Analysis of software development costs helps software organizations understand their own development process, improve estimation accuracy, control costs, and provide important decision-making information for improvement.
According to the difficulty and the depth of cost analysis, software development cost analysis can be divided into:
a) Cost estimation deviation: Analyze the deviation between estimated cost and actual cost and the reason.
b) Cost composition: To analyze the composition of development costs, you can classify costs in the following ways:
——For example, it is divided into direct costs and indirect costs, among which direct and indirect costs are further divided into human costs and non-human costs;
—— According to the life cycle, the cost is
divided into phases such as requirements, design, coding, testing, and delivery; ——Divided into development cost, project management cost, quality assurance cost, etc. according to the type of work.
Through the statistics of the above classification and combined with the information of "cost estimation deviation", we can understand the main reasons for the deviation of cost estimation.
c) Correlation analysis of key cost influencing factors: A variety of factors in the software development process have different effects on costs, and these factors have different degrees of influence in each organization. Therefore, through analysis of the relationship between various factors and costs in the development process Relationships can allow managers to understand the focus of cost control in the organization.
Correlation analysis commonly used methods include scatter plot and Pearson correlation analysis.
d) Regression analysis of cost estimation equation: In addition to understanding the degree of influence of key influencing factors of cost, managers of software development activities also hope to obtain a quantitative relationship between cost and key influencing factors, that is, the estimation equation of cost and key influencing factors.
After the project is over, costs and related data are of great value to the organization and should be collected and analyzed. The purpose and perspective of the analysis include:
a) Project evaluation: According to the cost estimation deviation and composition, evaluate the ability of the project team's budget control and the efficiency of process execution.
b) Establish or correct the cost estimation model: As mentioned above, the cost estimation equation regression analysis will produce new costs and related data after the project ends. These data can be used to evaluate the effect of the regression equation and help to continuously optimize the regression equation. .
c) Process improvement: understand the problems of the development process by analyzing the cost distribution ratio and the deviation rate of various activity cost estimates. Combining these data with experience and understanding of the organization can provide managers with process improvement information.
Project scale, workload, duration, cost and other estimates and actual data should also be stored in the benchmark database established within the organization for future project teams and organizations. The actual use includes:
a) Provide reference for similar project estimates;
b ) Establish, evaluate and optimize the cost estimation model;
c) Carry out correlation analysis on quality issues;
d) Calculate the benchmark cost per unit scale;
e) Analyze the cost ratio of each activity of the organization.
The organization can also submit the data of the project team to the industry benchmark database to provide support for the continuous update of industry benchmark data.