoted to the whole group ？

‍

The R & D efficiency management team analyzed a large number of historical data , After fully communicating with the front-line R & D team , It is considered reasonable to set only the lower limit of organization level , The value is the current per capita code equivalent 30%. The lower limit is designed to locate extreme zero production and low production , To urge team level improvement , It is not used for personal R & D performance appraisal .

‍

When the workload of a research and development team is lower than the baseline, more than a certain proportion of members 、 Or when the workload trend of the team declines significantly , Then ask the team to start self-examination , Explain why . Experts from the R & D efficiency management team will work with the team to analyze , Discuss the improvement plan .

‍

Research and development quality

‍

Based on the common problems found in previous research , Combine the index recognition and the convenience of tool support , The R & D quality management team sets the defect density as the first R & D quality baseline index .

‍

Baseline indicators will be gradually incorporated into the team assessment system , Guide the team to pay attention to R & D efficiency and quality , But not linked to personal performance .

‍

What needs to be noticed here is , Setting baselines from top to bottom is a management tool , Can we continuously improve our requirements , Until the R & D efficiency is improved ？

‍

The answer is No . The limitations of quantitative management objectively exist . A considerable part of the problem of R & D efficiency is caused by systematic and objective factors , Management means and systems cannot solve this part of the problem , It can only be solved through the improvement of engineering practice . If unlimited demands are made to improve R & D efficiency with subjective initiative , Become essentially “ Internal volume ”, Even at the cost of developer overload, it has achieved temporary results , It's not sustainable .

‍

（4） Continuous improvement , Reverse drive tool and practice improvements

‍

R & D efficiency

‍

In addition to using code equivalent indicators to measure resource R & D efficiency （ Value producer perspective ） Outside , The R & D efficiency management group also focuses on business oriented delivery R & D efficiency （ Value receiver perspective ）, Including demand 30 The sky line rate 、 On time delivery rate 、 Demand lead time, etc .

‍

Because the practice of each R & D team in the demand link is different , The difference of demand granularity will affect the validity of data , The delivery R & D efficiency index related to the demand will not set the baseline requirements for the time being , Each team puts forward the target value according to the actual situation . The R & D efficiency management team will regularly observe the achievement of goals and month on month trends , Pay attention to the teams with large fluctuations or downward trends in the delivery R & D efficiency index .

‍

meanwhile , Put forward requirements for the upper limit of granularity of requirements from the organizational level , And configure the threshold in the tool , Guide the R & D team in demand analysis 、 The review process is strictly controlled , Split requirements into reasonable granularity .

‍

Research and development quality

‍

Beyond the defect density index in the R & D stage , The R & D quality management team hopes to expand the R & D quality measurement to the whole process , Gradually cover different stages of the R & D life cycle , And under the following four guiding ideas, the panorama of R & D quality management baseline is designed .

‍

• The version is rhythmic ： combination “ Version train ” Mechanism , Specify the completion time of key nodes in the iteration , Set the version freeze point , Use the time after the version is sealed to conduct the production readiness review , Strictly control the time rhythm
• Design Standardization ： Define different types of design templates and content requirements , Prevent design omission
• Standardization of Review ： Strictly implement the review requirements of each stage , Do a good job of pre Review , Keep review records , Review problem tracking closed loop , Improve the effectiveness of the review
• The whole process of testing ： Test shift left , Participate in all links of R & D , Control the handling of abnormal extreme situations from the perspective of testing , Reverse scenario considerations ; Test move right , Assist the business in acceptance testing , Acceptance problem tracking closed loop

‍

With the automation of the tool chain and the recognition of the team for R & D performance data are further improved , The R & D quality management team will incorporate more indicators into the baseline , Lead the R & D team to pay attention to R & D quality assurance from the perspective of the whole process . According to the measurement requirements of key activities , Form tool support requirements , Reverse drive the tool chain for continuous improvement .

‍

R & D quality management baseline panorama

05 Practice of improving R & D efficiency

‍

R & D Efficiency ： Explore fine management in multiple directions

‍

In terms of R & D efficiency , The current strategic focus of the R & D efficiency management group is to maintain the baseline 、 Guarantee execution first . besides , The R & D efficiency management team also encourages front-line R & D teams to spontaneously explore data-driven refined management .

‍

The team spontaneously conducted research and effect practice pilot

‍

Within the front-line R & D team , Because the data is more comparable 、 The context information is more complete , part team leader Trying to set our team threshold , Put forward higher requirements for R & D efficiency , And carry out research and effect practice pilot . Practice includes ：

• Management
• encourage Top： For developers with high production capacity , Conduct positive advocacy within the team , Share code and work experience
• Focus on Bottom： For developers with low production capacity , Understand the workload 、 The difficulty of the task and other actual situations , Analyze problems and make targeted improvements
• Engineering
• R & D specifications ： Currently selected R & D efficiency measurement indicators , Submit small steps from the aspect of index design 、code review、 Good practices such as code reuse are more beneficial , Organize training within the team , Promote the implementation of R & D norms
• Industry research collaboration ： When insufficient upstream demand leads to low R & D workload , R & D personnel actively communicate with product managers , Mining products / The optimizable part of the technical framework

‍

These spontaneous practices not only improve the efficiency of team research and development , It also cultivates the ability to get feedback quickly 、 Timely checking 、 Self driven improved team culture .

‍

R & D quality ： Organization level and team level in-depth co construction

‍

According to the screening conditions given above , The R & D quality management team selected several R & D teams with dozens of people to pilot . The following takes a pilot team as an example , Introduce the R & D quality practice jointly built by the R & D quality management team and the team .

‍

2021 Annual volume data shows , The one-time smoke passing rate of the pilot team continued to decline , High defect density .

‍

After the pilot , The R & D quality management team goes deep into the pilot team as a coach , Cooperate to carry out the improvement practice of three iterations . in the meantime , The R & D quality management team maintains high-frequency communication with the team .

‍

Alignment cognition

‍

Based on previous research , The definition of serious defects and other concepts is well understood , The R & D quality management team shall clarify the improvement objectives .

‍

Lay stress on formulating norms

‍

According to the actual needs of business stages and characteristics , The R & D quality management team and the team jointly select the key points from the list of general R & D quality requirements , Formulate feasible process specifications 、 Templates 、checklist、 Version rhythm suggestions and other normative materials .

Take the requirements document as an example ： For high stability 、 Long running time 、 The core system relied on by many other systems , Focus on whether to describe in detail the impact of the change on the associated system in the demand analysis and review process 、 Impact on historical data 、 Exception handling logic 、 Authority control, etc .

Requirements inspection items , Highlight as a key concern

In the process of R & D , The R & D personnel take the initiative to conduct self-examination against the materials , Reduce dependence on personal abilities through standardization , Reduce the cost of practice .

‍

Dig the root cause , Targeted improvement

‍

The R & D quality management team and the R & D team participate in the iteration review and recheck , Trace the root cause of the problem . Relevant members need to formulate improvement measures and prevention plans for similar problems , The R & D quality management team will assist in reviewing the plan .

‍

for example , For the pilot team, the passing rate of smoking at one time decreased 、 The problem of high defect density , The review found that the development team proposed that the R & D quality fluctuated greatly , The time of the test is tight and the pressure is heavy , The quality of R & D is low . To address this issue , Mainly in the following 2 Implement improvement in three aspects ：

‍

• Code review ： Ask team members to review according to the review template and specification formulated in the previous link , The review task is clearly assigned to each developer . For more complex requirements 、 Modules with low smoke passing rate in the previous time , It will be reviewed again at the acceptance meeting . Hard requirements ： The evaluation can only be carried out after the evaluation problems are solved .
• Development self testing ： The scope of developer self-test includes unit tests , And the smoking case provided by the test , Strictly implement cross testing and review , Push the test to the left .

‍

The R & D team and the R & D quality management team will continue to follow up the implementation of improvement measures , Until the relevant R & D quality indicators tend to be normal .

‍

In the pilot cycle of three iterations , The participation of R & D quality management team is gradually reduced . After the pilot , R & D quality management team personnel withdraw , The team QA To take over , Lead the practice of R & D quality and continue to implement .

‍

According to the pilot situation , The awareness of the R & D team has been significantly improved , Be able to actively find problems and actively improve ： The team QA Be able to formulate improvement strategies independently 、 Sort out the guiding materials applicable to the team ; Developers actively participate in R & D quality building , Carry out R & D quality backtracking of the whole R & D process .

‍

The methodology is constantly precipitated through practice

‍

After the pilot , The R & D quality management team will pay a regular return visit to understand the landing situation , On the one hand, give timely suggestions , On the other hand, problems are constantly found through cases 、 Sort out the improvement plan , Continuously verify the effectiveness of improvement measures , Accumulate excellent practice , Prepare for a larger scale of R & D efficiency construction .

‍

It should be noted that , The so-called improvement cannot only focus on the fluctuation of data . The data only reflect the effect of improvement , Not the purpose of improvement .

‍

R & D quality management group reference MARI methodology , Established a common problem analysis and improvement practice framework ：

‍

• After finding problems with metrics , Analyze and interpret the data from multiple perspectives , Locate key weaknesses
• Combined with other related indicators and survey methods , Ask the root cause , Locate R & D performance bottlenecks and optimization opportunities
• Translate these insights into clarity 、 Executable 、 Verifiable improvement plan , Standardize the R & D process 、 Establish a good R & D culture
• Continuously measure and verify the improvement effect , Flexibly adjust the improvement plan

‍

The following three common problems in the R & D process are taken as examples , Show the above frame ：

‍

problem 1： Between the various links of R & D , Failed to align information about requirements

• Root cause ：
• The content of the requirements document is too simple , User scenarios 、 The business process 、 Function module 、 Exception handling and so on are not explained in detail
• The implementation of requirements review is not in place , Lack of record , The problems found in the review are not followed up
• Not synchronized in time after the demand changes
• Practical advice ：
• Based on GWT（Given+When+Then） Provide a clear description of the business scenario , Including forward flow and abnormal flow , Clarify the acceptance rules
• Add pre-trial link , Demand analysts need to think ahead , Prepare various instructions against the inspection items , Strengthen information tracking and confirmation in the review process
• Strengthen the standardization of requirements documents , For example, regularly recheck defects , If the root cause of the defect is the omission of information in the requirements document , Then it is suggested to add corresponding requirements inspection items
• Establish information synchronization channels for all links

‍

problem 2： Other teams responsible for the associated system , Failed to align information about requirements

• Root cause ：
• The requirement scheme was designed without sufficient communication with the associated system team
• The content of the requirements document is too simple , The possible impact of this change on the associated system is not described in detail
• The review process lacks the participation of related system teams in mutual evaluation
• Practical advice ：
• Fully communicate with the related system team in the demand analysis process , Jointly confirm the responsibilities of all parties 、 Boundaries and ways of collaboration
• The requirement review link clearly states whether it will affect other systems , Whether the demand scheme has reached a consensus with related parties , What is the calling relationship between systems , Who is the main responsibility system , Whether it can be launched independently , Relevant scheduling and responsible person of the associated system
• If no consensus has been reached , Then the demand analysis members need to continue to follow up , The person in charge of requirements or the project manager shall coordinate with the related system team , Inform that development and testing need to be jointly promoted

‍

problem 3： Insufficient self-test in the development stage , Cause too much Bug Flow into the test segment

• Root cause ： Complex business scenarios , Not enough consideration , The implementation does not meet the requirements , Low self-test coverage
• Practical advice
• Strengthen the whole process communication , Test shift left , Reduce the pressure in the testing process and the cost of problem repair in the later stage
• In the design review link , Development and testing work together to sort out and confirm with requirements , Clear business process - Functional design - Correspondence of data model , Clarify the consistency between the design scheme and the requirements
• Before the key requirements are tested, the development organization showcase, Confirm the consistency between implementation and requirements by requirements
• Root cause ： The quality of code testing and R & D is not high , Insufficient self-test
• It is suggested that developers strengthen code review before test + Static code scan
• It is required to develop smoke test cases for self-test coverage

‍

The R & D quality management team combed the common problems in all links of the R & D process 、 Root cause analysis and practical suggestions , And precipitate into the knowledge assets of the organization , Continue to iterate in cooperation with the front-line R & D team .

‍

06 expectation

‍

The insurance group will continue to explore the following aspects to continuously improve R & D efficiency ：

‍

• Expand the scope of R & D performance measurement to the whole process of software R & D and more positions

Hope to make the product 、 Operation and maintenance 、 More software R & D related positions such as data are included in the R & D Efficiency Measurement System , Further improve the transparency of R & D process .

‍

One side , Conduct a more in-depth evaluation of the R & D efficiency of each single point from a global perspective , Avoiding local optimization will have a negative impact on global optimization ; On the other hand , Focus on end-to-end value delivery of software development , avoid “ R & D efficiency shaft ”, Make each product 、 project 、 The R & D efficiency of the team is improved and the business value at the organizational level 、 Authors efficiency 、 Customer satisfaction and other business achievements , Drive business acceleration with lean thinking .

‍

• Optimize the data model of R & D efficiency , Output enlightening engineering improvement suggestions

It is hoped that the R & D efficiency special team and the front-line R & D team will continue to work closely , Continuous precipitation of excellent practice , Enrich the knowledge system of R & D effectiveness , Thus, a scenario based R & D efficiency data model is established , Not only output normative improvement suggestions , Enlightening suggestions for improvement are also given . for example , Comprehensively analyze the complexity of requirements 、 Change impact scale 、 Business priority and other dimensions , Assist the demander in prioritizing , Reasonably suggest the splitting and scheduling of requirements , Achieve more efficient iterative development .

‍

• Strengthen cooperation between R & D efficiency and R & D quality management team , According to the characteristics of the R & D team , Output quality and efficiency comprehensive improvement plan

At a lower cost 、 Higher R & D efficiency 、 The process is controllable 、 The result can be measured as the goal , Based on the main line of R & D efficiency , Take delivery inspection as the process , Take data measurement as the starting point , Implement special optimization , Create a benchmarking ecosystem for R & D Efficiency Management . Output practical improvement solutions that meet their personalized needs to the R & D team , Focus on delivering R & D efficiency 、 Deliver R & D quality 、 Ability to deliver , Build a comprehensive digital road of R & D efficiency .

------ end ------

Smart and easy Merico R & D effectiveness analysis platform , Committed to helping the R & D team solve R & D Efficiency 、 R & D quality and talent development Three big pain points , Improve R & D efficiency and software engineering quality ;

Welcome to communicate with us in the comment area ！