Anyone who has done a project knows that it is not easy to collect and clarify requirements , especially to dig out the detailed and deep-level requirements of the demand side.
When many companies conduct demand research, often due to differences in problem description and understanding between the two parties, large deviations occur in the process of continuous transmission of requirements, resulting in the final developed functions being quite different from the original business requirements. Business personnel can't explain it clearly, and technical personnel don't understand it, so the final development result cannot meet the needs of real business scenarios.
So, how can we do a detailed research on requirements so that the requirements in real business scenarios can be accurately communicated to the final developers?
To sum up, there are two points: grasp the overall thinking and principles, and do a good job in three key links.
By grasping the pain points layer by layer, let the management and business personnel clarify their needs, that is, the project boundary, so that the development of IT personnel will not deviate from the direction. Even if the BI system cannot guarantee a perfect fit for the requirements in the end, but the core requirements are met and the BI system can be used in the enterprise, the project is not considered a failure.
First share the complete plan with everyone, and if you need to read it in detail, go directly to:
Link 1: Investigate the analysis scenarios of business departments
Before investigating the analysis scenarios of business departments, the first thing to do is to determine the business departments that need to be investigated according to the users of the BI system. You can investigate all the business departments that want to use the BI system at one time, or multiple round-robin survey.
For each department that needs to be researched, a corresponding data linker and business linker should be designated, and of course the same person can also serve concurrently. The project manager needs to conduct research on the "data analysts" and "business personnel" of each business department separately, and cooperate to complete the data sorting.
Among them, specific business scenario demand research can be collected from the following three levels.
The first is the management level. It mainly investigates the analysis needs of indicators related to corporate strategy. The method is to disassemble the strategic goals of enterprises and departments layer by layer, and then analyze from the perspective of data supporting strategic goals to obtain demand data.
For example, from the dismantling of a company's corporate strategy to the departments that support strategic goals
- This department supports the OKR of strategic goals and the corresponding business actions;
- The measurement indicators and measurement dimensions formulated to measure the business action;
- Whether the data currently has a data storage location, etc.
Through step-by-step disassembly to confirm the data sheets that need to be developed, analyze and record:
The second is to investigate the needs of business departments in some daily analysis scenarios:
Finally, some hidden needs of business departments are investigated. These needs are different from daily analysis scenarios and need to be discovered through brainstorming or interviews:
After completing these requirements surveys, you can summarize all the collected scenario requirements based on the principles of scenario dimension indexing and data systematization.
For example, after the BI project team of a fashion company conducts demand research on various business departments, it summarizes the demand into an Excel table based on dimensions such as type, demand index, index definition and formula, data granularity, commodity/channel, data frequency, and data source. , and on the basis of indexing the scene dimension, sort out the data tables, and finally form the data index system of the enterprise.
Link 2: Investigate data quality
The data in the enterprise is mainly divided into business system data, manually collected data, external data, etc. according to the source .
- When conducting research on business system data, the BI project team needs to identify who is connected to each business system, and obtain relevant data interfaces and data dictionaries. If they cannot be obtained, they need to negotiate and formulate coping strategies.
- For manual data, the project team can first collect historical manual data, and this work can be carried out simultaneously with the demand research of the business department.
- For external data, you can refer to the research method of business system data, focusing on data availability and usage scenarios.
During this process, the project manager needs to communicate clearly with the IT staff of the information department on the following two points:
-
Co-organize existing database data and confirm data quality
-
Both parties identify what will meet the needs and what will not. Unsatisfiable needs need to be returned to the business for re-investigation, and satisfying needs are directly provided or new tables are provided.
In addition, it is worth emphasizing that at the stage of researching data quality, it is also necessary to clearly define the core architectural data of projects such as organizational structure, users, and data authority systems. Permissions include not only module function permissions, but also data permissions, that is, which data can be seen by different users and roles. For example, the city sales manager can see the sales data of the city he is responsible for, while the regional sales manager can only see the sales data of the area he is responsible for.
Among them, the allocation of authority by person supports allocation by department and role, which is suitable for companies with different structures.
Assigning data permissions can also be divided and marked through Excel. The permission requirements document of a group headquarters is as follows:
Link 3: Design, confirm and modify the data system
In addition, when designing the data system, two levels, the original table and the basic wide table, are mainly considered, combined with the minimum granularity of the data usage requirements considered in the previous research, and the data that may be used in the analysis. Dimensions and indicators, as far as possible to achieve full coverage of analysis scenarios, to meet various data granularity requirements.
The confirmation and modification of the data system mainly includes the addition/deletion/modification of data dimensions, indicators, and granularity, and the unification of field meanings and logical calibers. After completing the confirmation and modification, the project team also needs to output a demand research confirmation letter, which can enter the next stage only after being approved by the project leadership committee and each team.
Summary
Requirements collection is an important part of the project construction process. Doing it according to the above links can save at least 30% of time and manpower. Complete project construction includes what to do, who should do it, how to do it , and how to successfully operate the BI project in the enterprise to actual business value. It can be realized with specific BI construction tools .